Cri 2017 38 issue 1 by Hogrefe

Volume 38 / Number 1 / 2017

Editor-in-Chief Diego De Leo Associate Editors Maria A. Oquendo Ella Arensman

Crisis

The Journal of Crisis Intervention and Suicide Prevention

Published under the auspices of the International Association for Suicide Prevention (IASP)

An innovative and highly effective brief therapy for suicidal patients “ASSIP is perhaps the most significant innovation we have seen in the assessment and treatment of suicidal risk...” David A. Jobes, PhD, Professor of Psychology, The Catholic University of America, Washington, DC, USA Past President, American Association of Suicidology

Konrad Michel / Anja Gysin-Maillart

ASSIP – Attempted Suicide Short Intervention Program A Manual for Clinicians

2015, x + 114 pp. US $59.00 / € 41.95 ISBN 978-0-88937-476-8 Also available as eBook Attempted suicide is the main risk factor for suicide. The Attempted Suicide Short Intervention Program (ASSIP) described in this manual is an innovative brief therapy that has proven in published clinical trials to be highly effective in reducing the risk of further attempts. ASSIP is the result of the authors’ extensive practical experience in the treatment of suicidal individuals. The emphasis is on the therapeutic alliance with the suicidal patient, based on an initial patientoriented narrative interview.

www.hogrefe.com

The four therapy sessions are followed by continuing contact with patients by means of regular letters. This clearly structured manual starts with an overview of suicide and suicide prevention, followed by a practical, step-by-step description of this highly structured treatment. It includes numerous checklists, handouts, and standardized letters for use by health professionals in various clinical settings.

Crisis The Journal of Crisis Intervention and Suicide Prevention

Volume 38, No. 1, 2017 Published under the Auspices of the International Association for Suicide Prevention (IASP)

Editor-in-Chief

Diego De Leo, MD, PhD, DSc, FRANZCP, Australian Institute for Suicide Research and Prevention, WHO Collaborating Centre on Research and Training in Suicide Prevention, Griffith University, Mt. Gravatt Campus, 4122 Queensland, Australia (Tel. +61 7 3735-3382, Fax +61 7 3735-3450, E-mail D.DeLeo@griffith.edu.au)

Associate Editors

Maria A. Oquendo, MD, New York State Psychiatric Institute, 1051 Riverside Drive, New York, NY 10032, USA (Tel. +1 212 543-5835, Fax +1 212 543-6017, E-mail mao4@columbia.edu) Ella Arensman, PhD, National Suicide Research Foundation, Department of Epidemiology and Public Health, University College Cork,Western Gateway Building, Room 4.34,Western Road, Cork, Ireland (Tel. +353 21 420-5551 or +353 87 052-2284, E-mail earensman@ucc.ie)

Editorial Board

A. Apter, Tel Aviv, Israel R. Battegay, Basel, Switzerland M. P. Battin, Salt Lake City, UT, USA A. L. Beautrais, Christchurch, New Zealand A. L. Berman, Washington, DC, USA S. S. Canetto, Fort Collins, CO, USA J. Cutcliffe, Tyler, TX, USA O. Ekeberg, Oslo, Norway A. Erlangsen, Aarhus, Denmark M. Goldblatt, Boston, MA, USA D. Gunnell, Bristol, UK K. Hawton, Oxford, UK L. M. Hayes, Mansfield, MA, USA K. van Heeringen, Gent, Belgium H. Hjelmeland, Trondheim, Norway E. Isometsa, Helsinki, Finland G. Jessen, Odense, Denmark N. Kapur, Manchester, UK N. Kapusta, Vienna, Austria A. J. F. M. Kerkhof, Amsterdam, The Netherlands M. M. Khan, Karachi, Pakistan

Past Editors-in-Chief

Raymond Battegay, Annette L. Beautrais, David C. Clark, John F. Connolly, Ad J. F. M. Kerkhof, Hermann Pohlmeier

Responsible Organization Publisher

The journal is published under the auspices of the International Association for Suicide Prevention (IASP) (IASP Central Administrative Office, National Centre for Suicide Research and Prevention, Sogsvannsveien 21, Building 12, N-0372 Oslo, Norway, Tel. +47 22 923715, E-mail admin@iasp.info, www.iasp.info).

Production

Juliane Munson, Hogrefe Publishing, Merkelstr. 3, 37085 Göttingen, Germany Tel. +49 551 99950-0, Fax +49 551 99950-111, E-mail juliane.munson@hogrefe.com

Subscriptions

Hogrefe Publishing, Herbert-Quandt-Str. 4, 37081 Göttingen, Germany Tel. +49 551 99950-900, Fax +49 551 99950-998

Advertising/Inserts

Marketing, Hogrefe Publishing, Merkelstr. 3, 37085 Göttingen, Germany, Tel. +49 551 99950-0 Fax +49 551 99950-111, E-mail marketing@hogrefe.com

ISSN

ISSN-L 0227-5910, ISSN-Print 0227-5910, ISSN-Online 2151-2396

© 2017 Hogrefe Publishing. This journal as well as the individual contributions and illustrations contained within it are protected under international copyright law. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without prior written permission from the publisher. All rights, including translation rights, reserved.

Subscription prices

Annual subscription (2017, 6 issues): Institutions US $379.00 / € 298.00; Individuals US $179.00 / € 128.00. Postage and handling: US $24.00 / € 18.00. Single issues: US $63.00 / € 49.50 (+ postage and handling)

Payment

Payment may be made by check, international money order, or credit card, to Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, or, for North American customers, to Hogrefe Publishing Corp., Journals Department, 7 Bulfinch Place, 2nd floor, Boston, MA 02114, USA.

Electronic Full Text

The full text of Crisis is available online at www.econtent.hogrefe.com

Abstracting Services

The journal is abstracted/indexed in Index Medicus/Medline, Current Contents/Social and Behavioral Sciences (CC/S&BS), Social Sciences Citation Index (SSCI), Social SciSearch, PsycINFO, PSYNDEX, Scopus, and EMCare. Impact Factor (2015): 1.527.

Crisis (2017), 38(1)

K. Kõlves, Brisbane, Australia A. A. Leenaars, Windsor, Canada J. L. McIntosh, South Bend, IN, USA J. Mann, New York, NY, USA M. Marttunen, Helsinki, Finland L. Mehlum, Oslo, Norway B. L. Mishara, Montreal, Québec, Canada T. Niederkrotenthaler, Vienna, Austria S. Occhipinti, Nathan, Australia (Statistical Advisor) R. O’Connor, Glasgow, UK M. Phillips, Shanghai, China J. Pirkis, Melbourne, Australia M. Pompili, Rome, Italy A. Preti, Cagliari, Italy M. Silverman, Chicago, IL, USA S. Stack, Detroit, MI, USA L. Vijayakumar, Chennai, India M. Voracek, Vienna, Austria D. Wasserman, Stockholm, Sweden E. J. de Wilde, Rotterdam, Netherlands Paul Yip, Hong Kong SAR, China

Hogrefe Publishing GmbH, Merkelstr. 3, 37085 Göttingen, Germany Tel. +49 551 99950-0, Fax +49 551 99950-111, E-mail publishing@hogrefe.com,Web http://www.hogrefe.com North America: Hogrefe Publishing, 7 Bulfinch Place, Suite 202, Boston, MA 02114, USA, Tel. (866) 823-4726, Fax (617) 354-6875, E-mail customerservice@hogrefe.com, Web http://www.hogrefe.com

Published in 6 issues per annual volume.

Contents Editorial Research Trends

Suicide Prevention in an International Context: Progress and Challenges Ella Arensman

Relationship Satisfaction and Risk Factors for Suicide Benedikt Till, Ulrich S. Tran, and Thomas Niederkrotenthaler

Newspaper Reporting on a Cluster of Suicides in the UK: A Study of Article Characteristics Using PRINTQUAL Ann John, Keith Hawton, David Gunnell, Keith Lloyd, Jonathan Scourfield, Phillip A. Jones, Ann Luce, Amanda Marchant, Steve Platt, Sian Price, and Michael S. Dennis

Characteristics and Proximal Outcomes of Calls Made to Suicide Crisis Hotlines in California: Variability Across Centers Rajeev Ramchand, Lisa Jaycox, Pat Ebener, Mary Lou Gilbert, Dionne Barnes-Proby, and Prodyumna Goutam

Analysis of Internet Suicide Pacts Reported by the Media in Mainland China Fang-Fan Jiang, Hui-Lan Xu, Hui-Ying Liao, and Ting Zhang

Promoting Help Seeking to Veterans: A Comparison of Public Messaging Strategies to Enhance the Use of the Veterans Crisis Line Elizabeth Karras, Naiji Lu, Heather Elder, Xin Tu, Caitlin Thompson, Wendy Tenhula, Sonja V. Batten, and Robert M. Bossarte

Erratum Correction to Till et al., 2016 News, Announcements, and IASP

ÂŠ 2017 Hogrefe Publishing

63 64

Crisis (2017), 38(1)

A unique and comprehensive handbook presenting the state-of-art in suicide bereavement support t Ou April 2017

Karl Andriessen / Karolina Krysinska / Onja T. Grad (Editors)

Postvention in Action

The International Handbook of Suicide Bereavement Support 2017, approx. xvi + 422 pp. pre-order price US $94.00 / â&#x201A;Ź 74.95 ISBN 978-0-88937-493-5 Suicide is not merely the act of an individual; it always has an effect on others and can even increase the risk of suicide in the bereaved. The International Association for Suicide Prevention, the World Health Organisation, and others have recognized postvention as an important strategy for suicide prevention. This unique and comprehensive handbook, authored by nearly 100 international experts, including researchers, clinicians, support group facilitators, and survivors, presents the state-of-theart in suicide bereavement support. The first part examines the key concepts and the processes that the bereaved experience and illustrates them with illuminating clinical vignettes. The second and third parts

www.hogrefe.com

look in detail at suicide bereavement support in all the most relevant settings (including general practices, the workplace, online and many others) as well as in specific groups (such as health care workers). In the concluding section, the support provided for those bereaved by suicide in no less than 23 countries is explored in detail, showing that postvention is becoming worldwide strategy for suicide prevention. These chapters provide useful lessons and inspiration for extending and improving postvention in new and existing areas. This unique handbook is thus essential reading for anyone involved in suicide prevention or postvention research and practice.

Editorial Suicide Prevention in an International Context Progress and Challenges Ella Arensman President, International Association for Suicide Prevention (IASP) Director of Research, National Suicide Research Foundation, Department of Epidemiology and Public Health, University College Cork, Ireland WHO Collaborating Centre for Surveillance and Research in Suicide Prevention, Cork, Ireland

Strategic Global Developments in Suicide Prevention In recent years, the World Health Organization (WHO) Global Mental Health Action Plan, 2013–2020, has been a major step forward in pushing the agenda of suicide prevention globally (WHO, 2013; Saxena, Funk, & Chisholm, 2013). This plan was adopted by health ministers in all 194 WHO member states to formally recognize the importance of mental health, which was a remarkable achievement. Among WHO member states are 25 countries where suicide is currently still criminalized and an additional 20 countries where according to Sharia law suicide attempters may be punished with jail sentences (Mishara & Weisstub, 2016). The action plan covers specified actions to improve mental health and to contribute to the attainment of a set of agreed global targets, in particular to aim for (a) a 20% increase in service coverage for severe mental disorders, and (b) a 10% reduction of the suicide rate in countries by 2020. The subsequent publication of the WHO report Preventing Suicide: A Global Imperative, in 2014 (WHO, 2014), was strategically a major and timely next step to increase the commitment of national governments and health ministers to move from agreement to action in relation to suicide prevention. Many members of the International Association for Suicide Prevention (IASP) representing all regions in the world were involved in preparing this report. IASP in collaboration with WHO’s Department of Mental Health and Substance Abuse has initiated workshops inviting country representatives to discuss and share experiences in the development and implementation of national suicide prevention programs, during IASP world congresses © 2017 Hogrefe Publishing

and regional seminars. In addition, IASP is in the process of establishing an International Special Interest Group to support the development and implementation of national suicide prevention programs at a global level. In all six WHO regions, both IASP and WHO underline the importance of national suicide prevention programs on World Suicide Prevention Day on a yearly basis. The WHO report provides guidance in developing and implementing national suicide prevention programs while taking into account the different stages at which a country is, that is, countries where suicide prevention activities have not yet taken place, countries with some activities, and countries that currently have a national response. Within geographic regions, countries that have adopted a national suicide prevention program can impact positively on surrounding countries and increase prioritization of suicide prevention in countries that do not yet have a national program and do not want be an exception in a negative sense, that is, they do not want to be left behind!

Are We Making Progress in Suicide Prevention at Global Level? We are indeed! Since the publication of the WHO Global Mental Health Action Plan and the WHO report on preventing suicide, there are several indications that the development and the implementation of national suicide prevention programs have accelerated, in particular in countries and regions where so far little or no suicide prevention initiatives were present, such as Guyana (Ministry of Public Health, 2014), Suriname (Ministerie van VolksgezondCrisis (2017), 38(1), 1–6 DOI: 10.1027/0227-5910/a000461

Editorial

heid, 2016), and Bhutan (Royal Government of Bhutan, 2015). In addition, IASP has supported an initiative by the Ministry of Public Health in Afghanistan to develop a national suicide prevention program, supported by a multisectoral advisory group, which was again reinforced by the WHO global report on preventing suicide. However, liaising with partners and stakeholders in suicide prevention in Afghanistan, in particular arranging face-to-face meetings, is challenging due to infrastructural limitations and the ongoing adversity resulting from conflicts and war. A further incentive is the fact that the WHO report has been translated in all six UN languages. Regional launches have been held in Mexico with representatives from Spanish-speaking countries, in Cairo, with representatives from the WHO Eastern Mediterranean Region, and in Tokyo, with representatives from the WHO Western Pacific Region. Furthermore, a growing number of countries have recently completed their second national suicide prevention program, such as England (Department of Health, 2012), Scotland (The Scottish Government, 2013), Ireland (Department of Health, 2015), and the United States (US Department of Health and Human Services, 2012). A particular positive development was the recently published national suicide prevention plan for Guyana (2015– 2020; Ministry of Public Health, 2014). The overall age-standardized rate of suicide for Guyana in 2012 was 44.2 per 100,000, the highest reported suicide rate in the world. This represents an extremely high rate of 70.8 per 100,000 for men and a relatively high rate, within an international context, of 22.1 per 100,000 for women. Due to concerns about the current high suicide rates and the commitment of the Departments of Health and Public Health in Guyana, and the Pan American Health Organisation/ World Health Organization (PAHO/WHO), a comprehensive multisectoral 5-year national suicide prevention action plan was prepared. The strategy incorporates activities across the continuum of suicide prevention representing universal interventions, targeting the entire population to reduce access to means and reducing inappropriate media coverage of suicide. In addition, selective interventions, targeting high-risk groups of suicide, and indicated interventions, targeting individuals who show signs of symptoms that are strongly associated with suicide, for example, suicide prevention helplines and peer support networks for those with suicide ideation and suicide attempt, are included. The strategy is based on cross-cutting values and principles: (a) universal health coverage; (b) human rights; (c) evidence-based practice – and interventions for treatment and prevention; (d) life course approach; and (e) multisectoral approach. Considering that a suicide attempt in Guyana is still considered a criminal law offence with the consequence that the person involved may be liable to imprisonment for 2 Crisis (2017), 38(1), 1–6

years (Mishara & Weisstub, 2016), the publication of the national suicide prevention plan is a significant achievement. Hopefully, thanks to increased awareness and stigma reduction, which is a key objective of the national plan, the legal status of suicide and attempted suicide in Guyana will be revisited as a matter of urgency.

IASP-WHO Global Survey on Suicide and Suicide Prevention On the basis of a global survey, conducted by IASP and the WHO Department of Mental Health and Substance Abuse in 2013, IASP national representatives of 90 countries (57%) completed the survey questionnaire, attaining information on national strategies and activities in suicide prevention. In nearly two thirds (61%) of the responding countries, suicide was perceived as a significant public health concern. In 31% of the countries a comprehensive national strategy or action plan was adopted by the government. Among the countries that did not have a national strategy, a number of suicide prevention activities were carried out in just over half (52%) of the countries, which included training on suicide risk assessment and intervention (38%), training for general practitioners (26%), and suicide prevention training for non-health professionals including first responders, teachers, and journalists (37%). A unique contribution of this survey was that for some regions across the world, such as the WHO Eastern Mediterranean and African regions, where previously information on suicide prevention activities was limited or absent, new information was obtained. For example, in 40% of the responding countries in the Eastern Mediterranean Region a training program on suicide assessment and intervention for GPs was available, and in 20% of the countries in this region, training programs were available on suicide prevention for non-health professionals. A detailed overview of the survey outcomes for the different geographic regions is currently being prepared for publication in an IASP monograph.

Challenges Ahead Despite the progress in suicide prevention globally, we still face numerous challenges. The accuracy and reliability of suicide statistics are an ongoing issue of concern in a considerable number of countries (Tollefsen, Hem, & Ekeberg, 2012). In terms of implementing national suicide prevention programs and the sustainability of interventions, a number of challenges remain. On-going challenges include insufficient resources, ineffective co-ordination, lack of en© 2017 Hogrefe Publishing

Editorial

forcement of guidelines, limited access to surveillance data on suicide and attempted suicide or self-harm, and lack of independent and systematic evaluations (WHO, 2014). In addition, it would be important for a national suicide prevention program to address real-time developments, such as emerging suicide contagion and clustering, emerging methods of suicide, and new vulnerable and high-risk groups, such as migrants and refugees from Eastern Mediterranean countries, with increased risk of suicide and self-harm.

Typical Components of a National Suicide Prevention Program and the Evidence Base In terms of the content of a national suicide prevention program, the WHO global report recommends a systematic approach and summarizes typical components (WHO, 2014). Even though these components are supported by evidence, the strength and consistency of the evidence for some of the components/interventions in reducing suicide and attempted suicide or self-harm vary across different studies.

Surveillance Increasing the quality and timeliness of national data on suicide and suicide attempts/self-harm is a core component of a national suicide prevention program, in particular establishing integrated data collection systems that serve to identify vulnerable groups, individuals, and situations. WHO recently published a practice manual for establishing and maintaining surveillance systems for suicide attempts and self-harm. This manual aims to provide guidance to countries who wish to set up a public health surveillance system for suicide attempts and self-harm cases presenting to general hospitals (WHO, 2016). While there is a lack of reliable national data on the prevalence of suicide attempts/self-harm presentations to hospital emergency departments in low- and middle-income countries (LMICs) and the demographic and psychosocial profile of those involved, surveillance and follow-up of this high-risk group could be an initial step toward building a national suicide prevention program (Fleischmann et al., 2016).

Restricting Access to Means This involves implementation of measures to reduce availability of and access to frequently used means of suicide, ÂŠ 2017 Hogrefe Publishing

for example, pesticides, drugs, firearms, enhancing safety of bridges etc. Internationally, there is consistent evidence that restricting access to lethal means is associated with a decrease in suicide and that substitution to other methods is limited (Zalsman et al., 2016). In addition, evidence from 18 studies showed a consistent reduction of suicide following restricted access and increased safety of sites where people frequently took their lives (Pirkis et al., 2015). Fleischmann et al. (2016) further reported consistent findings supporting implementation of this intervention in LMICs.

Media This refers to implementing guidelines to enhance responsible reporting of suicide in print, broadcast, Internet, and social media. The role of the mass media has been shown to be effective in reducing stigma and increasing help-seeking behavior. There are also indications of promising results based on multilevel suicide prevention programs (Niederkrotenthaler, Reidenberg, Till, & Gould, 2014). A systematic review covering 30 studies on social media sites for suicide prevention (Robinson et al., 2016) showed that social media platforms can reach large numbers of individuals and may allow others to intervene following expression of suicidal behavior. However, reported challenges include lack of control over user behavior, possibility of suicide contagion, limitations in accurately assessing suicide risk, and issues relating to privacy and confidentiality. The importance of responsible media reporting of suicide in LMICs is underlined by Fleischmann et al. (2016). It must be noted, however, that evaluation of the effectiveness of this intervention in LMICs is required.

Training and Education Educating health-care and community-based professionals to recognize depression and early signs of suicidal behavior is important for determining the level of care and referral for treatment and subsequent prevention of suicidal behavior (Coppens et al., 2014; Wasserman et al., 2012). Sustainability and capacity building of trainers and benefits in terms of knowledge, attitudes, and confidence can be achieved via a train-the-trainer model (Coppens et al., 2014; Isaac et al., 2009). There are some indications for a link between improvements in intermediate outcomes (e.g., improved knowledge, attitudes, and confidence) among health-care and community-based professionals and primary outcomes, for example, reduced suicide and self-harm rates (Hegerl, Rummel-Kluge, VĂ¤rnik, Crisis (2017), 38(1), 1â&#x20AC;&#x201C;6

Editorial

Arensman, & Koburger, 2011; Mann et al., 2005; Zalsman et al., 2016). Fleischmann et al. (2016) underline the importance of training and education via the WHO mhGAP program in LMICs.

Treatment A recently published Cochrane systematic review on the effectiveness of psychosocial interventions for self-harm included 29 RCTs including 8,480 participants (Hawton et al., 2016). The most commonly evaluated intervention involved CBT-based psychological therapy with a duration of an average of 10 sessions. At follow-up, people who had received CBT were significantly less likely to have engaged in repeated self-harm compared with those receiving treatment as usual. For people with a history of multiple self-harm episodes, dialectical behavior therapy was identified as reducing the frequency of repeated selfharm, but did not reduce the proportion of individuals repeating self-harm. However, the number of RCTs conducted so far is relatively small. In addition, Zalsman et al. (2016) found consistent evidence for the effectiveness of lithium in reducing suicidal behavior among people with mood disorders. Even though there are indications for the eligibility of these treatments in LMICs, national implementation may not be feasible due to the costs and lack of trained mental health professionals (Fleischmann et al., 2016).

Awareness and Stigma Reduction This refers to increasing awareness via public information campaigns to support the understanding that suicides are preventable, and increasing public and professional access to information about all aspects of preventing suicidal behavior. Stigma reduction involves promoting the use of mental health services and services for the prevention of substance abuse and suicide as well as reducing discrimination against people using these services (WHO, 2014). There is emerging evidence for these interventions from community-based multilevel interventions to improve the care for people diagnosed with depression and simultaneously address awareness and skills in early identification of suicide risk among health-care and community-based professionals (Hegerl et al., 2013; Szekely et al., 2013), with proven synergistic effects of simultaneously implementing evidence-based interventions (Harris et al., 2016). Due to the feasibility of implementing community-based multilevel interventions in culturally different countries, this approach is also eligible for implementation in LMICs (Fleischmann et al., 2016). Crisis (2017), 38(1), 1–6

Postvention Improving the response to and caring for those affected by suicide and suicide attempts is considered a key component of national suicide prevention programs (WHO, 2014). There is emerging evidence supporting beneficial effects of a number of interventions, including counseling postvention for survivors and outreach at the scene of a suicide (Szumilas & Kutcher, 2011). In addition, evidence-based guidelines for responding to suicide in a secondary school setting have been published recently (Cox et al., 2016). However, further research is required into the effectiveness of postvention services and interventions on reducing suicide and attempted suicide/selfharm.

Crisis Intervention and Access to Services This involves increasing the capacity of communities to respond to crises, such as emerging suicide clusters or murder–suicide, with appropriate interventions, including access to emergency mental health care for individuals in a crisis situation, through telephone helplines or the Internet (WHO, 2014). The systematic review by Zalsman et al. (2016) found inconsistent effects for crisis and follow-up interventions on suicide attempts and ideation, and they recommend further investigation into the effectiveness. However, Fleischmann et al. (2016) considered implementation of these interventions in LMICs albeit with ongoing monitoring of the feasibility and evaluation of the effects on reducing suicide and attempted suicide or selfharm. In addition to the components as recommended in the WHO report (WHO, 2014), the review by Zalsman et al. (2016) also reported that the quality of evaluation studies involving school-based programs has improved over the past decade since the review by Mann et al. (2005). There is increasing evidence from RCTs addressing mental health literacy, suicide risk awareness, and skills training in a secondary school setting and their impact on reduced suicide attempts and severe suicidal ideation (Zalsman et al., 2016). The inclusion of the suicide mortality rate as an indicator of the UN sustainable development goals (SDGs) for 2030 directs further attention to suicide and its prevention. The progress of suicide prevention at a global level and the evidence base for national suicide prevention programs will be an ongoing priority for IASP and a key topic on the program of the forthcoming 29th IASP World Congress, “Preventing Suicide: A Global Commitment, from Communities to Continents,” July 28–22 2017, in Kuching, Sarawak, Malaysia (http://www.iasp2017.org). © 2017 Hogrefe Publishing

Editorial

Acknowledgments The author would like to thank Dr. Shekhar Saxena and Dr. Alexandra Fleischmann, Department of Mental Health and Substance Abuse, WHO, for their collaboration on many IASP-WHO collaborative activities in suicide prevention. Dr. Fleischmann and Vanda Scott, IASP International Advisor, had a key role in conducting the IASP-WHO Global Survey on Suicide Prevention. Niall McTernan, Research Support Officer NSRF, assisted in reviewing national suicide prevention programs.

References Coppens, E., Van Audenhove, C., Iddi, S., Arensman, E., Gottlebe, K., Koburger, N., … Hegerl, U. (2014). Effectiveness of community facilitator training in improving knowledge, attitudes, and confidence in relation to depression and suicidal behavior: Results of the OSPI-Europe intervention in four European countries. Journal of Affective Disorders, 165, 142–150. doi:10.1016/j. jad.2014.04.052 Cox, G. R., Bailey, E., Jorm, A. F., Reavley, N. J., Templer, K., Parker, A., … Robinson, J. (2016). Development of suicide postvention guidelines for secondary schools: A Delphi study. BMC Public Health, 16, 180. doi:10.1186/s12889-016-2822-6 Department of Health. (2012). Preventing suicide in England – a cross-government outcomes strategy to save lives. London, UK: Author. Department of Health. (2015). National strategy to reduce suicide in Ireland, connecting for life, 2015–2020. Dublin, Ireland: Author. Fleischmann, A., Arensman, E., Berman, A., Carli, V., De Leo, D., Hadlaczky, G., … Saxena, S. (2016). Evidence-based interventions for suicide prevention in a global perspective. Global Mental Health, 3, e5. doi:10.1017/gmh.2015.27 Harris, F. M., Maxwel, l. M., O’Connor, R., Coyne, J. C., Arensman, E., Coffey, C., … Hegerl, U. (2016). Exploring synergistic interactions and catalysts in complex interventions: Longitudinal, mixed methods case studies of an optimised multi-level suicide prevention intervention in four European countries (Ospi-Europe). BMC Public Health, 16, 268. doi:10.1186/s12889-016-2942-z Hawton, K., Witt, K. G., Taylor Salisbury, T. L., Arensman, E., Gunnell, D., Townsend, E., … Hazell, P. (2016). Self-harm: A systematic review of the efficacy of psychosocial treatments for adults. Lancet Psychiatry, 3(8), 740–750. doi:10.1016/S22150366(16)30070-0 Hegerl, U., Rummel-Kluge, C., Värnik, A., Arensman, E., & Koburger, N. (2013). Alliances against depression – a community based approach to target depression and to prevent suicidal behaviour. Neuroscience and Biobehavioral Reviews, 37(10 Pt. 1), 2404–2409. doi:10.1016/j.neubiorev.2013.02.009 Isaac, M., Elias, B., Katz, L. Y., Belik, S. L., Deane, F. P., Enns, M. W., & Sareen, J. (2009). Gatekeeper training as a preventative intervention for suicide: A systematic review. Canadian Journal of Psychiatry, 54(4), 260–268. Mann, J. J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., … Hendin, H. (2005). Suicide prevention strategies: A systematic review. Journal of the American Medical Association, 294(16), 2064–2074. Ministry of Public Health. (2014). National suicide prevention plan. Georgetown, Guyana: Author. Ministerie van Volksgezondheid. (2016). Nationaal suicide preventie – en interventieplan, 2016–2020 [National suicide preven© 2017 Hogrefe Publishing

tion – an intervention plan, 2016–2020]. Paramaribo, Suriname: Author. Mishara, B. L., & Weisstub, D. N. (2016). The legal status of suicide: A global review. International Journal of Law and Psychiatry, 44, 54–74. Niederkrotenthaler, T., Reidenberg, D. J., Till, B., & Gould, M. S. (2014). Increasing help-seeking and referrals for individuals at risk for suicide by decreasing stigma: The role of mass media. American Journal of Preventive Medicine, 47(3 Suppl. 2), S235– 243. doi:10.1016/j.amepre.2014.06.010 Pirkis, J., Too, L. S., Spittal, M. J., Krysinska, K., Robinson, J., & Cheung, Y. T. (2015). Interventions to reduce suicides at suicide hotspots: A systematic review and meta-analysis. Lancet Psychiatry, 2(11), 994–1001. doi:10.1016/S2215-0366(15)00266-7 Robinson, J., Cox, G., Bailey, E., Hetrick, S., Rodrigues, M., Fisher, S., & Herman, H. (2016). Social media and suicide prevention: A systematic review. Early Intervention in Psychiatry, 10(2), 103– 121. doi:10.1111/eip.12229 Royal Government of Bhutan. (2015). Suicide prevention in Bhutan – a three year action plan (2015–2018). Thimphu, Bhutan: Author. Saxena, S., Funk, M., & Chisholm, D. (2013). World Health Assembly adopts comprehensive Mental Health Action Plan 2013–2020. Lancet, 381(9882), 1970–1971. Szekely, A., Konkoly, T. B., Mergl, R., Birkas, E., Rozsa, S., Purebl, G., & Hegerl, U. (2013). How to decrease suicide rates in both genders? An effectiveness study of a community-based intervention (EAAD). PLoS One, 8(9), e75081. doi:10.1371/journal. pone.0075081 Szumilas, M., & Kutcher, S. (2011). Post-suicide intervention programs: A systematic review. Canadian Journal of Public Health, 102(1), 18–29. The Scottish Government. (2013). Suicide prevention strategy 2013–2016. Edinburgh, UK: Author. Tollefsen, I. M., Hem, E., & Ekeberg, O. (2012). The reliability of suicide statistics: A systematic review. BMC Psychiatry, 12, 9. doi:10.1186/1471-244X-12-9 US Department of Health and Human Services (HHS). (2012). National strategy for suicide prevention: Goals and objectives for action. Washington, DC: HHS, Office of the Surgeon General and National Action Alliance for Suicide Prevention. Wasserman, D., Rihmer, Z., Rujescu, D., Sarchiapone, M., Sokolowski, M., Titelman, D., … Carli, V. (2012). The European Psychiatric Association (EPA) guidance on suicide treatment and prevention. European Psychiatry, 27(2), 129–141. doi:10.1016/j.eurpsy.2011.06.003 World Health Organization. (2013). Global mental health action plan, 2013–2020. Geneva, Switzerland: Author. World Health Organization. (2014). Suicide prevention – a global imperative. Geneva, Switzerland: Author. World Health Organization. (2016). Practice manual for establishing and maintaining suicide attempts and self-harm surveillance systems. Geneva, Switzerland: Author. Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone, M., … Zohar, J. (2016). Suicide prevention strategies revisited: 10-year systematic review. Lancet Psychiatry, 3(7), 646–659. doi:10.1016/S2215-0366(16)30030-X

Accepted January 11, 2017 Published online March 3, 2017

Crisis (2017), 38(1), 1–6

Editorial

About the author Prof. Arensman is President of the International Association of Suicide Prevention, Director of Research at the National Suicide Research Foundation, and Research Professor in the Department of Epidemiology and Public Health, University College Cork, Ireland. In all three roles, she actively encourages international collaboration in suicide research and prevention.

Crisis (2017), 38(1), 1â&#x20AC;&#x201C;6

Prof. Ella Arensmann 4.34 Western Gateway Building University College Cork Cork Ireland

ÂŠ 2017 Hogrefe Publishing

Research Trends

Relationship Satisfaction and Risk Factors for Suicide Benedikt Till1,2, Ulrich S. Tran2, and Thomas Niederkrotenthaler1 Suicide Research Unit, Institute of Social Medicine, Center for Public Health, Medical University of Vienna, Austria Department of Basic Psychological Research and Research Methods, School of Psychology, University of Vienna, Austria

1 2

Abstract. Background: Previous studies suggest that troubled romantic relationships are associated with higher risk factors for mental health. However, studies examining the role of relationship satisfaction in suicide risk factors are scarce. Aims: We investigated differences in risk factors for suicide between individuals with high relationship satisfaction, individuals with low relationship satisfaction, and singles. Furthermore, we explored patterns of experiencing, and dealing with, conflicts in the relationship and examined associations with suicide risk factors. Method: In this cross-sectional study, we assessed relationship status, relationship satisfaction, specific types of relationship conflicts, and suicide risk factors (i.e., suicidal ideation, hopelessness, depression) with questionnaires among 382 individuals in Austria. Results: Risk factors for suicide were higher among singles than among individuals in happy relationships, but lower among those with low relationship satisfaction. Participants reporting a high number of unsolved conflicts in their relationship had higher levels of suicidal ideation, hopelessness, and depression than individuals who tend to solve issues with their partner amicably or report no conflicts. Conclusion: Relationship satisfaction and relationship conflicts reflect risk factors for suicide, with higher levels of suicidal ideation, hopelessness, and depression reported by individuals who mentioned unsolved conflicts with their partner and experienced low satisfaction with their relationship. Keywords: suicidal ideation, romantic relationship, relationship satisfaction, relationship conflicts

An individual’s suicide is not determined by one single stressor (Wassermann, 2001). Suicidal behavior is considered to be the result of a combination of several risk factors (Mann et al., 2005), including biological, social, psychological, and cultural influences (Sonneck, Kapusta, Tomandl, & Voracek, 2012). Support from social relationships has been found to be an essential determinant of the maintenance of both physical and mental health (Holt-Lunstad, Birmingham, & Jones, 2008; House, Umberson, & Landis, 1988; Logan, Hall, & Karch, 2011; Rook, 1984; Whisman & Baucom, 2012), and social isolation has been identified as a major risk factor for morbidity and mortality comparable to well-established contributors such as smoking, obesity, and high blood pressure (Holt-Lunstad et al., 2008; House et al., 1988; Kiecolt-Glaser & Newton, 2001). Social support from a spouse appears particularly important, exceeding the positive impact of support from children or friends on an individual’s well-being (Antonucci, Lansford, & Akiyama, 2001; Okabayashi, Liang, Krause, Akiyama, & Sugisawa, 2004; Okun & Lockwood, 2003). Thus, suicidal behavior may be reflected in whether or not an individual is currently living in a happy romantic relationship. Happiness in a romantic relationship is known to be an important contributor to an individual’s global happiness and life satisfaction (Glenn & Weaver, 1981; Markey, Markey, & Gray, 2007) and one of the most powerful predictors of mental health (Gove, Hughes, & Style, 1983; Logan © 2016 Hogrefe Publishing

et al., 2011). Several studies have demonstrated that individuals who are married or live in marriage-like relationships are in better mental and physical health (Gove et al., 1983; Holt-Lunstad et al., 2008; Kolves, Ide, & De Leo, 2012), happier, and less inclined to suicide (Batterham et al., 2014; Gove et al., 1983; Kolves et al., 2012) and have lower morbidity and mortality (Kiecolt-Glaser & Newton, 2001) than individuals who are single. Accordingly, the loss of a spouse is known to have the potential to induce negative mental or physical health changes (Kiecolt-Glaser & Newton, 2001). Many authors argue that spouse and family provide personal private satisfactions that make life meaningful and rewarding for adults (Gove et al., 1983; Stack, 2000). It is also believed that support from a spouse can protect from the harmful effects of stress, highlighting that romantic relationships may serve as an effective means for psychological well-being (Coyne & DeLongis, 1986; Markey et al., 2007). Accordingly, several studies (Batterham et al., 2014; Stack, 1990; Wyder, Ward, & De Leo, 2009) demonstrated that rates of suicide and suicide attempts are higher among divorced or separated women and men in nearly all age groups compared with married individuals. The prevalence of divorce was also found to be the strongest predictor for suicide rates in many countries (Stack, 1992). Moreover, divorced people have higher levels of depression, morbidity, financial pressure, and risk of alcohol abuse Crisis (2017), 38(1), 7–16 DOI: 10.1027/0227-5910/a000407

(Stack, 2000). Of note, a study using data from 17 national surveys (Stack & Eshleman, 1998) showed that married persons are happier than individuals who cohabit but are not married. Stack and Eshleman (1998) concluded that marriage may affect happiness through the promotion of financial satisfaction and the improvement of health. More recent research (Whisman & Uebelacker, 2006) suggests that cohabiting individuals experience more distress than married couples, but do not differ in terms of suicidal ideation or perceived health. However, many researchers point out that being in a romantic relationship per se is not necessarily protective for mental or physical health (Carr, Freedman, Cornman, & Schwarz, 2014; Gove et al., 1983; Kazan, Calear, & Batterham, 2016; Kiecolt-Glaser & Newton, 2001). Troubled relationships are a prime source of stress and may also limit the ability to seek support in other relationships (Coyne & DeLongis, 1986; Kiecolt-Glaser & Newton, 2001). In several studies (e.g., Carr et al., 2014; Glenn & Weaver, 1981), psychological distress and low life satisfaction were associated with unhappiness in the relationship, and several authors (e.g., Antonucci et al., 2001; Beach, Fincham, & Katz, 1998; Whisman & Uebelacker, 2009; Whisman, Uebelacker, Tolejko, Chatav, & McKelvie, 2006) reported an association between depression and discord in the relationship, while trust in a spouse was associated with high life satisfaction. Furthermore, several studies demonstrated higher risk factors for physical and mental health, including higher levels of anxiety and depression, among individuals who reported low relationship quality than among those with high relationship quality (e.g., Gallo, Troxel, Matthews, & Kuller, 2003; Grewen, Girdler, & Light, 2005; Leach, Butterworth, Olesen, & Mackinnon, 2013; Robles & Kiecolt-Glase, 2003). Low-quality intimate partner relationships (Arcel, Mantonakis, Petersson, Jemos, & Kaliteraki, 1992), interpersonal conflict (Choi et al., 2010; Logan et al., 2011), and negative life events related to a romantic partner (Bagge, Glenn, & Lee, 2013; Logan et al., 2011) are also considered to be precipitating factors to suicide. The findings that low-quality relationships, often characterized by conflict, problems, and arguments, contribute to an increased risk of suicidality is also evidenced by the fact that these factors can create an environment of abuse and violence, which often lead to an increased risk of suicide (Kazan et al., 2016). Stressful romantic relationships were also found to be associated with increased mortality risk (Lund, Christensen, Nilsson, Kriegbaum, & Rod, 2014). In a recent study (Ivan, Koyanagi, Tyrovolas, & Haro, 2015), negative partner interactions in a romantic relationship were found to be associated with increased likelihood of depression, anxiety, and suicidal ideation in a nationally representative sample in Ireland, while high perceived Crisis (2017), 38(1), 7â&#x20AC;&#x201C;16

B. Till et al.: Relationship Satisfaction and Suicidality

support from the spouse was associated with low scores of depression and low perceived stress among married individuals from US American college sample (Dehle, Larsen, & Landers, 2001). A study among married and cohabiting US American adults demonstrated that individuals in discordant relationships experienced higher levels of distress, perceived their health as poorer, and were more likely to report suicidal ideation than people who were not in discordant relationships (Whisman & Uebelacker, 2006). An association between poor romantic relationship quality and increased risk of depression was also found in a longitudinal study among 4,642 US American adults (Teo, Choi, & Valenstein, 2013). Research also shows that individuals in troubled marriages often report lower happiness and life satisfaction than singles do (Glenn & Weaver, 1981). In a review of literature on marital integration and suicide, Stack (2000) noted that in a majority of suicide cases it was determined that the relationship quality of marriages was deteriorating in the year before the suicide. It is also interesting to note that in a longitudinal study, Bruce and Kim (1992) found that divorce increased the risk of a major depression in men not only by a factor of 45 times compared with happily married men, but also by a factor of 14.3 times compared with unhappily married men. Despite evidence for an association of relationship status and quality with physical and mental health, research on the association of relationship satisfaction with suicidality is scarce (Kazan et al., 2016). Most of the few extant studies used aggregate data to assess associations between suicide and divorce rates (e.g., Stack, 1990, 1992) or focused on associations between mental health and relationship quality or specific partner interactions, but not relationship satisfaction (e.g., Arcel et al., 1992; Ivan et al., 2015). While these terms are often used interchangeably (Fincham & Rogge, 2010), relationship satisfaction refers more to the subjective evaluation of oneâ&#x20AC;&#x2122;s actual relationship with regard to the prototype of a good relationship, whereas relationship quality is commonly referred to as the objective assessment of the relationship in terms of the presence or absence of specific characteristics or qualities (Hassebrauck & Fehr, 2002). Furthermore, to the best of our knowledge, there are no studies available that explored risk factors for suicide with regard to specific types or areas of conflicts in romantic relationships (e.g., sexuality, communication, finances, etc.). Some relationship conflicts may be more relevant for personal distress than others. The present study aimed to explore differences in terms of risk factors for suicide (i.e., suicidal ideation, hopelessness, depression) between individuals with high and with low relationship satisfaction as well as in individuals who are currently single. We hypothesized that scores for suicidal ideation, hopelessness, and depression are lowest among individuals in happy romantic relationships, higher among ÂŠ 2016 Hogrefe Publishing

B. Till et al.: Relationship Satisfaction and Suicidality

singles, and highest among individuals in unhappy relationships. Furthermore, we investigated whether suicide risk factors are linked to specific types or areas of conflicts in relationships.

Method Participants A total of 382 participants, including 208 females and 174 males, living in Austria at the time of data collection were recruited by undergraduate psychology students at the University of Vienna for course credit. Mean age was 32.64 years (SD = 15.16). In terms of highest completed school level, 4.5% (n = 17) of the participants had compulsory education, 12.8% (n = 49) completed apprenticeship training, 11.5% (n = 44) completed intermediate technical and vocational school, 52.9% (n = 202) were secondary/high school graduates, and 18.3% (n = 70) completed college or university. Compared with the Austrian population, individuals with college and high school education were over-represented in our sample (Austrian population: 23.2% compulsory education, 27.7% apprenticeship, 18.1% intermediate technical and vocational school, 15.0% high school, and 16.0% college; Statistik Austria, 2011).

Measures Relationship Status We asked the participants to indicate whether they were currently in a romantic relationship: yes = 1, no = 0. Relationship Assessment Scale This scale includes seven items (e.g., “In general, how satisfied are you with your relationship?”) that are rated on a 5-point scale ranging from 1 = low satisfaction to 5 = high satisfaction to measure general relationship satisfaction (Hendrick, 1988). This scale was only completed by participants in romantic relationships. Problem List This questionnaire consists of 23 self-report items that represent 23 different areas of potential conflict in romantic relationships (e.g., sexuality, communication, or finances). Participants rate the intensity of each potential problem with 0 = no conflicts in this area; 1 = conflicts in this area, but we can usually solve them; 2 = conflicts in this area, no solutions, frequent arguments; or 3 = conflicts in this area, but we rarely talk about it. All items on this scale are analyz© 2016 Hogrefe Publishing

ed individually; thus, no sum score is calculated (Hahlweg, 1996). This scale was only completed by participants in romantic relationships. Suicide Probability Scale This self-report measure assesses suicidal ideation with 36-items (e.g., “I feel people would be better off if I were dead”) that are rated on a 4-point scale ranging from 1 = none or a little of time to 4 = most or all at the time. A total weighed score was calculated based on the manual of the questionnaire (Cull & Gill, 1988). Beck Hopelessness Scale (BHS) The BHS is a 20-item scale that assesses hopelessness (Beck & Steer, 1988) and is a robust predictor of suicidal behavior (Beck, Brown, Berchick, Stewart, & Steer, 1990; McMillan, Gilbody, Beresford, & Neilly, 2007). All items (e.g., “My future seems dark to me”) were rated on a 6-point scale ranging from 1 = very false to 6 = very true. Erlanger Depression Scale This scale includes eight self-report items (e.g., “I want to cry”) that are rated on a 4-point scale ranging from 0 = completely wrong to 4 = exactly right to assess symptoms of depression (Lehrl & Gallwitz, 1983).

Data Analysis For each variable, sum scores across all respective items were calculated. We split our sample into one group with high relationship satisfaction (n = 104, Mdn = 33, IQR = 2, Min. = 30, Max. = 35) and one group with low relationship satisfaction (n = 117, Mdn = 26, IQR = 6.5, Min. = 7, Max. = 29) by the median (Mdn = 29) of the Relationship Assessment Scale scores (Hendrick, 1988). These two groups were contrasted with those participants who were currently not in a relationship (n = 156). See Table 1 for an overview of the groups’ means and standard deviations in terms of suicidal ideation, hopelessness, and depression. Mean differences between the three groups in each dependent measure were examined with F tests (analysis of variance), and significant differences between individual groups were investigated with Tukey tests. Ratings of items identifying areas of potential conflict in romantic relationships were subjected to latent class analysis to classify participants into distinct groups, based on their observed response patterns. LatentGOLD 4.5 was used for this analysis, treating item ratings as nominal categories. Latent class analysis is specifically suited to nominal data and avoids problems inherent in traditional factor analysis, such as having to rotate factors and having to assume continuous observed variables (Garson, 2009). We Crisis (2017), 38(1), 7–16

B. Till et al.: Relationship Satisfaction and Suicidality

Table 1. Descriptive statistics for suicide risk factors among participants with high and low relationship satisfaction and participants currently not in a relationship High relationship satisfaction (n = 104) Suicide risk factors

Low relationship satisfaction (n = 117) M

No relationship (n = 156) M

Suicidal ideation (α = 0.89)

42.64

8.03

51.69

18.10

49.85

12.70

Hopelessness (α = 0.89)

45.39

11.28

55.49

15.65

50.92

13.24

4.43

3.45

8.19

6.22

6.58

4.61

Depression (α = 0.82)

Note. Values are means (M), standard deviations (SD), and Cronbach’s α of the variables estimated with SPSS.

planned to fit models with one, two, three, and four latent classes to the data. However, models with four or more latent classes required the estimation of more parameters than were testable with regard to overall model fit resulting in a negative number of degrees of freedom. Hence, no model with four or more latent classes was finally fit to the data. Model fit was assessed with the Bayesian information criterion (BIC), the likelihood-ratio goodness of fit statistic (L2), and percentages of classification (see Breslau, Reboussin, Anthony, & Storr, 2005; Garson, 2009). The model with the best data fit was kept as final model. Latent class analysis also computes the probability of assignment to each of the latent classes for each participant. Based on the highest probability, participants are then assigned to one of the latent classes. Mean assignment probabilities (so-called posterior mean assignment probabilities) of class members in turn are informative with regard to classification reliability. High posterior mean assignment probabilities are indicative of high classification reliability, which suggests that a large majority of the individuals can be assigned to a particular class with high probability. Posterior mean assignment probabilities are reported for the final model. Differences in terms of suicidal ideation, hopelessness, and depression between latent classes were then assessed with analyses of variance using F tests. Tukey tests were conducted to investigate significant differences between individual groups.

Results The average duration of the participants’ relationships was 12.6 years (SD = 13.5) and ranged from less than 1 month to 64 years. There was no difference between individuals with high and low relationship satisfaction in terms of relationship duration, t(217) = 1.31, p = .19, and there was no association between relationship duration and suicidality, r(209) = −.05, p = .51, hopelessness, r(217) = .10, p = .15, and depression, r(217) = −.07, p = .30. The analyses of variance revealed a significant difference between the three groups (high relationship satisCrisis (2017), 38(1), 7–16

faction, low relationship satisfaction, no relationship) in terms of suicidal ideation, F(2, 351) = 13.26, p < .001, hopelessness, F(2, 368) = 14.93, p < .001, and depression, F(2, 368) = 15.96, p < .001. Inspection of the means and contrast tests indicated that suicidal ideation was lower among individuals with high relationship satisfaction than among those with low relationship satisfaction (p < .001) and among individuals who were currently not in a relationship (p < .001). Hopelessness (p < .01) and depression (p < .01) were higher among individuals who were currently not in a relationship than among individuals with high relationship satisfaction. By contrast, hopelessness (p < .05) and depression (p < .05) were lower among singles than among individuals with low relationship satisfaction. Overall, suicidal ideation, hopelessness, and depression were lowest among individuals with high relationship satisfaction, higher among individuals who were currently not in a relationship, and highest among individuals with low relationship satisfaction. Overall, temperament (n = 44, 19.4%), communication (n = 39, 17.4%), habits (n = 39, 17.3%), sexuality (n = 36, 16.0%), and housekeeping (n = 35, 15.5%) were the areas of relationship conflicts where respondents reported most frequently unsolved or unspoken conflicts. Assault was the least frequently reported area with such conflicts (n = 8, 3.5%). Sadly, this still corresponded to eight individuals who were currently in a violent relationship. Table 2 provides an overview of the frequencies of conflicts in all areas of potential conflicts in a romantic relationship assessed in the present study. A latent class model with three clusters fitted best based on model fit indices in order to differentiate clusters of conflict areas. See Table 3 for an overview of the fit indices of the latent class models. The latent class analysis revealed that participants could be differentiated less with regard to conflicts in specific areas of relationships, but rather with regard to the overall quantity and quality of conflicts in the relationships: Participants of Cluster 1 (n = 95, 42.0%; mean posterior assignment probability = 0.97) had a high probability of reporting no conflicts in any area of their relationship indicated by a rating of 0 on all or most items of the problem list (Hahlweg, 1996). Thus, we con© 2016 Hogrefe Publishing

B. Till et al.: Relationship Satisfaction and Suicidality

Table 2. Overview of conflict areas in relationships among study participants (n = 226) Conflict type

No conflicts

Conflicts solved

Conflicts unsolved

Finances

169 (75.8%)

37 (16.6%)

14 (6.3%)

No talk about conflict 3 (1.3%)

Professional life

160 (71.1%)

41 (18.2%)

18 (8.0%)

6 (2.7%)

Housekeeping

100 (44.2%)

91 (40.3%)

29 (12.8%)

6 (2.7%)

Parenting

125 (56.6%)

67 (30.3%)

16 (7.2%)

13 (5.9%)

86 (38.1%)

111 (49.1%)

20 (8.8%)

9 (4.0%)

Social activities

134 (59.3%)

67 (29.6%)

16 (7.1%)

9 (4.0%)

Temperament

108 (47.8%)

74 (32.7%)

34 (15.0%)

10 (4.4%)

Devotion

120 (53.1%)

75 (33.2%)

16 (7.1%)

15 (6.6%)

Attractiveness

170 (75.6%)

46 (20.4%)

4 (1.8%)

5 (2.2%)

Leisure activities

Trust

175 (77.4%)

37 (16.4%)

7 (3.1%)

Jealousy

126 (55.8%)

75 (33.2%)

16 (7.1%)

9 (4.0%)

Freedom

127 (56.2%)

74 (32.7%)

10 (4.4%)

15 (6.6%)

Sexuality

126 (56.0%)

63 (28.0%)

13 (5.8%)

23 (10.2%)

Extramarital relationships

194 (87.0%)

17 (7.6%)

6 (2.7%)

Family

128 (56.6%)

65 (28.8%)

22 (9.7%)

11 (4.9%)

86 (38.1%)

101 (44.7%)

33 (14.6%)

6 (2.7%)

Communication

104 (46.4%)

81 (36.2%)

26 (11.6%)

13 (5.8%)

Wish for a child

180 (80.7%)

25 (11.2%)

7 (3.1%)

11 (4.9%)

Lack of acceptance/support

145 (64.4%)

57 (25.3%)

13 (5.8%)

10 (4.4%)

Demands

126 (56.0%)

76 (33.8%)

17 (7.6%)

6 (2.7%)

Illness/handicap

179 (79.2%)

36 (15.9%)

4 (1.8%)

7 (3.1%)

Individual habits

Alcohol/drug use

163 (72.1%)

42 (18.6%)

11 (4.9%)

10 (4.4%)

Assaults

207 (91.6%)

11 (4.9%)

5 (2.2%)

3 (1.3%)

Note. Values are frequencies (n) and percentages (%) of conflict types across all potential areas of relationship conflicts.

ceptualized Cluster 1 as the no conflicts class. Participants of Cluster 2 (n = 87, 38.5%; mean posterior assignment probability = 0.96) tended to report conflicts in some areas of their relationships, but also reported that these conflicts were usually amicably solved as indicated by predominant ratings of 0 or 1 on the items of the problem list. On the basis of this observation, we conceptualized Cluster 2 as the conflicts with solutions class. Participants of Cluster 3 (n = 44, 19.5%; mean posterior assignment probability = 0.99) had a high probability of reporting conflicts in the relationship that caused frequent arguments or remained unaddressed as indicated by frequent ratings of 2 and 3 on the items of the problem list. We conceptualized this cluster as the conflicts without solutions class. Devotion of the partner (Item 8), communication (Item 17), lack of acceptance or support by the partner (Item 19), and demands by the partner (Item 20) contributed most to the separation of the three clusters, as these areas of conflict were among the most frequently reported relationship problems in Clusters 2 and 3. The analyses of variance revealed a significant difference between the three clusters in terms of suicidal ideaÂŠ 2016 Hogrefe Publishing

tion, F(2, 212) = 29.64, p < .001, hopelessness, F(2, 220) = 22.38, p < .001, and depression, F(2, 220) = 21.12, p < .001. Inspection of the means and contrast tests indicated that suicidal ideation, hopelessness, and depression were higher among participants reporting unsolved conflicts in their relationship (Cluster 3) than among participants of Cluster 1 (suicidal ideation = p < .001, hopelessness = p < .001, depression = p < .001) and Cluster 2 (suicidal ideation = p < .001, hopelessness = p < .001, depression = p < .001) who showed similar scores (suicidal ideation = p = .21, hopelessness = p = .11, depression = p = .47). See Table 4 for an overview of the cluster means and standard deviations in terms of suicidal ideation, hopelessness, and depression. Of note, analysis of crosstabs revealed a significant correlation between being in an unhappy relationship and Cluster 3. Individuals who reported unsolved conflicts in their relationship had a higher probability of experiencing low satisfaction with their partner (j = .47, n = 221, p < .001).

Crisis (2017), 38(1), 7â&#x20AC;&#x201C;16

B. Till et al.: Relationship Satisfaction and Suicidality

Table 3. Fit of the latent class models Model 1-Cluster

BIC

Classification error, %

9710.27

7088.97

157

< .001

–

2-Cluster

9352.96

6352.22

< .001

0.78

3-Cluster

9320.28

5940.10

< .001

3.01

Note. BIC = Bayes information criterion. L2 = likelihood ratio test statistic. Cell entries list the fit of alternative 1-, 2-, and 3-cluster models with regard to 2 three indices of model fit. Lower values of BIC, L , and classification error indicate a better model fit.

Table 4. Descriptive statistics for suicide risk factors among participants of the three clusters of relationship conflicts Cluster 2: conflicts with solutions (n = 87)

Cluster 1: no conflicts (n = 95) Suicide risk factors

Cluster 3: conflicts without solutions (n = 44) M

Suicidal ideation

42.37

8.24

45.74

9.53

60.93

23.67

Hopelessness

46.08

11.89

50.11

11.15

62.49

19.19

4.91

4.56

5.80

4.03

10.76

7.17

Depression

Note. Values are means (M) and standard deviations (SD) of the variable estimated with SPSS.

Discussion The results of the present study show that relationship status and relationship satisfaction are both reflected in risk factors for suicide. In our study, suicidal ideation, hopelessness, and depression were higher among singles than among individuals in happy romantic relationships, but lower than among those in unhappy relationships. These findings are concordant with evidence of previous studies that people in romantic relationships are generally happier, less depressed, and in better mental health than individuals who are single (e.g., Carr et al., 2014; Glenn & Weaver, 1981; Gove et al., 1983; Holt-Lunstad et al., 2008) or divorced or live separated (Batterham et al., 2014; Kolves et al., 2012; Stack, 1990, 1992, 2000; Wyder et al., 2009), and that this association is influenced by relationship quality, with levels of happiness being lower among individuals in unhappy relationships than among individuals in happy relationships and in singles (e.g., Arcel et al., 1992; Bagge et al., 2013; Choi et al., 2010; Gallo et al., 2003; Glenn & Weaver, 1981; Grewen et al., 2005; Leach et al., 2013; Logan et al., 2011; Lund et al., 2014; Whisman & Uebelacker, 2006, 2009). A theory that may explain the association between low relationship satisfaction and suicidality is the interpersonal theory of suicidal behavior (Joiner, 2005; Van Orden et al., 2010). According to this theory, an individual must have both the desire and the ability to die by suicide in order to complete suicide. The desire to die is developed when people perceive themselves as a burden to family, friends, or society and when they feel socially alienated, also referred to as low belongingness (Joiner, 2005; Van Orden et al., 2010). Living in an unhappy relationship Crisis (2017), 38(1), 7–16

may generate or exacerbate a feeling of social alienation and may intensify perceived burdensomeness and result in increased suicidality (Kazan et al., 2016). Accordingly, relationship problems were associated with increased suicide risk in previous studies (Stack, 2000). The finding of the present study suggests that being in a romantic relationship can positively contribute to life satisfaction and protect from suicidal ideation, but may also constitute a major stressor when individuals become dissatisfied with their relationship. Although individuals in romantic relationships may have numerous psychological and social advantages over singles, these benefits may be limited to individuals in happy relationships (Coyne & DeLongis, 1986; Kiecolt-Glaser & Newton, 2001), as risk factors for suicide were higher among individuals in unhappy relationships than among singles. It is also important to note that many epidemiological studies investigating risk factors for suicide control for individuals’ relationship status, but are unable to include data on relationship satisfaction (e.g., Agerbo, Nordentoft, & Mortensen, 2002; Niederkrotenthaler, Floderus, Alexanderson, Rasmussen, & Mittendorfer-Rutz, 2012). The present study highlights that relationship status per se is not a sufficient indicator for suicidality. In our study, we also identified three different types of relationships in terms of quantity and quality of conflicts between the partners (i.e., relationships with solutions to conflicts, relationships without solutions to conflicts, relationships without conflicts). Individuals who frequently experience conflicts with their partners and have no solutions for their conflicts reported higher levels of suicidal ideation, hopelessness, and depression than did individu© 2016 Hogrefe Publishing

B. Till et al.: Relationship Satisfaction and Suicidality

als who rarely have conflicts with their partners or amicably solve their conflicts. Individuals experiencing unsolved conflicts with their partner were also less satisfied with their relationship. Fincham and Beach (1999) noted that marital conflict has profound implications for individual well-being and mental health, including increased probability of developing major depression, eating disorders, or addictions. However, the findings of the present study suggest that conflicts in relationships are per se not necessarily detrimental; only conflicts that are not adequately solved by the partners appear to convey a clear negative effect. Unsolved issues in a relationship likely increase the number of arguments and disputes between partners and may lead to continuous dissatisfaction and frustration, which may subsequently increase risk factors for suicide. Similarly, withdrawal during conflict by either or both partners was associated with more negativity in relationships, which correlated with low relationship quality and thoughts about divorce in a large sample of an American survey (Stanley, Markman, & Whitton, 2002). The strong and consistent association between intimate partner violence and suicidality was also recently highlighted by Kazan et al. (2016). These findings highlight that positive communication and the ability to handle conflicts constructively are essential components for relationship satisfaction. The provision of a safe place to address issues in treatment for couples as well as efforts to help couples develop reliable methods for talking safely and openly at home has been identified as a central aspect of couple therapy in previous research (Stanley et al., 2002). In the present study it seemed that, overall, whether couples were able to solve their issues was more relevant for their relationship satisfaction and well-being than the particular area of the conflicts. However, communication problems and disagreements in terms of the partner’s displayed devotion, support, and demands contributed most to the separation of the three clusters, suggesting that relationship issues related to these topics are most detrimental for relationship satisfaction among those conflicts that were assessed in the present study. Future research on relationship satisfaction should particularly pay attention to how partners rate solutions to conflicts related to communication and disagreements in terms of the partner’s displayed devotion, support, and demands. The present study has some limitations. First, the participants were not representative of the total population, with an overrepresentation of female and young individuals who had finished secondary or higher education. Thus, the findings of the present study may not generalize to the general public. Furthermore, owing to the cross-sectional and correlational design of this study, causality related to associations between relationship satisfaction and suicide risk factors cannot be assessed. A further limitation of the © 2016 Hogrefe Publishing

study may be that some items of the scale used to assess suicidal ideation are related to the concept of burdensomeness (Joiner et al., 2009), a key determinant of suicidality in Joiner’s interpersonal theory of suicidal behavior (Joiner, 2005; Van Orden et al., 2010), which may explain the association between low relationship satisfaction and suicidality. Future studies on the associations between relationship satisfaction and suicidality may want to exclude items related to burdensomeness from their measure of suicidal ideation and use these items as separate predictor variables. Another limitation was that the term romantic relationship was not explicitly explained to the participants and that there was no opportunity for participants to indicate that they are currently living in open or multiple relationships. Moreover, we did not collect data on participants’ sexual orientation. These factors may influence the prevalence of different types of conflicts and their associations with relationship satisfaction and suicidality. Therefore, the results of the present study should be interpreted with caution with regard to generalization to individuals living in nontraditional romantic relationships. Finally, no data were collected assessing which particular type of relationship (married, engaged, seriously dating, etc.) the participants were currently involved in. However, recent research suggests that associations of relationship satisfaction and conflicts with suicide risk factors do not vary with regard to type of relationship (Whisman & Uebelacker, 2006).

Conclusion Our findings indicate that individuals’ relationship status and their satisfaction with their partner may be associated with risk factors for suicide. Suicidal ideation, hopelessness, and depression were higher among singles than among individuals in happy relationships, but lower than among those with low relationship satisfaction. The present study highlights that being single may be more beneficial in terms of psychological well-being and suicide risk than staying in an unhappy relationship. Furthermore, risk factors for suicide were higher among partners who reported unsolved conflicts with their partners, compared with those who rarely have conflicts or tend to solve their conflicts amicably. These findings suggest that well-functioning relationships may be a protective factor for suicide, while unsolved conflicts and dissatisfaction in the relationship may be linked to increased suicide risk. From the standpoint of the goal of attaining a high level of psychological well-being, it is a dilemma that humans depend so heavily upon romantic relationships for their happiness, which are known to be particularly fragile (Glenn & Weaver, 1981). At least, this seems to be true for young, eduCrisis (2017), 38(1), 7–16

cated females, who represented the majority of our study sample. The ability to amicably solve conflicts seems to be a key factor for happiness and life satisfaction, as unsolved relationship conflicts were associated with low relationship satisfaction and high levels of risk factors for suicide in the present study. Public health interventions may benefit from focusing on romantic relationship satisfaction and dysfunction in order to prevent a range of mental health problems (Ivan et al., 2015), including suicidal ideation. Couple therapy and other strategies to improve relationship satisfaction (Jacobson & Addis, 1993; Jacobson, Christensen, Prince, Cordova, & Eldridge, 2000) may have the potential to lead to improvements in couple interaction and mental well-being. Further investigations examining the role of social relationships on risk factors for suicide are warranted, especially studies that use a longitudinal approach to assess the impact of relationship satisfaction on suicidality and suicidal behaviors. Furthermore, more research is necessary to evaluate associations between relationship satisfaction and suicide risk in different sociodemographic groups and to investigate differences with regard to sexual orientation, since young, educated female populations were overrepresented in the study sample and no data on sexual orientation were collected. Intimate partner relationships need to be emphasized more in suicide research. Finally, as low relationship satisfaction is associated with increased suicide risk, follow-up studies exploring how to prevent suicide among individuals in unhappy relationships are warranted.

References Agerbo, E., Nordentoft, M., & Mortensen, P. B. (2002). Familial, psychiatric, and socioeconomic risk factors for suicide in young people: Nested case-control study. British Medical Journal, 325, 74. Antonucci, T. C., Lansford, J. E., & Akiyama, H. (2001). Impact of positive and negative aspects of marital relationships and friendships on well-being of older adults. Applied Developmental Science, 5, 68–75. Arcel, L. T., Mantonakis, J., Petersson, B., Jemos, J., & Kaliteraki, E. (1992). Suicide attempts among Greek and Danish women and the quality of their relationships with husbands or boyfriends. Acta Psychiatrica Scandinavica, 85, 189–195. Bagge, C. L., Glenn, C. R., & Lee, H.-J. (2013). Quantifying the impact of recent negative life events on suicide attempts. Journal of Abnormal Psychology, 122, 359–368. Batterham, P. J., Fairweather-Schmidt, A. K., Butterworth, P., Calear, A. L., Mackinnon, A. J., & Christensen, H. (2014). Temporal effects of separation on suicidal thoughts and behaviours. Social Science & Medicine, 111, 58–63. Beach, S. R. H., Fincham, F. D., & Katz, J. (1998). Marital therapy in the treatment of depression: Toward a third generation of therapy and research. Clinical Psychology Review, 18, 635–661. Beck, A. T., Brown, G., Berchick, R. J., Stewart, B. L., & Steer, R. A. (1990). Relationship between hopelessness and ultimate suiCrisis (2017), 38(1), 7–16

B. Till et al.: Relationship Satisfaction and Suicidality

cide: A replication with psychiatric outpatients. American Journal of Psychiatry, 147, 190–195. Beck, A. T., & Steer, R. A. (1988). Manual for the Beck Hopelessness Scale. San Antonio, TX: Psychological Corporation. Breslau, N., Reboussin, B. A., Anthony, J. C., & Storr, C. L. (2005). The structure of posttraumatic stress disorder: Latent class analysis in 2 community samples. Archives of General Psychiatry, 62, 1343–1351. Bruce, M. L., & Kim, K. M. (1992). Differences in the effects of divorce on major depression in men and women. American Journal of Psychiatry, 149, 914–917. Carr, D., Freedman, V. A., Cornman, J. C., & Schwarz, N. (2014). Happy marriage, happy life? Marital quality and subjective well-being in later life. Journal of Marriage and Family, 76, 930–948. Choi, K. H., Wang, S. M., Yeon, B., Suh, S. Y., Oh, Y., Lee, H. K., & Lee, K. U. (2013). Risk and protective factors predicting multiple suicide attempts. Psychiatry Research, 3, 957–961. Coyne, J. C., & DeLongis, A. (1986). Going beyond social support: The role of social relationships in adaptation. Journal of Consulting and Clinical Psychology, 54, 454–460. Cull, J. G., & Gill, W. S. (1988). Suicide Probability Scale (SPS). Los Angeles, CA: WPS. Dehle, C., Larsen, D., & Landers, J. E. (2001). Social support in marriage. American Journal of Family Therapy, 29, 307–324. Fincham, F. D., & Beach, S. R. H. (1999). Conflict in marriage: Implications for working with couples. Annual Review of Psychology, 50, 47–77. Fincham, F. D., & Rogge, R. (2010). Understanding relationship quality: Theoretical challenges and new tools for assessment. Journal of Family Theory & Review, 2, 227–242. Gallo, L. C., Troxel, W. M., Matthews, K. A., & Kuller, L. H. (2003). Marital status and quality in middle-aged women: Associations with levels and trajectories of cardiovascular risk factors. Health Psychology, 22, 453–463. Garson, G. D. (2009). Statnotes: Topics in multivariate analysis: Latent class analysis. Retrieved from http://faculty.chass.ncsu. edu/garson/PA765/latclass.htm Glenn, N. D., & Weaver, C. N. (1981). The contribution of marital happiness to global happiness. Journal of Marriage and the Family, 43, 161–168. Gove, W. R., Hughes, M., & Style, C. R. (1983). Does marriage have positive effects on the psychological well-being of the individual? Journal of Health and Social Behavior, 24, 122–131. Grewen, K. M., Girdler, S. S., & Light, K. C. (2005). Relationship quality: Effects on ambulatory blood pressure and negative affect in a biracial sample of men and women. Blood Pressure Monitoring, 10, 117–124. Hahlweg, K. (1996). Fragebogen zur Partnerschaftsdiagnostik (FDP). Partnerschaftsbogen PFB, Problemliste PL und Fragebogen zur Lebensgeschichte und Partnerschaft FLP [Inventory for marital assessment (FDP). Partnership Questionnaire PFB, Problem List PL, and Biography and Partnership Questionnaire FLP]. Göttingen, Germany: Hogrefe. Hassebrauck, M., & Fehr, B. (2002). Dimensions of relationship quality. Personal Relationships, 9, 253–270. Hendrick, S. S. (1988). A generic measure of relationship satisfaction. Journal of Marriage and Family, 50, 93–98. Holt-Lunstad, J., Birmingham, W., & Jones, B. Q. (2008). Is there something unique about marriage? The relative impact of marital status, relationship quality, and network social support on ambulatory blood pressure and mental health. Annals of Behavioral Medicine, 35, 239–244. House, J. S., Umberson, D., & Landis, K. R. (1988). Structures and processes of social support. Annual Review of Sociology, 14, 293–318. Ivan, S. Z., Koyanagi, A., Tyrovolas, S., & Haro, J. M. (2015). The association of relationship quality and social networks with depres© 2016 Hogrefe Publishing

B. Till et al.: Relationship Satisfaction and Suicidality

sion, anxiety, and suicidal ideation among older married adults: Findings from a cross-sectional analysis of the Irish longitudinal study on aging (TILDA). Journal of Affective Disorders, 179, 134–141. Jacobson, N. S., & Addis, M. E. (1993). Research on couples and couple therapy: What do we know? Where are we going? Journal of Consulting and Clinical Psychology, 61, 85–93. Jacobson, N. S., Christensen, A., Prince, S. E., Cordova, J., & Eldridge, K. (2000). Integrative behavioral couple therapy: An acceptance-based, promising new treatment for couple discord. Journal of Consulting and Clinical Psychology, 68, 351–355. Joiner, T. E. (2005). Why people die by suicide. Cambridge, MA: Harvard University Press. Joiner, T. E., Van Orden, K. A., Witte, T. K., Selby, E. A., Ribeiro, J. D., Lewis, R., & Rudd, M. D. (2009). Main predictions of the interpersonal–psychological theory of suicidal behavior: Empirical tests in two samples of young adults. Journal of Abnormal Psychology, 118, 634–646. Kazan, D., Calear, A. L., & Batterham, P. J. (2016). The impact of intimate partner relationships on suicidal thoughts and behaviours: A systematic review. Journal of Affective Disorders, 190, 585–598. Kiecolt-Glaser, J. K., & Newton, T. L. (2001). Marriage and health: His and hers. Psychological Bulletin, 127, 472–503. Kolves, K., Ide, N., & De Leo, D. (2012). Fluctuations of suicidality in the aftermath of a marital separation: 6-Month follow-up observations. Journal of Affective Disorders, 142, 256–263. Leach, L. S., Butterworth, P., Olesen, S. C., & Mackinnon, A. (2013). Relationship quality and levels of depression and anxiety in a large population-based survey. Social Psychiatry and Psychiatric Epidemiology, 48, 417–425. Lehrl, S., & Gallwitz, A. (1983). Erlanger Depressions-Skala EDS [Erlangen depression scale EDS]. Vaterstetten, Germany: Vless. Logan, J., Hall, J., & Karch, D. (2011). Suicide categories by patterns of known risk factors: A latent class analysis. Archives of General Psychiatry, 68, 935–941. Lund, R., Christensen, U., Nilsson, C., J. Kriegbaum, M., & Rod, N. H. (2014). Stressful social relations and mortality: A prospective cohort study. Journal of Epidemiology and Community Health, 68, 720–727. Mann, J. J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Hass, A., … Hendin, H. (2005). Suicide prevention strategies – a systematic review. Journal of the American Medical Association, 294, 2064–2074. Markey, C., Markey, P., & Gray, H. (2007). Romantic relationships and health: An examination of individuals’ perceptions of their romantic partners’ influences on their health. Sex Roles, 57, 435–445. McMillan, D., Gilbody, S., Beresford, E., & Neilly, L. (2007). Can we predict suicide and non-fatal self-harm with the Beck Hopelessness Scale? A meta-analysis. Psychological Medicine, 37, 769–778. Niederkrotenthaler, T., Floderus, B., Alexanderson, K., Rasmussen, F., & Mittendorfer-Rutz, E. (2012). Exposure to parental mortality and markers of morbidity, and the risks of attempted and completed suicide in offspring: An analysis of sensitive life periods. Journal of Epidemiology and Community Health, 66, 233–239. Okabayashi, H., Liang, J., Krause, N., Akiyama, H., & Sugisawa, H. (2004). Mental health among older adults in Japan: Do sources of social support and negative interaction make a difference? Social Science & Medicine, 59, 2259–2270. Okun, M. A., & Lockwood, C. M. (2003). Does Level of assessment moderate the relation between social support and social negativity?: A meta-analysis. Basic and Applied Social Psychology, 25, 15–35. © 2016 Hogrefe Publishing

Robles, T. F., & Kiecolt-Glaser, J. K. (2003). The physiology of marriage: Pathways to health. Physiology & Behavior, 79, 409–416. Rook, K. S. (1984). The negative side of social interaction: Impact on psychological well-being. Journal of Personality and Social Psychology, 46, 1097–1108. Sonneck, G., Kapusta, N., Tomandl, G., & Voracek, M. (Eds.). (2012). Krisenintervention und Suizidverhütung [Crisis intervention and suicide prevention]. Vienna, Austria: Facultas. Stack, S. (1990). New micro-level data on the impact of divorce on suicide, 1959–1980: A test of two theories. Journal of Marriage and Family, 52, 119–127. Stack, S. (1992). Marriage, family, religion, and suicide. In R. W. Maris, A. L. Berman, J. T. Maltsberger, & R. I. Yufit (Eds.), Assessment and prediction of suicide (pp. 540–552). New York, NY: Guilford. Stack, S. (2000). Suicide: A 15-year review of the sociological literature part II: Modernization and social integration perspectives. Suicide and Life-Threatening Behavior, 30, 163–176. Stack, S., & Eshleman, J. R. (1998). Marital status and happiness: A 17-nation study. Journal of Marriage and Family, 60, 527–536. Stanley, S. M., Markman, H. J., & Whitton, S. W. (2002). Communication, conflict, and commitment: Insights on the foundations of relationship success from a national survey. Family Process, 41, 659–675. Statistik Austria. (2011). Educational attainment of the Austrian population between 25 and 64 years, from 1971 to 2012 [Data file]. Retrieved from http://www.statistik.at/web_de/statistiken/ bildung_und_kultur/bildungsstand_der_bevoelkerung/020912. html Teo, A. R., Choi, H., & Valenstein, M. (2013). Social relationships and depression: Ten-year follow-up from a nationally representative study. PLOS ONE, 8, e62396. Van Orden, K. A., Witte, T. K., Cukrowicz, K. C., Braithwaite, S. R., Selby, E. A., & Joiner, T. E. (2010). The interpersonal theory of suicide. Psychological Review, 117, 575–600. Wassermann, D. (2001). Suicide – an unnecessary death. London, UK: Dunitz. Whisman, M. A., & Baucom, D. (2012). Intimate relationships and psychopathology. Clinical Child and Family Psychology Review, 15, 4–13. Whisman, M. A., & Uebelacker, L.A. (2006). Impairment and distress associated with relationship discord in a national sample of married or cohabiting adults. Journal of Family Psychology, 20, 369–377. Whisman, M. A., & Uebelacker, L.A. (2009). Prospective associations between marital discord and depressive symptoms in middle-aged and older adults. Psychology and Aging, 24, 184–189. Whisman, M. A., Uebelacker, L. A., Tolejko, N., Chatav, Y., & McKelvie, M. (2006). Marital discord and well-being in older adults: Is the association confounded by personality? Psychology and Aging, 21, 626–631. Wyder, M., Ward, P., & De Leo, D. (2009). Separation as a suicide risk factor. Journal of Affective Disorder, 116, 208–213.

Received November 5, 2015 Revision received February 24, 2016 Accepted February 24, 2016 Published online July 22, 2016

About the authors Benedikt Till, PD DSc, is a psychologist and Assistant Professor at the Suicide Research Unit, Institute of Social Medicine, Center for Public Health, Medical University of Vienna, Austria. He works in the field of

Crisis (2017), 38(1), 7–16

media psychology, health communication, and suicide research and is board member of the Wiener Werkstaette for Suicide Research. Ulrich S. Tran, PD DSc, is a clinical psychologist and Senior Lecturer of Methods in Psychology at the Department of Basic Psychological Research and Research Methods, School of Psychology, University of Vienna, Austria. He is a board member of the Wiener Werkstaette for Suicide Research. Thomas Niederkrotenthaler, MD PhD MMS, is Associate Professor at the Suicide Research Unit of the Institute of Social Medicine, Vienna, Austria. He is currently co-chair of the Council of National Representatives to IASP, and founding chairman of the Wiener Werkstaette for Suicide Research (http://www.suizidforschung.at).

Crisis (2017), 38(1), 7â&#x20AC;&#x201C;16

B. Till et al.: Relationship Satisfaction and Suicidality

Benedikt Till Suicide Research Unit Institute of Social Medicine Center for Public Health Medical University of Vienna Kinderspitalgasse 15 A-1090 Vienna Austria Tel. +43 1 40160-34615 Fax +43 1 40160-934882 E-mail benedikt.till@meduniwien.ac.at

ÂŠ 2016 Hogrefe Publishing

Research Trends

Newspaper Reporting on a Cluster of Suicides in the UK A Study of Article Characteristics Using PRINTQUAL Ann John1,7, Keith Hawton2, David Gunnell3, Keith Lloyd1, Jonathan Scourfield4, Phillip A. Jones1, Ann Luce5, Amanda Marchant1, Steve Platt6, Sian Price7, and Michael S. Dennis1 Swansea University Medical School, Institute of Life Sciences 2, Swansea University, Swansea, UK Centre for Suicide Research, Department of Psychiatry, Warnford Hospital, Oxford, UK 3 School of Social and Community Medicine, University of Bristol, Bristol, UK 4 School of Social Sciences, Cardiff University, UK 5 Institute for Media and Communication Research, Bournemouth University, UK 6 Centre for Population Health Sciences, University of Edinburgh Medical School, UK 7 Public Health Wales National Health Service Trust, Cardiff, Wales, UK 1 2

Abstract. Background: Media reporting may influence suicide clusters through imitation or contagion. In 2008 there was extensive national and international newspaper coverage of a cluster of suicides in young people in the Bridgend area of South Wales, UK. Aims: To explore the quantity and quality of newspaper reporting during the identified cluster. Method: Searches were conducted for articles on suicide in Bridgend for 6 months before and after the defined cluster (June 26, 2007, to September 16, 2008). Frequency, quality (using the PRINTQUAL instrument), and sensationalism were examined. Results: In all, 577 newspaper articles were identified. One in seven articles included the suicide method in the headline, 47.3% referred to earlier suicides, and 44% used phrases that guidelines suggest should be avoided. Only 13% included sources of information or advice. Conclusion: A high level of poor-quality and sensationalist reporting was found during an ongoing suicide cluster at the very time when good-quality reporting could be considered important. A broad awareness of media guidelines and expansion and adherence to press codes of practice are required by journalists to ensure ethical reporting. Keywords: suicide cluster, newspaper reporting, guidelines

There is growing recognition that suicides may occur in clusters. Evidence suggests that mass clusters – when there is a temporary increase in the total frequency of suicides for a population relative to the time preceding and after the cluster, but with no spatial element – typically follow the reporting of actual or fictional suicides (Haw, Hawton, Niedzwiedz, & Platt, 2013). Notable examples have occurred following the fictional portrayal of suicide or self-harm in TV programs (Hawton et al., 1999) and celebrity suicides (Niederkrotenthaler et al., 2012). This copycat/imitation phenomenon (or Werther effect) is a modeling of suicidal behavior, with the media acting as a vehicle for contagion. Such suicides frequently involve the same method (Hawton & Williams, 2002; Pirkis & Blood, 2001; Sisask & Varnik, 2012; Stack, 2003). Currently the evidence for this effect is strongest in newspaper reporting (Hawton & Williams, 2005; Stack, 2005). The impact is most evident within the first 2 days of a report and over the next week (Bollen & Phil© 2016 Hogrefe Publishing

lips, 1982), although occasionally lasting longer (Fu & Yip, 2007). The prominence of the story and repetition of the reporting may be particularly influential (Niederkrotenthaler et al., 2010). Identification with the individual in the report or holding them in particular esteem can also influence impact (Pirkis & Nordentoft, 2011). Most importantly, particular population groups may be more vulnerable, especially younger people and those suffering from depression (Cheng et al., 2007). The effects of newspaper reporting on suicidal behavior, however, may not be entirely harmful. In a study of newspaper reporting in Austria, Niederkrotenthaler and colleagues (2010) showed that the reporting of suicidal ideation, not associated with subsequent attempted or completed suicide, may have a protective effect (Papageno effect). Recent evidence using space–time models suggests that up to 2% of probable suicides may occur in point clusters, with an excessive number of suicides occurring in close temporal and geographical proximity (Larkin & Beautrais, Crisis (2017), 38(1), 17–25 DOI: 10.1027/0227-5910/a000410

2012). Temporal–spatial suicide clusters are thought to be two to four times more common among young people (aged 15–24 years) than among other age groups (Niedzwiedz, Haw, Hawton, & Platt, 2014). There has been limited research on media influences in point clusters, although a recent case–control study (Gould, Kleiman, Lake, Forman, & Midle, 2014) of 48 suicide clusters in young people in the US showed that a variety of newspaper report characteristics were associated with the initiation of clusters. This study identified a variety of newspaper report characteristics that were associated with the initiation of the cluster (between first and second deaths), namely, front-page placement, detailed descriptions of the suicidal individual and act, and headlines containing the word suicide. In light of earlier evidence that highlights the important negative associations between reporting and suicide clusters, recommendations for responsible reporting by journalists have been promoted by the World Health Organization, 2005), and in many countries by suicide prevention organizations. In the UK the Samaritans have produced guidance (2008, revised 2013). Press Codes of Ethics similarly encourage ethical reporting and are able to hold organizations and journalists accountable for their actions (in the UK the Press Complaints Commission Code of Practice and the Independent Press Standards Organisation). We previously conducted a study of national mortality data identifying point suicide clusters in Wales over a 10year period, 2000–2009 (Jones et al., 2013). There was statistical evidence of a single, cluster of 10 deaths in young people aged 15– 34 years (primary cluster) in Bridgend and the surrounding area for the period December 27, 2007, to February 19, 2008. This cluster was smaller, shorter in duration, and predominantly later than the phenomenon that was widely reported in national and international print media in early 2008. No other statistical clusters were identified in Wales over the study period and there was no evidence of previous clusters in the Bridgend area indicating specific community vulnerability. Five other clusters of possible suicides were identified in the temporospatial analysis in 15–34-year-olds across Wales during 2000– 2009, but these were not significant at the .05 level (secondary clusters). Of these, two occurred at roughly similar time periods to the primary cluster, December 27, 2007, to March 17, 2008, and included cases from the primary cluster but related to a larger geographical area in the same locality (Jones et al., 2013). Combining the primary and secondary cases in this area for the period December 27, 2007, to March 17, 2008, extends the size of the cluster to a possible 18 cases. Given the high profile of reporting of deaths in the Bridgend locality, the objectives of our current study were to examine the quantity and quality of newspaper articles relating to the cluster, in particular in relation to guidelines on reporting. Crisis (2017), 38(1), 17–25

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

Method Search Strategy Searches were conducted of two specialist news reports databases (Nexis and Newsbank), the Internet search engine Google, and individual newspaper websites (including News UK), using the terms suicide and Bridgend. Original newspaper articles were retrieved either via the Internet and individual newspaper subscription, or from local or British library archives. On-line versions can be updated after the original date of print publication or the presentation may change, particularly in relation to the photographs that were originally published in paper editions (Luce, 2010); when this was evident, the original articles were obtained from library archives. For newspapers accessed via library archives, additional hand searching was undertaken. This search included reports during a period of 6 months prior to the identified commencement of the primary cluster (first death) and 6 months following the cessation of the secondary clusters (last death), that is, June 26, 2007, to September, 16, 2008. The newspapers included in the study were as follows: −− Local/regional: South Wales Echo; Western Mail; Wales on Sunday −− National broadsheets: The Times; The Guardian; The Independent; The Telegraph; The Observer; The Times on Sunday; The Independent on Sunday; The Sunday Telegraph −− National tabloids: The Daily Mail; The Mirror; The Sun; The Mail on Sunday; The Sunday Mirror; Daily Express; The Daily Star; People; News of the World; The Sunday Express; The Star on Sunday All newspaper articles and editorials using the terms suicide and Bridgend were included but letters were excluded.

Quality Assessment An instrument designed to assess the quality of reporting of newspaper articles (PRINTQUAL) was used in this study. The development and characteristics of this rating scale have been reported in detail previously (John et al., 2014). In summary, PRINTQUAL was based on the UK Samaritans guidelines for reporting suicide and self-harm (Samaritans, 2008) and on published evidence concerning the relationship between suicide and media reporting (Hagihara, Tarumi, & Abe, 2007; Hamilton, Metcalfe, & Gunnell, 2011; Hawton & Williams, 2002; Niederkrotenthaler et al., 2010; Stack, 2000, 2003; Thom, McKenna, Edwards, O’Brien, & Nakarada-Kordic, 2012). The instrument was initially developed by the research team, and © 2016 Hogrefe Publishing

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

then further advice on the items to be included was sought from an international group of experts in the area of suicide, suicide reporting, and suicide clusters. PRINTQUAL comprises two subscales of quality of print media reporting on suicide: negative/poor, and positive/good. A list of each characteristic is coded (1) or (0) depending on its presence or absence. The maximum possible poor-quality individual item count is 19 and good-quality item count is 4. Cronbach’s α (internal consistency) for the poor quality subscale was 0.96 (excellent) and for the good quality subscale, 0.69 (acceptable; John et al., 2014). Once training of investigators had taken place, the agreement between coders on a subset of 30 articles was acceptably high (Cohen’s κ ≥ .75) for most individual items (John et al., 2014) except the use of recommended phrases or phrases to be avoided in reporting, identification of suicide hotspots, and the use of explicit details of method used; although agreement for these items was still acceptable (κ ≥ .60). The use of sensational language has been identified in other studies as difficult to define (Hamilton et al., 2011) which may account for the latter score. Although a weighted scoring system has been developed for PRINTQUAL (John et al., 2014), for the purposes of this study only the frequencies of nonweighted items are described.

Data Extraction A data extraction sheet was specifically designed for the study, which, in addition to general descriptive items regarding the articles, included all the items from PRINTQUAL. Two investigators (A. M., P. J.) received training in the use of PRINTQUAL and further instruction on the rating of individual items.

Data Analysis Newspaper Reporting We calculated the daily frequency of published newspaper reports and plotted them against the incidence of possible suicide deaths for the primary and secondary clusters. We also calculated for each day of the study the number of newspaper reports in the preceding 2 and 7 days, highlighting the days when possible suicide deaths for the primary and secondary clusters occurred, since these would be the articles people would be exposed to immediately prior to their deaths and, according to the literature, are considered to be when the impact of reporting is most evident (Bollen & Phillips, 1982). Items indicating sensationalist reporting or those directly and unequivocally contravening guidelines were specifically recorded. These were identified by the fol© 2016 Hogrefe Publishing

lowing characteristics: main headline front page with the method specifically mentioned or the word suicide; explicit details of method used (i.e., more detail than just stating the method) within the report; mention of a suicide hotspot; repeated reporting of earlier suicides; technical details of an unusual method that in the context of this study was any method other than hanging; and the number of photographs included in the article. Some of these items of sensationalist reporting were specified within PRINTQUAL. Other items recorded in our study were additional to those used in PRINTQUAL, for example, the number of photographs in an article was removed from the final PRINTQUAL score because of inclusion in the scale of other items relating to photographs and, although the main headline and front page are items in PRINTQUAL, they are independent of each other. We calculated counts of PRINTQUAL items by newspaper type and in total. We calculated Pearson’s chi-square to explore the association between newspaper type and total poor-quality and good-quality items.

Results Reporting Quantity and Type We identified 577 newspaper reports concerning suicide in Bridgend during the study period June 26, 2007, to September 16, 2008. A total of 347 (60.1%) articles were in national newspapers (133 in broadsheets, 214 in tabloids), and 230 (39.9%) were in regional newspapers. Figure 1 displays the number of newspaper articles per day for the duration of the primary and secondary clusters (December 27, 2007, to March 16, 2008). The days when deaths occurred, of those aged 15–34 years in the primary or secondary clusters, are highlighted with circles. Two deaths occurred on certain days resulting in a total of 15 circles indicating 18 deaths. Days are numbered from the start of the primary cluster (December 27, 2007) to protect the identity of individuals. Other dates included relate to actual newspaper reports or press activity. The first report in a regional newspaper describing a cluster of suicides in Bridgend appeared on the January 17, 2008. The first report in a national newspaper describing a cluster appeared on the January 23 following a story released by the Wales News Service (a wire service similar to Reuters) on January 21 and 22. There was a large increase in the volume of reporting from January 23. Figure 2 displays the number of newspaper articles per day in the previous 2 and 7 days for the duration of the primary cluster and secondary clusters, with days when the deaths of those aged 15–34 years occurred highlighted. Crisis (2017), 38(1), 17–25

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

Figure 1. Number of articles per day from the start of the primary cluster with days of deaths highlighted.

Figure 2. Number of articles in the previous 2 and 7 days with days of deaths highlighted.

Sensationalist Reporting Table 1 shows the frequency of each poor-quality and positive-quality items in PRINTQUAL overall. A total of 92 (15.9%) articles mentioned the method in the headline on any page (12 were on the front page) and 223 (38.6%) mentioned suicide in the headline (six were on the front page). Just over a quarter of articles (155; 26.9%) referred to a suicide hotspot, including the use of terms such as suicide town. Nearly two thirds (350; 60.7%) of reports included photographs. The number of photographs per report ranged from one to 24; 49 (8.5% of all articles, 14.0% of articles carrying photographs) had more than Crisis (2017), 38(1), 17â&#x20AC;&#x201C;25

four photographs. Notably, 247 (42.8% of all articles, 70.5% of articles carrying photographs) included a photograph of the deceased. Thirteen (2.3% of all articles, 3.7% of articles carrying photographs) included photographs of the actual scene of the death (e.g., inside house, tree), 35 (6.1% of articles, 10% of articles carrying photographs) of the location (e.g., outside house, street, wood, locality), and 210 (36.4% of all articles, 60.0% of articles carrying photographs) republished photographs from earlier cases. The republishing of photographs of earlier deaths often accounted for the high number of photographs associated with an article.

ÂŠ 2016 Hogrefe Publishing

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

Table 1. Frequency of each item in PRINTQUAL in total and by newspaper type Frequency present (%) All, n = 577

Broadsheet national (%) n = 133

Tabloid national (%) n = 214

Regional (%) n = 230

10 (4.7)

23 (10.0)

Poor-quality item Is the article on the front page?

42 (7.3)

Is it the main headline on the front page?

26 (4.5)

4 (3.0)

5 (2.3)

17 (7.4)

Is the method mentioned in the headline?

92 (15.9)

16 (12.0)

36 (16.8)

40 (17.4)

195 (33.8)

47 (35.3)

74 (34.6)

74 (32.2)

32 (5.5)

8 (6.0)

2 (0.9)

22 (9.6)

Does the article cover over 50% of the page? Is it on Page 3?

9 (6.8)

Does the article use phrases to be avoided as listed in Samaritans guidelines?

250 (43.3)

68 (51.1)

117 (54.7)

65 (28.3)

Are explicit or technical details of the method described?

114 (19.8)

36 (27.1)

62 (29.0)

16 (7.0)

Are technical details of an unusual method for the locality described?

5 (0.9)

5 (3.8)

0 (0.0)

Are the contents of a suicide note described?

34 (5.9)

10 (7.5)

13 (6.1)

11 (4.8)

Does it mention or refer to a suicide hotspot?

155 (26.9)

41 (30.8)

102 (47.7)

12 (5.2)

9 (1.6)

2 (1.5)

7 (3.3)

0 (0.0)

Does it report positive outcomes from the death? Is the cause of the suicide attributed to a single factor?

127 (22)

30 (22.6)

63 (29.4)

34 (14.8)

Is there repeated reporting of earlier suicides in the article?

273 (47.3)

70 (52.6)

130 (60.7)

73 (31.7)

Does the article report the person knew previous suicides or that the timing implies a link?

240 (41.6)

61 (45.9)

106 (49.5)

73 (31.7)

Does the article highlight community expressions of grief?

201 (34.8)

50 (37.6)

86 (40.2)

65 (28.3)

Does the article include interviews with the bereaved?

214 (37.1)

45 (33.8)

95 (44.4)

74 (32.2)

Does the article include photographs of the scene, location, or method? Does the article include a photograph of the deceased? Does the article mention a celebrity suicide?

46 (8.0)

18 (13.5)

17 (7.9)

11 (4.8)

247 (42.8)

49 (36.8)

117 (54.7)

81 (35.2)

9 (1.6)

5 (3.8)

4 (1.9)

0 (0.0)

Good-quality item Does the article include recommended language as based on guidelines?

210 (36.4)

60 (45.1)

44 (20.6)

106 (46.1)

Does the article describe complex or multifactorial causes of the death?

60 (10.4)

27 (20.3)

16 (7.5)

17 (7.4)

Does it include sources of information or advice?

75 (13.2)

19 (14.3)

26 (12.1)

30 (13.0)

109 (18.9)

36 (27.1)

20 (9.3)

53 (23.0)

Does it take the opportunity to educate the reader?

Reporting Quality The range of poor-quality items per article was 0–13, with only four (0.7%) articles having no poor-quality items. The median was 3 and interquartile range (IQR) was 2–6. The range of good quality items was 0–4 (Mdn = 1, IQR = 0–1), with almost half of the articles (281, 48.7%) having no good-quality items and only eight (1.4%) having all four. Only 76 (13.2%) articles included sources of information or advice, 53 signposted to the Samaritans, one to the National Health Service, and a further 22 (3.8%) to other sources of advice. A total of 347 (60.1%) articles were in national newspapers (133 in broadsheets, 214 in tabloids) and 230 (39.9%) were in regional newspapers. All broadsheet articles combined contained 574 poor-quality items out of a possible total of 2,527 (22.7%), tabloids 1,046 of 4,066 (25.7%), and regional newspapers 691 of 4,370 (15.8%). There © 2016 Hogrefe Publishing

was a small effect but significant association between the type of newspaper and poor-quality reporting items, χ(2) = 130, p = .0001, Cramer’s V = .10. All broadsheet articles combined contained 142 good-quality items out of a possible total of 532 (26.7%), tabloids 46 of 856 (5.4%), and regional newspapers 206 of 920 (20.0%). There was a moderate effect but significant association between the type of newspaper and poor-quality reporting items, χ(2) = 136, p = .0001, Cramer’s V = .24.

Discussion We examined the quality of newspaper reporting for the duration of a community point cluster of suicides. Importantly, we found concerning evidence of poor-quality reporting during a probable suicide cluster; this is the very time when Crisis (2017), 38(1), 17–25

good-quality reporting could be considered most essential in the public health response to a cluster. In particular, nearly half of reports referred to earlier suicides, 43% displayed a photograph of the deceased, and 44% used phrases that in the light of research evidence and suicide prevention guidelines should be avoided. Conversely, only 13% included sources of information or advice. There was a high level of sensationalist reporting, indicated by approximately one in seven articles mentioning the method in the headline and extensive republishing of photos from previous suicide deaths. Many newspapers reported on a large number of probable suicides in young people that occurred in Bridgend in 2007 although there is no statistical evidence of an excess of deaths during that time. The first regional newspaper report of an excess of suicides in Bridgend on January 17, 2008, reported on deaths unrelated to the actual identified primary or secondary cluster deaths defined in our statistical analysis (Jones et al., 2013) and that had occurred several months previously. This article coincided with the fourth primary cluster death (sixth death of primary and secondary clusters combined). The initial four deaths of the identified primary cluster were no more than would normally be expected at this stage, being in keeping with the number of suicide deaths in this age group of 15â&#x20AC;&#x201C;34-yearolds for this time period in previous years (Jones et al., 2013). The subsequent six primary cluster deaths and three secondary cluster deaths were accompanied by a large increase in the volume of reporting from January 23, 2008. Only three further deaths (from the secondary cluster) followed a second and larger peak in reporting on February 20, 2008. This might argue against any possible causal link. However, this second peak in reporting could have been stimulated by an international press conference held on February 19, 2008, to highlight the potentially damaging role of the media, and the thematic content of reports in this second peak may therefore have been less likely to maintain any contagion process. We plan, in the future, to conduct a more in-depth thematic analysis of the two peaks of reporting to explore whether any differences in content had an impact on any contagion process There was no clear relationship between the frequency of newspaper reports and deaths when examining 2- and 7-day rolling periods preceding each suicide. However, in interpreting the influence of the volume of reporting, it is difficult to account for other factors. For example, intervention and support provided by health and voluntary agencies during the later period of the cluster may have reduced the risk of subsequent deaths. Other studies have found two waves of reporting following suicide deaths (Balazs et al., 2013).

Crisis (2017), 38(1), 17â&#x20AC;&#x201C;25

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

Strengths and Limitations The focus of this study was on newspaper articles only, rather than other types of reporting, such as radio, television, or Internet. This was for a number of reasons. Firstly, the initial reports of a possible suicide cluster among young people in Bridgend occurred in local newspapers, and the main focus of subsequent reporting was in the print media. Secondly, the evidence for possible contagion from media reporting still remains strongest for print media (Hawton & Williams, 2005; Stack, 2005). Studies suggest that television is less likely to produce a copycat effect than newspaper reports are (Hawton & Williams, 2005; Stack, 2005). This may be because of stricter regulation practices or simply because it is more difficult to study as an exposure. Studies comparing the effects of modes of reporting suggest that the impact of Internet reporting is of lower magnitude than that of print media (Hagihara et al., 2007). However, with the huge increase in use of new media since the time of the Bridgend deaths, Internet and social media influences might now be more relevant to clustering and contagion of suicide (Daine et al., 2013). There was no evidence of social media being a significant factor in the deaths associated with the Bridgend cluster following a police investigation (P. Hurley, personal communication, May 21, 2015). Robertson, Skegg, Poore, Williams, and Taylor (2012) have recently described a point cluster in adolescents when SMS text messaging and online social networking were possibly an important mode of contagion. The Internet may also be a mechanism for cyber-bullying and encourage self-harm behavior, although its influence may also be positive by encouraging positive coping and help-seeking (Daine et al., 2013). Previous studies have been limited by lacking an appropriate quality measure. We have attempted to minimize this by using an assessment instrument (PRINTQUAL), which was developed on the basis of widely quoted guidelines and other evidence together with a consensus weighting system that was devised in collaboration with experts in the field of suicidology.

Meaning and Implications One approach to support the media in responsible reporting of suicide has been to produce guidelines. Such guidance is an integral part of suicide prevention strategies around the world. Our findings give further emphasis to the importance of inclusion of particular items in press codes and recommendations on suicide reporting. Gould et al. (2014) found that front-page placement, detailed descriptions of the suicidal individual and act, and headlines containing the word suicide or a description of the method ÂŠ 2016 Hogrefe Publishing

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

used were frequently present in articles associated with suicide clusters. In our study of Bridgend suicide reporting, 10% of all regional papers had a report on the front page, 16% mentioned the method in the headline, and 20% included explicit details of the act. Repetitive suicide reporting and reference to suicide epidemics have also been found to influence suicide rates (Niederkrotenthaler et al., 2010). Over a quarter of articles concerning Bridgend deaths referred to a suicide hotspot, nearly half included reference to earlier suicides, over 40% made links between the suicides, and over a third reproduced photographs related to previous local deaths. There is a general consensus that information on help and support needs to be included in press codes and reporting recommendations (Maloney et al., 2013) – only 13% of articles we identified included such information. It is difficult to demonstrate whether media guidelines improve the quality of reporting (Bohanna & Wang, 2012; Hawton & Williams, 2002). In 2006 the UK Press Complaints Commission (PCC) added a clause to the Editors’ Code of Practice explicitly recommending that the media avoid excessively detailed reporting of suicide methods. We found little evidence of this advice being followed in 2008, with a wide interpretation of the term excessively detailed as stipulated in the 2006 code. However, reporting guidelines in Australia have been generally well received, although there are difficulties in interpretation of recommendations that require subjective judgements (Machlin et al., 2012). Likewise, in New Zealand, Thom et al. (2012) found that adherence to Ministry of Health guidelines has overall resulted in good-quality reporting, although, as we have found, there was a paucity of articles referencing sources of help or people managing to overcome their suicidal ideation. Suicides by younger people (as in Bridgend), involving violent methods, or occurring in public places or medical and residential facilities are particularly likely to attract the media’s attention (Machlin, Pirkis, & Spittal, 2013), and so it is particularly important that responsible reporting occurs in these circumstances. Other countries have examined newspaper reporting of suicide by newspaper type (Cheng & Yip, 2012) but there are no studies examining this based in the UK. Our study suggests that poor-quality reporting in relation to suicide may be more of an issue for national newspapers than regional ones, which has implications for ensuring they are engaged in initiatives to improve the adoption of guidelines. This was particularly in relation to phrases to be avoided, technical details, hotspots, repeated reporting of earlier suicides, and the use of photographs. The UK 2009 edition of the PCC Editors’ Codebook highlights the distress that can be caused by insensitive and inappropriate graphic illustrations accompanying media reports of suicide and the re-publication of photographs of people who © 2016 Hogrefe Publishing

have died by suicide when reporting other suicide deaths in the same area. The results of our study highlight the prudence of these recommendations. They also commend the inclusion of details of local support organizations and helplines with any coverage of suicide deaths. In future, researchers should further explore the content of newspaper reporting using qualitative methods. This could include interviewing those with near-fatal self-harm at the time of an apparent suicide cluster to explore their understanding of their own behavior, and the role of Internet reporting and social media. Finally, it is worth asking (and investigating) whether the national media reporting of the apparent cluster in Bridgend had an impact on UK national suicide rates, as arguably the volume of reporting of the method could have contributed to a mass cluster. No evidence of a mass cluster was found at a regional (South Wales) or Wales geographical level (Jones et al., 2013).

Conclusion We have described the quantity and quality of newspaper reporting during a suicide point cluster of young people in South Wales. There were high levels of sensationalist reporting. This gives credence to suggestions that increased awareness, collaboration, training, and ownership by journalists of guidelines for reporting of suicide are required to improve the adoption of guidelines and improve the quality of reporting (Bohanna & Wang, 2012). Recommendations on the reporting of suicide should be balanced with an awareness of tackling stigma in relation to suicide and selfharm, signposting sources of help, encouraging help-seeking behavior, and educating the public both to understand the complexity of reasons why someone may take their own life and in how to respond to people in crisis. Acknowledgments This study was funded by a grant from the National Institute for Social Care and Health Research, Welsh Government. We thank Dr. Babar Kamran for his help in obtaining original newspaper articles and Detective Superintendent Paul Hurley, Specialist Crime Investigations, for confirming via e-mail that social media were not a significant factor in the cluster of suicides in the Bridgend area during the relevant period. There are no conflicts of interest. D. G. and K. H. are National Institute for Health Research (England) Senior Investigators.

Crisis (2017), 38(1), 17–25

References Balazs, J., Kereszteny, A., Pelbat, G., Sinka, L., Szilvas, F., & Torza, T. (2013). Online media report on a Hungarian Double suicide case: Comparison of consecutively published articles. Psychiatria Danubina, 25(3), 248–254. Bohanna, I., & Wang, X. (2012). Media guidelines for the responsible reporting of suicide: A review of effectiveness. Crisis, 33, 190–198. Bollen, K., & Phillips, D. (1982). Imitative suicides: A national study of the effects of television news stories. American Sociological Review, 47, 802–809. Cheng, A., Hawton, K., Chen, T., Yen, A., Chang, J., Chong, M., … Chen, L. (2007). The influence of media reporting of a celebrity suicide on suicidal behaviour in patients with history of depressive disorder. Journal of Affective Disorders, 103, 69–75. Cheng, Q., & Yip, P. S. F. (2012). Suicide news reporting accuracy and stereotyping in Hong Kong. Journal of Affective Disorders, 141(2–3), 270–275. doi:10.1016/j.jad.2012.03.036 Daine, K., Hawton, K., Singaravelu, V., Stewart, A., Simkin, S., & Montgomery, P. (2013). The power of the Web: A systematic review of the studies of the influence of the internet on selfharm and suicide in young people. PLOs oNE, 8(10), e77555. doi:10.1371/journal.pone.0077555 Fu, K., & Yip, P. (2007). Long-term impact of celebrity suicide on suicidal ideation: Results from a population based study. Journal of Epidemiology and Community Health, 61, 540–546. Gould, M., Kleiman, M., Lake, A., Forman, J., & Midle, J. (2014). Newspaper coverage of suicide and initiation of suicide clusters in teenagers in the USA, 1988-96: A retrospective, population-based, case-control study. The Lancet Psychiatry. doi:10.1016/S2215-0366(14)70225-1 Hagihara, A., Tarumi, K., & Abe, T. (2007). Media suicide reports, internet use and the occurrence of suicides between 1987 and 2005 in Japan. BMC Public Health, 7, 321. doi:10.1186/14712458-7-321 Hamilton, S., Metcalfe, C., & Gunnell, D. (2011). Media reporting and suicide: A time-series study of suicide from Clifton suspension bridge, UK 1974-2007. Journal of Public Health, 33, 511–517. Haw, C., Hawton, K., Niedzwiedz, C., & Platt, S. (2013). Suicide clusters: A review of risk factors and mechanisms. Suicide and Life-Threatening Behavior, 43, 97–108. Hawton, K., Simkin, S., Deeks, J., O’Connor, S., Keen, A., Altman, D., … Bulstrode, C. (1999). Effects of a drug overdose in a television drama on presentations to hospital for self poisoning: Time series and questionnaire study. BMJ, 318, 972–977. Hawton, K., & Williams, K. (2002). Influences of the media on suicide. BMJ, 325, 1374–1375. Hawton, K., & Williams, K. (2005). Media influences on suicidal behaviour: Evidence and prevention. In K. Hawton (Ed.), Prevention and treatment of suicidal behaviour (pp. 340–356). Oxford, UK: Oxford University Press. John, A., Hawton, K., Lloyd, K., Luce, A., Platt, S., Scourfield, J., … Dennis, M. (2014). PRINTQUAL: The development of a measure for assessing the quality of newspaper reporting of suicide. Crisis. doi:10.1027/0227-5910/a000276 Jones, P., Gunnell, D., Platt, S., Scourfield, J., Lloyd, K., Huxley, P., … Dennis, M. (2013). Identifying probable suicide clusters in Wales using national mortality data. PLoS ONE, 8(8), e71713. doi:10.1371/journal.pone.0071713 Larkin, G., & Beautrais, A. (2012). Geospatial mapping of suicide clusters. Te Pou o Te Whakaaro Nui. Retrieved from http:// www.tepou.co.nz/library/tepou/geospatial-mapping-of-suicide-clusters Luce, A. (2010). The infantilisation and stigmatization of suicide: A multi-modal analysis of British press reporting of the Bridgend Crisis (2017), 38(1), 17–25

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

suicides (Doctoral dissertation). Bournemouth University, Cardiff, UK. Machlin, A., Pirkis, J., & Spittal, M. (2013). Which suicides are reported in the media – and what makes them ‘newsworthy’. Crisis, 34, 305–313. Machlin, A., Skehan, J., Sweet, M., Wake, A., Fletcher, J., Spittal, M., & Pirkis, J. (2012). Reporting suicide: Interpreting media guidelines. Australian Journalism Review, 34, 45–56. Maloney, J., Pfuhlmann, B., Arensman, E., Coffey, C., Gusmao, R., Postuvan, V., … Sisask, M. (2013). Media recommendations on reporting suicidal behaviour and suggestions for optimisation. Acta Psychiatrica Scandinavica, 128, 314–315. Niederkrotenthaler, T., Fu, K., Yip, P., Fong, D., Stack, S., Cheng, O., & Pirkis, J. (2012). Changes in suicide rates following media reports on celebrity suicide: A meta-analysis. Journal of Epidemiology and Community Health, 66, 1037–1042. Niederkrotenthaler, T., Voracek, M., Herbeth, A., Till, B., Strauss, M., Etzerdorfer, E., … Sonneck, G. (2010). Role of media reports in completed and prevented suicide: Werther v. Papageno effects. British Journal of Psychiatry, 197(3), 234–243. doi:10.1192/bjp. bp.109.074633 Niedzwiedz, C., Haw, C., Hawton, K., & Platt, P. (2014). The definition and epidemiology of clusters of suicidal behaviour: A systematic review. Suicide and Life-Threatening Behavior. doi:10.1111/ sltb.12091 Pirkis, J., & Blood, R. (2001). Suicide and the media. Part 1: Reportage in non-fictional media. Crisis, 22, 146–154. Pirkis, J., & Nordentoft, M. (2011). Media influences on suicide and attempted suicide. In R. O’Conner, S. Platt, & J. Gordon (Eds.), International handbook of suicide prevention research, policy and practice (pp. 531–544). Chichester, UK: John Wiley & Sons. Robertson, L., Skegg, K., Poore, M., Williams, S., & Taylor, B. (2012). An adolescent suicide cluster and the possible role of electronic communication technology. Crisis, 33, 239–245. Samaritans. (2008). Media guidelines for reporting suicide and selfharm. Retrieved from http://www.samaritans.org/media-centre/media-guidelines-reporting-suicide Sisask, M., & Varnik, A. (2012). Media roles in suicide prevention: A systemtic review. International Journal of Environmental Research and Public Health, 9, 123–128. Stack, S. (2000). Media impacts on suicide: A quantitative review of 293 findings. Social Science Quarterly, 81, 957-971. Stack, S. (2003). Media coverage as a risk factor in suicide. Journal of Epidemiology and Community Health, 57, 238–240. Stack, S. (2005). Suicide in the media: A quantitative review of studies based on non-fictional stories. Suicide and Life-Threatening Behavior, 35, 121–133. Thom, K., McKenna, B., Edwards, G., O’Brien, A., & Nakarada-Kordic, I. (2012). Reporting of suicide by the New Zealand media. Crisis, 33, 199–207. World Health Organization. (2005). Preventing suicide. A resource for media professionals. Retrieved from http://www.who.int/ mental_health/prevention/suicide/resource_media.pdf

Received July 6, 2015 Revision received February 28, 2016 Accepted March 2, 2016 Published online July 22, 2016 About the authors Ann John, MBBS, MD, is Associate Professor of Public Mental Health at Swansea University Medical School, UK. She is a public health physician with a research focus on suicidal behaviors and the translation of evidence into policy and practice. © 2016 Hogrefe Publishing

A. John et al.: Newspaper Reporting on a Cluster of Suicides in the UK

Keith Hawton, FMedSci, is Professor of Psychiatry and Director of the Centre for Suicide Research at the Centre for Suicide Research, University of Oxford, UK. His team has been involved in a wide range of studies related to epidemiology, causes, prevention, and treatment of suicidal behaviors, including media influences. David Gunnell, MB, ChB, DSc, is Professor of Epidemiology at the University of Bristol, UK. He is a public health physician and epidemiologist with a longstanding research interest in the etiology and prevention of suicide and in improving population mental health. Keith Lloyd, MRCPsych, MD, is Dean and Head of Swansea University Medical School, UK, and Professor of Psychiatry. His research interests are psychiatric epidemiology and the management of mental disorders in primary care and community settings especially suicide and selfharm. Jonathan Scourfield, PhD, is Professor of Social Work at Cardiff University, UK, and Deputy Head of the School of Social Sciences. He has conducted a range of studies on the social context of suicide and selfharm and is the co-author of Understanding Suicide: A Sociological Autopsy (Palgrave Macmillan, 2011). Phillip Jones, PhD, was formerly a researcher in the Population Psychiatry, Suicide and Informatics group at Swansea University Medical School, UK. Amanda Marchant, MSc, is a researcher in the Population Psychiatry, Suicide and Informatics group at Swansea University Medical School, UK.

ÂŠ 2016 Hogrefe Publishing

Professor Stephen Platt, PhD, was Professor of Health Policy Research in the Centre for Population Health Sciences at the University of Edinburgh, UK. For over 30 years he has maintained a research interest in epidemiological, social, and cultural aspects of suicidal behavior. Ann Luce, PhD, is a journalist-turned-academic and is the author of a forthcoming book on this topic. She is also the author of several book chapters and journal articles about disability in the media and representations of midwifery and the media. Sian Price, MPH, is a public health specialist and head of the Public Health Wales Observatory Evidence Service, UK. Mick Dennis, MRCPsych, is Professor of Psychiatry for Older People at Swansea University Medical School, UK, and was previously a consultant liaison psychiatrist. He has been involved in many areas of mental health research, particularly the epidemiology of suicide and selfharm, and social and psychological factors in suicidal behaviors.

Ann John Swansea University Medical School 3rd Floor, Institute of Life Sciences 2 Swansea University, Singleton Park Swansea, SA2 8PP UK Tel. +44 (0) 1792 602-568 Fax +44 (0) 1792 513-430 E-mail a.john@swansea.ac.uk

Crisis (2017), 38(1), 17â&#x20AC;&#x201C;25

Research Trends

Characteristics and Proximal Outcomes of Calls Made to Suicide Crisis Hotlines in California Variability Across Centers Rajeev Ramchand1, Lisa Jaycox1, Pat Ebener2, Mary Lou Gilbert2, Dionne Barnes-Proby2, and Prodyumna Goutam2 RAND Corporation, Arlington, VA, USA RAND Corporation, Santa Monica, CA, USA

1 2

Abstract. Background: Suicide hotlines are commonly used to prevent suicides, although centers vary with respect to their management and operations. Aims: To describe variability across suicide prevention hotlines. Method: Live monitoring of 241 calls was conducted at 10 suicide prevention hotlines in California. Results: Call centers are similar with respect to caller characteristics and the concerns callers raise during their calls. The proportion of callers at risk for suicide varied from 3 to 57%. Compliance with asking about current suicide risk, past ideation, and past attempts also ranged considerably. Callers to centers that were part of the National Suicide Prevention Lifeline (NSPL) were more likely to experience reduced distress than callers to centers that were not part of the NSPL. Conclusion: Because callers do not generally choose the center or responder that will take their call, it is critical to promote quality across call centers and minimize the variability that currently exists. Accrediting bodies, funders, and crisis centers should require that centers continuously monitor calls to ensure and improve call quality. Keywords: suicide prevention, crisis hotlines, live monitoring, evaluation, California

Communities commonly use suicide prevention hotlines to identify persons at risk for suicide (e.g., screening callers), reduce immediate risk (e.g., encouraging callers to remove lethal means from the vicinity or dispatching emergency personnel), and provide referrals to behavioral health care (Acosta, Ramchand, Jaycox, Becker, & Eberhart, 2012; Gould, Munfakh, Kleinman, & Lake, 2012). When voters in California approved Proposition 63 (the Mental Health Services Act) and the Mental Health Oversight and Accountability Commission (MHOAC) selected preventing suicide as one of its statewide initiatives (Clark et al., 2013), 12 grant recipients were funded to create, expand (i.e., to new counties or populations), or enhance (i.e., by offering crisis chat and/or text services) the hotline services they provide. To date, evaluations of suicide prevention hotlines have examined both the process of calls and proximal outcomes. With respect to process, studies have evaluated call responders’ behaviors on the call such as their response styles (Gould, Cross, Pisani, Munfakh, & Kleinman, 2013; Mishara, Chagnon, & Daigle, 2007; Mishara et al., 2007; Mishara & Daigle, 1997) and their adherence to recomCrisis (2017), 38(1), 26–35 DOI: 10.1027/0227-5910/a000401

mendations that all callers be asked about current suicide ideation, past suicide ideation, and past suicide attempts (Joiner et al., 2007). These studies have shown that the majority of call responders are able to establish good contact and rapport with callers (Gould et al., 2013; Mishara et al., 2007; Mishara & Daigle, 1997). In addition, a recent evaluation found that over 90% of callers were asked about current suicidal thoughts; however, inquiries about past thoughts and past attempts was under 50% (Gould et al., 2013), a marginal improvement from 5 years earlier (Mishara et al., 2007). In terms of outcomes, callers generally improve over the course of the call as measured by improved mental state and reduced suicide ideation (Gould et al., 2013; Gould, Kalafat, Harrismunfakh, & Kleinman, 2007; Gould et al., 2012; Kalafat, Gould, Munfakh, & Kleinman, 2007; King, Nurcombe, Bickman, Hides, & Reid, 2003; Mishara & Daigle, 1997). Research on predictors of call outcomes have focused primarily on call responders’ styles and behaviors. Mishara et al. (1997) found that those call responders who exhibited “Rogerian” versus “directive” styles were more likely to have callers who exhibited improvements over © 2016 Hogrefe Publishing

R. Ramchand et al.: California Crisis Hotlines

the course of the call. Ten years later, Mishara et al. (2007) found that callers who exhibited positive changes during the call spoke with responders who exhibited empathy, respect, a supportive approach, who had good contact, and used collaborative problem solving , but that active listening was not related to call outcomes. Finally, Gould et al. (2013) randomized call responders (by call center) so that half received Applied Suicide Intervention Skills Training (ASIST), a training to recognize risk factors for suicide and respond to increase a person’s immediate safety and link them to appropriate resources (Cross et al., 2014; Smith, Silva, Covington, & Joiner, 2014). Callers who spoke with ASIST-trained responders were rated as becoming less depressed, less overwhelmed, less suicidal, and more hopeful by the end of the call (Gould et al., 2013). Although past evaluations have each included several crisis centers, all have examined calls in aggregate. However, call centers in the United States vary considerably. Some call centers are stand-alone operations while others are affiliated with a behavioral health-care center; they also vary in the ratio of volunteer to paid staff. Some offer crisis chat and/or text services in addition to telephone, some respond to multiple lines (e.g., 211, domestic violence, or gambling addiction), some specialize in working with particular populations (e.g., the Veterans Crisis Line for veterans; Knox, Kemp, McKeon, & Katz, 2012; or the Trevor Lifeline for lesbian, gay, bisexual, transgender, and questioning youth and young adults). And finally, only a subset of centers are part of the National Suicide Prevention Lifeline (NSPL). Presumably, call center characteristics could

Table 1. Call center characteristics among centers participating in live monitoring evaluation (N = 10) New call center created by CalMHSA funding

Affiliated with health center (vs. stand-alone operation)

Operation of other lines 211 line Other lines Accrediteda

1 8 10

Staffing Exclusively paid staff

Exclusively volunteer staff

Paid/volunteer mix

Offers crisis chat and/or text

Member of NSPL

Electronic data management

Note. CalMHSA = California Mental Health Services Authority. NSPL = National Suicide Prevention Lifeline. a Although all centers were accredited when we visited them, five achieved accreditation during their period of CalMHSA funding, and many used CalMHSA funds specifically to achieve accreditation.

be importantly related to outcomes. A recent study showed that paid responders at crisis centers produce superior outcomes to volunteers (Gould et al., 2016); another noted that “call centers differ greatly” with respect to call quality and suggested that future research understand such differences (Mishara et al., 2007). Indeed, similar research has examined structural characteristics of other care settings (e.g., hospitals and outpatient settings) and associated outcomes related to suicide risk (Ho, 2006; Lee & Lin, 2009; Lin, Lee, Kuo, & Chu, 2008; Qin & Nordentoft, 2005). In this study, we use our evaluation of 10 call centers in California to describe variability in call centers with respect to caller and call characteristics, call responder behaviors, and call outcomes. Because of an agreement between RAND (the evaluator), participating call centers, and the California Mental Health Services Authority (CalMHSA), we do not refer to centers by name, nor do we present any identifiable information about them.

Method Sampling All procedures were approved by RAND’s Institutional Review Board. Two raters visited 10 call centers during the spring and summer of 2014. Both raters visited the first two call centers to double-code calls to improve reliability (inter-rater reliability is presented in Jaycox, Ramchand, Ebener, Barnes-Proby, & Gilbert, 2015); the remaining eight centers were visited by only one monitor and all call responders knew that their calls could be monitored. Call monitors listened to eligible inbound calls during several shifts over 2–4 days, corresponding with days of the week with highest call volume, with a goal of at least 20 calls per center. All call responders working the shifts in which monitoring took place consented to be monitored. Call responders informed callers that the call could be monitored at the outset of the call if crisis lines did not already have such an announcement in place. Characteristics of the 10 call centers varied, although all were accredited by the American Association for Suicidology (see Table 1). One was created with Proposition 63 funds, the others were in existence prior to receiving additional funding. Other ways they varied were in their affiliation (i.e., two were stand-alone hotlines, eight were affiliated with a health provider), in whether they received other hotline (N = 8) or 211 calls (N = 1), relied exclusively on volunteers (N = 3) or a mixture of volunteers and paid call responders (N = 7), offered and responded to crisis chats or texts (N = 4), and were part of the NSPL (N = 6).

Crisis (2017), 38(1), 26–35

R. Ramchand et al.: California Crisis Hotlines

Measures The monitoring protocol with full supporting documentation (Jaycox et al., 2015) is available for free download at: http://www.rand.org/pubs/tools/TL150.html. It was developed from existing protocols (Gould et al., 2013; Gould et al., 2012) and from the Lifeline QI Monitoring Tool.1 Briefly, monitors recorded: Call characteristics measured include callers’ sex, start and end time, whether the caller was put on hold, type of call (i.e., repeat caller), and reasons for ineligibility for monitoring (i.e., minor under 18 years old, counselor not participating in evaluation, non-English-speaking caller, obscene/prank caller). Call content included the types of issues and problems raised during the call listed in Table 2 regardless of whether the caller mentioned them spontaneously or in response to call responders’ questions. Suicide risk assessment measured adherence to NSPL guidelines that telephone responders ask, unless offered by callers spontaneously: (1) Are you currently thinking of suicide? (current ideation); (2) Have you thought about suicide in the last 2 months? (recent ideation); and (3) Have you ever attempted to kill yourself? (attempt). It is recommended that all three questions be asked of all callers (Joiner et al., 2007). Call challenges measured the degree to which the situation and the caller were challenging, on a scale from 1 (not at all challenging) to 5 (very challenging). Telephone counselor response measured the types of actions and behaviors that call responders engaged in, adapted from other protocols (Gould et al., 2013). Changes during call were assessed by judging caller distress at the beginning of the call and at the end of each call, similar to prior studies (Mishara et al., 2007) but simplified into a single global 4-point distress item ranging from not at all distressed to extremely distressed. Overall ratings of the call were used to judge the overall quality of the call and how the call ended. For contact with callers, there were three ratings: “good contact,” “good with some weaknesses,” and “did not establish good contact or important weaknesses.” Problem solving was not always needed; when it was, calls were coded as using a “collaborative approach” or a “collaborative approach with some weaknesses.” Referrals were also not always needed; when they were, calls were coded as “referrals provided” or “referrals provided with some weaknesses or incom-

pletely.” Finally, monitors rated, on a scale from 1 to 5, how satisfied the caller appeared to be at the end of the call.

Data Analysis Descriptive statistics are presented across all domains, including the overall mean or proportion of calls, as well as the range across centers. Comparisons between specific call centers and averages for all sites combined were conducted using regression modeling; however, because of our agreement with call centers, we only present the number of call centers that differed from the average when presenting our results. A significant difference was defined as a p value less than .05. We assessed the relationship between both center characteristics and call/caller characteristics with changes in caller distress and call satisfaction. Call/caller variables included in the regression modeling were call content, call duration, challenging caller, and challenging situation. Center-level variables were whether the call center took crisis chats/texts in addition to calls (although we only monitored calls), whether it was a member of the NSPL, and whether it was staffed exclusively by volunteers. We estimated regression models with robust variance estimators in which both call/caller variables and center-level variables were entered into the regression models simultaneously. The relatively low number of sites (N = 10) and observations (N = 241) prohibited us from entering center-level fixed effects to guard against omitted variables, but our inclusion of three center-level covariates should account for cross-site heterogeneity.

Results Calls were screened (N = 273) and 241 were monitored, excluding calls in which callers lacked capacity to give consent, calls conducted in a language other than English, callers who were less than 18 years of age, obscene/prank callers, or, in a very few cases, those who objected to having their call monitored. On average, we evaluated 24 calls per center (range = 14–33), with only one center having fewer than 20 calls monitored.

1 The Lifeline QI Monitoring Tool is designed for crisis hotlines that are part of the NSPL to “facilitate supervision of counselors and guide the appraisal of their behaviors during calls.” It is available to members of the NSPL and was provided to us when we were developing our quality monitoring form. Crisis (2017), 38(1), 26–35

R. Ramchand et al.: California Crisis Hotlines

Call Characteristics

Call Content

As shown in Table 2, 35% of calls (range = 0–72%) were from repeat callers, although the proportion was significantly greater at two centers. Female callers accounted for an average of 57% of callers (range = 30–77%). Five of the centers never put callers on hold; however, the average proportion of all calls put on hold was 7% (range = 0=26%), owing largely to two centers that had significantly greater number of calls put on hold. Calls tended to last 14 min. On average, 13% (range = 0–33%) of calls were from someone concerned about another individual (i.e., third parties).

Among non-third-party calls, mental health problems or substance use problems were exhibited by 51% of callers (range = 28–86%), and physical health challenges were exhibited by 25% of callers (5–50%; see Table 2). Interpersonal problems (referenced by 45% of callers, range = 24–69%) and work, housing, and financial problems (27%, range = 14–43%) were also commonly mentioned.

Table 2. Call and caller characteristics of calls made to California crisis hotlines (N = 241) M

Min.

Max.

# Significantly greater

# Significantly lower

24.1

Third-party calls

13%

33%

Repeat callers

35%

Female callers

57%

72%

30%

77%

7% 14

26%

Interpersonal problems Mental illness/substance abuse

45%

24%

69%

51%

28%

86%

Work, housing, and financial problems

27%

14%

43%

Physical illness

25%

50%

Suicide

26%

71%

Empathy

20%

40%

Trauma or loss

13%

38%

Service use

14%

Other (sexuality, military)

14%

21%

57%

12%

Extremely distressed

17%

Moderately distressed

30%

57%

A little distressed

41%

13%

64%

Not at all distressed

20%

54%

Calls/center Call characteristics

Calls put on hold Call duration (M, in min) a,b

Call content

Suicidal intent

Caller currently thinking of suicide Caller concerned about somebody else thinking about suicide a

Caller distress at beginning of call

Distribution differed from average at 6 of the 10 call centers

Note. Based on 210 calls (excludes third-party calls). Interpersonal problems include relationship problems (e.g., partner), family conflict problems (someone other than a partner), concern about a family member, concern about a friend, or interaction problems (e.g., with neighbors). Mental illness/ substance abuse includes depression/anxiety/PTSD/other mental illness or current alcohol or drug use problems. Work, housing, and financial problems include work problems, financial problems, fear of losing housing, or homelessness. Physical illness includes illness/injury/disability or chronic pain. Suicide includes suicidal thoughts or intentions, or concerns that the caller might have about such thoughts. Empathy includes wanting to talk, checking-in, or sad/stressed/lonely/having no companion. Trauma or loss includes exposure to violence/trauma or loss of a family member/friend. Service use includes calls about problems accessing or utilizing services such as being unable to reach a counselor or care provider. Other includes personal issues not described in the aforementioned categories, mostly problems related to one’s sexual orientation or veteran or military issues, including accessing help at the Department of Veterans Affairs.

Crisis (2017), 38(1), 26–35

R. Ramchand et al.: California Crisis Hotlines

Calls With Suicide Ideation Involving Self or Others

Challenging Situations and Callers

Calls with suicide intent were defined as involving either callers who were currently thinking about suicide themselves or callers who were concerned about suicide for someone else. Therefore, more callers brought up suicide during a call (i.e., 26%, see Table 2) than were currently thinking about taking their own lives (21%, range = 3–57%), with the proportion at one center significantly greater. Callers concerned about somebody else they thought may be currently suicidal were less common – occurring in 5% (range = 0–12%) of calls with no statistically significant differences across centers.

Suicide Risk Assessment Figure 1 presents adherence to NSPL-recommended suicide risk assessment standards. Asking about current ideation occurred frequently, but was quite variable (M = 69%, range = 13–100%). Other standards were less commonly met: only 25% (range = 4–77%) of callers were asked about recent ideation and only 21% (range = 0–60%) were asked about past attempts. However, two centers were significantly more likely to ask about recent ideation and three were significantly more likely to ask about past attempts. However, further investigation revealed differences by NSPL membership: In bivariate analyses comparing NSPL with non-NSPL crisis centers, NSPL centers were more likely to ask about current ideation (77% vs. 52%, OR =3.6, p < .01), recent ideation (31% vs. 16%, OR = 2.5, p = .02), and past attempts (27% vs. 10%, OR = 3.7, p < .01). Revised Figure 1

In general, situations were rated as more challenging (M = 2.5, range = 1.8–3.2) than callers (M = 1.7, range = 1.2–2.2). Monitors rated situations in two centers as more challenging, and in two centers, on average, as less challenging. Monitors rated callers as more challenging than average in one center, and rated callers as less challenging than average in another.

Caller Distress In all, 35% of callers were in extreme or moderate (mostly moderate) distress, although at three centers, over half of callers were rated as moderately or extremely distressed (Table 2). On the other hand, at two centers very few callers were extremely or moderately distressed.

Responses to Callers Each call was rated across 12 responder responses: seven positive response styles and five negative response styles. All response styles are presented in Table 3, each of which was rated on a 3-point scale ranging from 0 (not at all) to 2 (a lot). In general, positive behaviors occurred frequently, particularly allowing callers to talk about feelings/situations, connecting/establishing rapport with callers, being sensitive/receptive to callers, and being respectful. Two centers averaged lower than 1.5 on connecting/establishing rapport with callers, two averaged lower on being sensitive/receptive to callers problems, and one averaged lower on being respectful. Less frequently endorsed were reflecting back callers’ feelings, situations, and showing empathy/validating callers. Negative behaviors occurred very infrequently. Figure 1. Adherence to NSPL suicide risk assessment standards (N = 210). Caller offered = caller offered, among all calls. Responder asked = responder asked, among all calls in which caller did not offer this information. Error bars represent minimum and maximum proportions across centers.

100% 90% 80% 68%

70% 60% 50% 40%

25%

30% 20%

12%

10%

21% 5%

0% Current Ideation

Recent Ideation Caller Offered

Crisis (2017), 38(1), 26–35

Past Attempts

R. Ramchand et al.: California Crisis Hotlines

Table 3. Call responder behaviors among calls made to California crisis hotlines (N = 210) M

Min.

Max.

# Significantly greater

# Significantly lower

Positive behaviors Allowed caller(s) to talk about feelings/situation

1.8

1.5

2.0

Reflected back caller(s)’ feelings

1.0

0.5

1.5

Reflected back caller(s)’ situation

1.2

1.6

0.8

Connected/established rapport with caller(s)

1.5

1.8

1.2

Overall, was sensitive/receptive to caller(s)’ problems

1.6

1.8

1.2

Was respectful

1.8

2.0

1.3

Showed empathy/validated caller(s) (i.e., “it must be hard for you”)

1.0

1.4

0.5

Challenged caller(s) (in negative way)

0.1

Was condescending

Displayed inappropriate behavior (i.e., fell asleep, laughed at caller[s])

0.1

Was judgmental

0.2

Preached or forced his/her opinions on caller(s)

0.2

Good contact

66%

43%

88%

Good contact with some weaknesses

24%

46%

17%

Negative behaviors

Contact with callers

Did not establish good contact/important weaknesses Problem solving Collaborative approach

10%

30%

Collaborative approach with some weaknesses

52%

38%

69%

Referrals provided

15%

46%

Referrals provided with some weaknesses

15%

46%

Decreased distress

43%

28%

64%

No change

30%

11%

54%

15%

Provision of referrals

Change in distress

Increased distress

Note. Percentages do not add to 100% because coders did not rate six calls in this area.

Contact With Callers Most centers displayed very good contact with callers (M = 66%, range = 43–88%) and few calls were rated as having important limitations in contact between callers and responders.

Problem Solving In all, 52% (range = 38–69%) of all calls were rated as exhibiting some weaknesses in using collaborative problem solving with callers while 10% (range = 0–30%) of calls were rated as having no weaknesses (the remainder did © 2016 Hogrefe Publishing

not require collaborative problem solving). Among the 151 calls for which problem solving was needed, 16% of calls used a collaborative approach with no major weaknesses: Two centers had more calls with no weaknesses, and at two centers, all calls had some weaknesses.

Provision of Referrals In total, 30% of all calls required referrals and, on average, 15% (range = 0–46%) of all calls were rated as having no weaknesses in the provision of referrals and a similar proportion and range were rated as having some weaknesses.

Crisis (2017), 38(1), 26–35

R. Ramchand et al.: California Crisis Hotlines

Table 4. Call, caller, and center-level predictors of call satisfaction and decreased distress (N = 210) Caller satisfaction β

Decreased distress OR

Call(er) characteristics Duration

.03

< .001

1.06

.029

Challenging situation

−.04

.554

0.75

.170

Challenging caller

−.30

.001

0.71

.068

Interpersonal problems

.28

.017

1.48

.260

Mental illness/substance abuse

.21

.097

0.97

.937

Work, housing, and financial problems

.04

.753

0.71

.400

−.11

.426

1.11

.801

.04

.797

1.82

.159

Call content

Physical illness Suicide Empathy

−.14

.274

0.81

.686

Trauma or loss

.08

.642

0.99

.974

Service use

.01

.986

0.67

.600

−.08

.857

0.10

.206

−.03

.855

2.72

.024

Offers crisis chat and/or text

.42

.014

0.85

.752

Exclusively volunteer

.23

.153

1.16

.747

Other Center characteristics Member of NSPL

Caller Satisfaction The average level of caller satisfaction was 3.4 with very little variability across centers (3.2–3.9).

Change in Caller Distress All callers who were categorized as “not at all distressed” at the beginning of the call (range across centers = 4–54%) were similarly categorized as not at all distressed at the end of the call (i.e., in no instance did a call start out as not at all distressed and experience increased distress). Among the remaining calls, calls were categorized as having no change, decreased distress (the anticipated effect), and increased distress. In general, just under half of these calls experienced reductions in distress with only one center with a statistically significant difference (Table 3). On the overwhelming majority of remaining calls, there was no change in level of distress, ranging from 11% to 54% of calls, although one center was rated as having 15% of calls with increased distress.

Crisis (2017), 38(1), 26–35

Individual and Center-Level Predictors of Call Outcomes We examined whether call/caller characteristics (i.e., duration, challenging situation, challenging caller, call content) and center-level characteristics (i.e., part of the NSPL, whether it also received chat and/or text, and whether it was run exclusively by volunteers) were correlated with level of satisfaction, and among calls that began with callers at some level of distress, with decreased distress. Across both outcomes, calls that were longer were rated as having callers with higher levels of satisfaction (β = .03, p < .001) and a higher likelihood of reduced distress (OR = 1.06, p = .029; Table 4). There was no relationship between challenging situation and either outcome, although challenging callers were rated as having lower levels of satisfaction (β = −.30, p = .001). Call content was not related to decreased distress, but callers who brought up interpersonal problems (β = .28, p = .017) exhibited, on average, higher levels of satisfaction with the call. With respect to center characteristics, those centers that offer crisis chat and text had, on average, higher levels of caller satisfaction (β = .42, p = .014); centers that were part of the NSPL also displayed, on average, a higher likelihood of having calls with decreased caller distress (OR = 2.72, p = .024). © 2016 Hogrefe Publishing

R. Ramchand et al.: California Crisis Hotlines

Discussion This study describes variability in content and quality across 10 call centers in California that were funded as part of a statewide initiative to prevent suicide. The call centers were similar to one another with respect to caller characteristics and the concerns callers raised during their calls. While centers varied somewhat in the degree to which situations described on calls were challenging, they did not vary as much with respect to how challenging callers were themselves. However, just as call centers varied in their structure and operating procedures, they differed in important ways. The proportion of callers at risk for suicide varied considerably from 3% to 57% of calls. These low rates of suicide calls to some of the crisis centers means that some call counselors, particularly volunteers who work part-time hours, may not encounter callers with suicidal risk very often and thus will have less experience with these types of calls. The low rates of suicidal content may be partly due to differences in adherence to suicide risk assessment guidelines – persons may only indicate risk for suicide if they are asked directly about it. Compliance with asking about current suicide risk ranged from 13% to 100%, while compliance to asking about past ideation and past attempts ranged from 4% to 77%, and 0% to 60%, respectively. Adherence to these guidelines is currently the only quality metric available for evaluating calls made to suicide hotlines (Joiner et al., 2007) and although call centers affiliated with NSPL tend to have higher rates of compliance, fewer than 50% of NSPL centers ask about recent ideation or attempts, suggesting a persistent problem among suicide hotlines (Gould et al., 2013; Mishara et al., 2007). Training call responders in ASIST did not enhance adherence in the most recent randomized trial of NSPL call centers (Gould et al., 2013), indicating specific training or protocols to promote adherence may be needed. Alternatively, new standards may need to be developed if the current recommendation of asking all callers about risk is impractical. Overall, call responders exhibited positive behaviors and at most centers never displayed negative behaviors. In addition, across all centers, 66% of calls were rated as having good contact between callers and call responders, similar to past evaluations (Gould et al., 2013; Mishara & Daigle, 1997; Mishara et al., 2007). However, there is room for improvement with respect to collaborative problem solving, which was noted as having “some weaknesses” in 38–69% of calls. One way to promote collaborative problem solving may be to encourage more reflective listening styles in which call responders reflect back callers’ situations and feelings. As previously mentioned, one evaluation found that Rogerian (reflective listening) versus © 2016 Hogrefe Publishing

directive therapeutic approaches were significantly associated with greater decreases in caller depression during the course of the call (Mishara & Daigle, 1997). For promoting these behaviors, training in ASIST, which uses Rogerian principles to guide intervention styles, may be appropriate and result in higher-quality calls (Gould et al., 2013). An important conclusion from our live monitoring study is that a large proportion (i.e., on average, 70% across centers) of calls to crisis hotlines do not require referrals to be made. In these instances, callers may already know, and be connected with, the most relevant resources in their community. However, for the 30% of calls that needed referrals, half of the referrals provided exhibited some weaknesses. We identified two reasons for such weaknesses. First, the referrals required (i.e., local resources to address financial or housing challenges) may be beyond the reach of counselors’ typical knowledge or training, which focuses on crisis intervention. Second, at some centers, resource guides were hard to navigate or out of date. In addition to improving these resources, other solutions might be to establish crisis centers with expertise in certain topics or populations (e.g., the Veterans Crisis Line; Knox et al., 2012) to serve as referral sources, or to develop strategies to continue engagement with the caller (e.g., asking for contact information for a follow-up call) until an appropriate referral is found rather than within the limits of a 15min call. Across centers, our monitors rated most callers as appearing to be satisfied at the end of the call, and 43% experienced a reduction in distress over the course of the call. In general, longer calls were positively linked to both outcomes; in the recent trial, call responders trained in ASIST had, on average, longer calls than those not trained (Gould et al., 2013), providing further evidence of the potential value of the training. Call content was not associated with decreased distress, but calls dealing with interpersonal problems and mental illness/substance use tended to have higher levels of satisfaction – perhaps because they are the most common calls received and thus responders have familiarity in these areas. It is noteworthy that centers that were part of the NSPL were more likely to experience reduced distress than those that were not. Individuals calling a crisis hotline may not have the opportunity to choose which call center answers the call nor which call responder they speak to. Thus, promoting quality across all call centers and minimizing variability is critical to ensure a consistent approach to suicide prevention. Currently, there is no system that authorizes or licenses suicide hotlines to operate; thus, accrediting bodies, funders, and crisis centers’ own motivation to improve call quality are the only policy vehicles by which such improvements can be made. Additional incentives and mandates to ensure ongoing call monitoring, either live or via Crisis (2017), 38(1), 26–35

recorded calls, would likely improve the consistency and quality across centers. The call-monitoring protocol used in this study is available at http://www.rand.org/pubs/ tools/TL150.html, as a potential resource for such ongoing monitoring. A limitation of the study is that we only sampled 241 calls, an average of 24 per center. The lack of evidence of significant differences may be due to an actual lack of difference or limited statistical power to discern a difference, thus, effects are statistically significant only where differences are relatively large. We also do not know how representative the calls we monitored are of calls made to the centers that participated in our evaluation, to all call centers in California, or to all call centers nationally. However, in many cases our proportions align with data from a subset of California hotlines that comprise the California Suicide Prevention Network (California Suicide Prevention Network, 2014). Thus, the results are likely to generalize somewhat beyond the current sample. Another limitation is that many of our ratings are subjective. Although we conducted extensive training and inter-rater reliability checks to ensure consensus (Jaycox et al., 2015), subjective ratings can be biased. The most practical way to overcome this limitation is for future research to code recorded calls rather than conduct live monitoring; unfortunately none of the centers in this study currently do so. Finally, all responders knew that they were being monitored during our evaluation, and may have exhibited behaviors or practices different from routine, day-to-day practice. In conclusion, the call centers that were part of this evaluation perform a valuable service to California residents. However, there is variation across centers and some areas for improvement at most centers. Ongoing quality improvement efforts to identify center-specific strengths and weaknesses would ensure that callers in distress will receive optimal, and beneficial, care no matter who picks up the phone. Acknowledgments This work was funded by The California Mental Health Services Authority (CalMHSA), an organization of county governments working to improve mental health outcomes for individuals, families, and communities. Prevention and early intervention programs implemented by CalMHSA are funded by counties through the voter-approved Mental Health Services Act (Prop. 63). Prop. 63 provides the funding and framework needed to expand mental health services to previously underserved populations and all of California’s diverse communities. The authors thank the directors and staff of the participating crisis lines for their participation in the study, including the gracious welcome provided to the study staff who visited their call centers. Crisis (2017), 38(1), 26–35

R. Ramchand et al.: California Crisis Hotlines

References Acosta, J., Ramchand, R., Jaycox, L. H., Becker, A., & Eberhart, N. (2012). Interventions to prevent suicide: A literature review to guide evaluation of california’s mental health prevention and early intervention initiative (TR-1317). Santa Monica, CA: RAND Corporation. California Suicide Prevention Network. (2014). Common metrics data report. Los Angeles, CA: Didi Hirsch Mental Health Services. Clark, W., Welch, S. N., Berry, S. H., Collentine, A. M., Collins, R., Lebron, D., & Shearer, A. L. (2013). California’s historic effort to reduce the stigma of mental illness: The Mental Health Services Act. American Journal of Public Health, 103(5), 786–794. doi:10.2105/AJPH.2013.301225 Cross, W. F., Pisani, A. R., Schmeelk-Cone, K., Xia, Y., Tu, X., McMahon, M., … Gould, M. S. (2014). Measuring trainer fidelity in the transfer of suicide prevention training. Crisis, 35(3), 202–212. doi:10.1027/0227-5910/a000253 Gould, M. S., Cross, W., Pisani, A. R., Munfakh, J. L., & Kleinman, M. (2013). Impact of applied suicide intervention skills training on the national suicide prevention lifeline. Suicide and Life-Threatening Behavior, 43(6), 676–691. doi:10.1111/sltb.12049 Gould, M. S., Kalafat, J., Harrismunfakh, J. L., & Kleinman, M. (2007). An evaluation of crisis hotline outcomes. Part 2: Suicidal callers. Suicide and Life-Threatening Behavior, 37(3), 338–352. Gould, M. S., Munfakh, J. L., Kleinman, M., & Lake, A. M. (2012). National suicide prevention lifeline: Enhancing mental health care for suicidal individuals and other people in crisis. Suicide and Life-Threatening Behavior, 42(1), 22–35. doi:10.1111/j.1943278X.2011.00068.x Gould, M. S., Lake, A. M., Munfakh, J. L., Galfalvy, H., Kleinman, M., Williams, C., … McKeon, R. (2016). Helping callers to the National Suicide Prevention Lifeline who are at imminent risk of suicide: Evaluation of caller risk profiles and interventions implemented. Suicide and Life-Threatening Behavior, 46(2), 172–190. doi:10.111/sltb.12182 Ho, T. P. (2006). Duration of hospitalization and post discharge suicide. Suicide and Life-Threatening Behavior, 36(6), 682–686. doi:10.1521/suli.2006.36.6.682 Jaycox, L. H., Ramchand, R., Ebener, P., Barnes-Proby, D., & Gilbert, M. L. (2015). RAND’s silent monitoring protocol for assessing suicide crisis line call content and quality. Santa Monica, CA: RAND Corporation. Joiner, T., Kalafat, J., Draper, J., Stokes, H., Knudson, M., Berman, A. L., & McKeon, R. (2007). Establishing standards for the assessment of suicide risk among callers to the national suicide prevention lifeline. Suicide and Life-Threatening Behavior, 37(3), 353–365. doi:10.1521/suli.2007.37.3.353 Kalafat, J., Gould, M. S., Munfakh, J. L., & Kleinman, M. (2007). An evaluation of crisis hotline outcomes. Part 1: Nonsuicidal crisis callers. Suicide and Life-Threatening Behavior, 37(3), 322–337. King, R., Nurcombe, B., Bickman, L., Hides, L., & Reid, W. (2003). Telephone counselling for adolescent suicide prevention: Changes in suicidality and mental state from beginning to end of a counselling session. Suicide and Life-Threatening Behavior, 33(4), 400–411. Knox, K. L., Kemp, J., McKeon, R., & Katz, I. R. (2012). Implementation and early utilization of a Suicide Hotline for veterans. American Journal of Public Health, 102(Suppl. 1), S29–32. doi:10.2105/AJPH.2011.300301 Lee, H. C., & Lin, H. C. (2009). Are psychiatrist characteristics associated with postdischarge suicide of schizophrenia patients? Schizophrenia Bulletin, 35(4), 760–765. doi:10.1093/schbul/ sbn007 Lin, H. C., Lee, H. C., Kuo, N. W., & Chu, C. H. (2008). Hospital characteristics associated with post-discharge suicide of severely de© 2016 Hogrefe Publishing

R. Ramchand et al.: California Crisis Hotlines

pressed patients. Journal of Affective Disorder, 110(3), 215–221. doi:10.1016/j.jad.2008.01.030 Mishara, B. L., Chagnon, F., & Daigle, M. (2007). Which helper behaviors and intervention styles are related to better short-term outcomes in telephone crisis intervention? Results from a silent monitoring study of calls to the U.S. 1-800-SUICIDE Network. Suicide and Life-Threatening Behavior, 37(3), 308–321. doi:10.1521/suli.2007.37.3.308 Mishara, B. L., Chagnon, F., Daigle, M., Balan, B., Raymond, S., Marcoux, I., … Berman, A. (2007). Comparing models of helper behavior to actual practice in telephone crisis intervention: A silent monitoring study of calls to the U.S. 1-800-SUICIDE Network. Suicide and Life-Threatening Behavior, 37(3), 291–307. doi:10.1521/suli.2007.37.3.291 Mishara, B. L., & Daigle, M. S. (1997). Effects of different telephone intervention styles with suicidal callers at two suicide prevention centers: An empirical investigation. American Journal of Community Psychology, 25(6), 861–885. Qin, P., & Nordentoft, M. (2005). Suicide risk in relation to psychiatric hospitalization: Evidence based on longitudinal registers. Archives of General Psychiatry, 62(4), 427–432. doi:10.1001/ archpsyc.62.4.427 Smith, A. R., Silva, C., Covington, D. W., & Joiner, T. E. (2014). An assessment of suicide-related knowledge and skills among health professionals. Health Psychology, 33(2), 110–119. doi:10.1037/ a0031062

Received September 8, 2015 Revision received February 3, 2016 Accepted February 20, 2016 Published online June 24, 2016 About the authors Rajeev Ramchand is a senior behavioral scientist and epidemiologist at the RAND Corporation, Arlington, VA, USA, who studies the prevalence, prevention, and treatment of mental health and substance use

disorders in adolescents, service members and veterans, and minority populations. Lisa Jaycox is a senior behavioral scientist and clinical psychologist at the RAND Corporation, Arlington, VA, USA. She has combined clinical and research expertise in the areas of child and adolescent mental health problems, including depression and reactions to violence exposure such as posttraumatic stress disorder. Patricia Ebener is a senior survey director at the RAND Corporation, Santa Monica, CA, USA, with expertise in designing, conducting, and managing survey projects for RAND’s Survey Research Group. Dionne Barnes-Proby is a senior research project manager at the RAND Corporation, Santa Monica, CA, USA, and has been a child welfare social worker for many years. Her experience includes direct clinical practice with at-risk children as a foster care case worker as well as research including juvenile justice, education, health, welfare reform, and substance abuse. Mary Lou Gilbert is a senior project associate at the RAND Corporation, Santa Monica, CA, USA, with extensive background in community-based services and field research. She has experience with a variety of field methods, including structured survey methodology, semistructured interviews, as well as qualitative case studies and ethnographies. Prodyumna Goutam is an assistant policy analyst at the RAND Corporation, Santa Monica, CA, USA, and a doctoral candidate at the Pardee RAND Graduate School with research interests in education and health policy. Rajeev Ramchand RAND Corporation 1200 South Hayes Street Arlington, VA 22202-5050 USA Tel. +1 (703) 413-1100 ext. 5096 Fax +1 (703) 413-8111 E-mail Ramchand@rand.org

Crisis (2017), 38(1), 26–35

Research Trends

Analysis of Internet Suicide Pacts Reported by the Media in Mainland China Fang-Fan Jiang, Hui-Lan Xu, Hui-Ying Liao, and Ting Zhang XiangYa School of Public Health, Central South University, Changsha, China Abstract: Background: In mainland China, frequent Internet suicide pacts in recent years have raised strong concerns from several social sectors and the influence of social networks on suicide is constantly growing. Aims: To identify the epidemiological characteristics of media-reported Internet suicide pacts in mainland China. Method: Our study comprised 62 Internet suicide pacts involving 159 victims in mainland China before June 1, 2015. Kendall’s randomness test, a trend test, and a circular distribution test were applied to identify the rising or concentrated trends in the time of occurrence of Internet suicide pacts. Results: The overall male-to-female ratio was 2.3:1. Suicide victims were mainly people in their 20s to 30s (84.1%). In all, 87.1% suicide victims completed suicide in sealed hotels or rental housing, and charcoal-burning suicide accounted for 80.6% of cases. Conclusion: People who complete suicide as part of an Internet suicide pact are more likely to be males, aged 20–30 years. Charcoal-burning suicide in sealed hotels or rental housing was the commonest way of dying. Keywords: Internet suicide pacts, charcoal burning, epidemiological characteristics, media, China

Internet suicide pacts, a new type of suicide also known as net suicide, refer to suicide pacts that are mutual agreements between two or more strangers who meet online to complete suicide together at a given place and time (Naito, 2007). Since the world’s first Internet suicide pact was exposed in Japan in 2000, Internet suicide pacts have rapidly increased in popularity in that country. On average, as many as 60 victims (nearly 20 cases) a year have died in Internet suicide pacts and the number continues to rise (Naito, 2007). This worrying new trend of Internet suicide pacts, however, is not limited to Japan and has gradually spread to other parts of the world (Naito, 2007). In mainland China, frequent Internet suicide pacts in recent years have attracted a great deal of attention from several social sectors and have been reported on extensively by the mass media. Despite much research related to “traditional” suicide pacts, in which the victims are people in close relationships, such as spouses, lovers, and friends (Rajagopal, 2004), far less is known about Internet suicide pacts. This is the first study to explore the epidemiological characteristics of media-reported Internet suicide pacts using a content analysis method, compared with previous qualitative case studies or purely theoretical reports.

Crisis (2017), 38(1), 36–43 DOI: 10.1027/0227-5910/a000402

Method Case Finding We conducted an extensive search for media reports covering incidents of Internet suicide pacts in mainland China through four popular search engines (i.e., Google China, Yahoo China, Baidu, and Sogou) and in the relevant literature from the three main Chinese electronic databases (e.g., CNKI, Wanfang, CqVip), Web of Science, Pubmed/ Medline, Embase (in June 2015) in Changsha City, Hunan province, China. The keywords in the search were Internet suicide pacts, Internet suicide, suicide pacts, charcoal-burning suicide, net suicide, Internet group suicide, and suicide QQ group. (QQ is an important social media platform in China, which allows multiple users to communicate instantly. A message sent by a member can be immediately received by all group members. A QQ suicide group is formed by members with suicidal ideation, and suicide-related issues are the most important topics of chatting.) Additional reports of Internet suicide pacts in mainland China in printed media were also collected. The study included all reports published until June 1, 2015. We collected reports of all cases of Internet suicide pacts during the study period in mainland China from the literature, media reports, forums, and other resources. Each media-reported Internet suicide pact case included in the study met the following © 2016 Hogrefe Publishing

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

criteria: (1) there was overwhelming evidence (e.g., the method used, signs supporting coercion or violence, suicide notes or statement, motive etc.) proving an incident of natural death, suicide, or homicide–suicide; (2) there was suicide incident that can be regarded as a suicide pact; (3) there were suicide clusters, that is, strangers who met over the Internet and had never met before completing suicide; (4) the suicide incident occurred in mainland China. All the reports of Internet suicide pacts were read carefully and judged by two of the authors separately. If the two authors’ opinions on a suicide incident were consistent, the suicide incident would be included or excluded; otherwise, we sought the assistance of a tutor. Audits and checks of the keyword search were carried out to ensure the completeness of the case-finding procedures.

Measures All data in the present study were obtained from media-reported Internet suicide pacts that occurred before June 1, 2015, in mainland China. When members of an Internet suicide pact die, a death certificate including the date, time, and cause of death is issued by medical doctors or coroners. Reportable deaths with an external cause code of X60–X84 that occurred in the study period were classified as suicide according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10; World Health Organization, 1992). In the present study, suicide from carbon monoxide poisoning due to charcoal burning was classified as ICD-10 code X67 because of no specific ICD code for charcoal-burning suicides (American Psychiatric Association, 2013). A victim in a media-reported Internet suicide pact who had a diagnosis by a psychiatrist or general practitioner, or a recorded past episode of suicide attempt and evidence of another attempt, and if the psychiatric disorder was reported in the media in accordance with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), was regarded as having a mental disorder. For all of the media-reported Internet suicide pacts, information based on death certiﬁcates and other evidentiary material – i.e., the time, place, suicide method used, as well as personal information such as name, gender, age, marital status, occupation, education level, native place, family background, precipitating factors of suicide, suicide signs, history of suicide attempts, and psychiatric disorders – was extracted and tabulated in Excel. Key information that was not included in the report was supplemented by searching other Internet resources with repeated checks.

Data Analysis The Statistical Package for the Social Sciences (SPSS, version 17.0) was used to analyze the data. Descriptive statistical analysis was used to explore the epidemiological characteristics of media-reported Internet suicide pacts (numbers of suicide clusters, distribution of suicide methods, distribution of suicide places, as well as distribution of age, profession, precipitating factors, suicide signs, psychiatric disorders, and a history of suicide attempts or self-harm). Numerical variables are expressed as means and standard deviations (SD), and categorical variables as numbers and percentage in brackets. Moreover, Kendall’s randomness test, trend tests, and circular distribution tests were also applied to identify the rising or concentrated trends in the time of occurrence of the media-reported Internet suicide pacts. A statistical map revealing the regional distribution of media-reported Internet suicide pacts was plotted using ArcGIS statistical software.

Results Demographic Characteristics of Suicide Victims This study included 62 media-reported Internet suicide pacts involving 159 suicide victims. There were 111 males (69.8%) and 48 females (30.2%), and the ratio of males to females was approximately 2.3:1. Among the 159 suicide victims, 52.2% (83/159) were suicides, 29.6% (47/159) were suicide attempters, and 18.2% (29/159) gave up suicide after being persuaded by their friends, relatives, or police. The average age was 23.1 ± 3.9 years, with a minimum of 13 years and a maximum of 40 years. The age range for men was 17–40 years (average, 23.3 years; SD = 3.8 years) and for women, 13–36 years (average, 22.9 years; SD = 4.2 years). Victims of media-reported Internet suicide pacts were mainly people aged 20–30 years (84.1%); six individuals (4.8%) were younger than 18 years. Regarding the careers of the suicide victims, peasant workers accounted for the greatest number of suicide victims (20/44, 45.5%), followed by students (11/44, 25.0%), career professionals (enterprise employees, teachers, nurses; 8/44, 18.2%), unemployed (3/44, 6.8%), and others (taxi drivers, business people; 2/44, 4.5%). However, the career of the remaining 115 suicide victims was not mentioned (Table 1).

Crisis (2017), 38(1), 36–43

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

Table 1. Demographic characteristics of victims of media-reported Internet suicide pacts Variables

Males, N (%)

Females, N (%)

Total, N (%)

Age group (years) < 20 20–30 > 30

9 (10.5)

5 (12.5)

14 (11.1)

73 (84.9)

33 (82.5)

106 (84.1)

4 (4.6)

2 (5.0)

6 (4.8)

5 (17.9)

6 (37.5)

11 (25.0)

Profession Students

13 (46.4)

7 (43.7)

20 (45.5)

Career professionals

Peasant workers

5 (17.9)

3 (18.8)

8 (18.2)

Other

2 (7.1)

(0)

2 (4.5)

Unemployed

3 (10.7)

(0)

3 (6.8)

Table 2. Number of victims in media-reported Internet suicide clusters Internet suicide pact victims, n

Internet suicide pact occurrences, N (%)

35 (56.5)

20 (32.2)

6 (9.7)

1 (1.6)

Number of Victims in Media-Reported Internet Suicide Pact Clusters The majority of media-reported Internet suicide pact clusters involved two suicide victims (35 occurrences, 56.5%), followed by three (20 occurrences, 32.2%), four (six occurrences, 9.7%), and five suicide victims (one occurrence, 1.6%; Table 2).

Time of Occurrence of Media-Reported Internet Suicide Pacts The number of suicide victims in reports was highest in 2013 with 48 suicide victims (48/148, 32.4%), and lowest in 2012 with 12 suicide victims (12/148, 8.1%; see Figure 1). The number of suicide victims in media reports was 26 (17.6%) in 2015, 25 (16.9%) in 2014, 24 (16.2%) in 2010, and 13 (8.8%) in 2011. Kendall’s randomness test showed that the numbers of suicide victims over the years were in random sequence (U = 0.386, p > .05), and the rising trend was analyzed via a trend test (U=6.572, p < .05). In terms of months, a peak was observed in May (24/148, 16.2%), followed by March (20/148, 13.5%) and April (19/148, 12.8%), with no suicide pact occurring in September (see Figure 2). There was a concentrated trend of completing suicide between January 25 and July Crisis (2017), 38(1), 36–43

18 as analyzed by the circular distribution test (r = .325, p < .05). Internet suicide pacts occurred at a peak on April 22. There was an obvious seasonal concentrated trend in summer (60/148, 40.5%) and spring (45/148, 30.4%). The time of occurrence of the remaining 11 suicide victims was unknown.

Regional Distribution of Media-Reported Internet Suicide Pacts All media-reported Internet suicide pacts were distributed in 16 provinces (municipalities), and the top three were located in the coastal areas of China, that is, Zhejiang province (10 occurrences, 17.2%), Fujian province (seven occurrences, 12.1%), and Guangdong province (six occurrences, 10.3%). The numbers of media-reported Internet suicide pacts in the Beijing municipality, Gansu province, Jiangxi province, and Shanxi province were the lowest, each with only one occurrence (1.7%). With regard to the seven geographical areas, there was an obvious regional concentrated trend in East China (Shanghai municipality, Shandong, Jiangsu, Anhui, Jiangxi, Zhejiang, Fujian province) with 33 occurrences (56.9%). However, the area of the remaining four occurrences was unknown (Figure 3).

Methods of Suicide in Media-Reported Internet Suicide Pacts In the present study, carbon monoxide poisoning due to charcoal burning was found to be the most popular method of dying (50/62, 80.6%), followed by taking sleeping pills (13/62, 21.0%). There were 50 Internet suicide pacts in which both members used the same suicide method. Taking sleeping pills and carbon monoxide poisoning by © 2016 Hogrefe Publishing

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

Figure 1. Annual number of suicide victims

Figure 2. Monthly number of suicide victims

Figure 3. Map of regional distribution of media-reported Internet suicide pacts. ÂŠ 2016 Hogrefe Publishing

Crisis (2017), 38(1), 36â&#x20AC;&#x201C;43

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

Table 3. Methods of suicide in media-reported Internet suicide pacts Suicide methods

Internet suicide pact occurrences, N (%)

Charcoal burning (CO poisoning)

50 (80.6)

Taking sleeping pills

13 (21.0)

Taking poison

4 (6.5)

Drowning

2 (3.2)

Hanging

2 (3.2)

Cutting veins

2 (3.2)

Other (home-made hydrogen sulfide)

1 (1.6)

Table 4. Sites of suicide in media-reported Internet suicide pacts Suicide sites

Internet suicide pact occurrences, N (%)

Hotels

42 (67.7)

Rental housing

12 (19.4)

Homes

2 (3.2)

Other

6 (9.7)

other places (6/62; Internet cafes, seaside, gardens, roofs, cars, bridges; Table 4).

Reasons for Suicide in Media-Reported Internet Suicide Pacts In the present study, the top precipitating factor for Internet suicidal behavior was introversion along with withdrawn and pessimistic mood (26/98, 26.5%); disharmony within the family and lack of family care and emotional problems were the second leading causes for suicide (25/98, 25.5% each); 25 victims completed suicide because of disharmony within the family and lack of family care, and 16 of them (64.0%) came from broken families (separation, divorce, or death of parents). However, 61 suicide victims did not give the reasons for their suicidal behavior (Table 5).

Signs of Suicide in Media-Reported Internet Suicide Pacts

Table 5. Reasons for suicide in media-reported Internet suicide pacts Reasons for suicide

Victims of Internet suicide pacts, N (%)

Disharmony within family and lack of family care

25 (25.5)

Emotional problems

25 (25.5)

Introversion, withdrawn, and pessimistic mood

26 (26.5)

Setbacks in work and life

21 (21.4)

Self-abasement

8 (8.2)

Physical illnesses

7 (7.1)

Financial problems

6 (6.1)

Death of relatives

5 (5.1)

Be discriminated, abused, or bullied

5 (5.1)

Pressure in school work

2 (2.0)

Other

1 (1.0)

charcoal burning were the commonest methods in pacts in which two ways of dying were used (Table 3).

Sites of Suicide in Media-Reported Internet Suicide Pacts Of the Internet suicide pacts reported in the media, 67.7% were carried out in sealed hotels (42/62), 19.4% in rental housing (12/62), 3.2% at homes (2/62), and 9.7% in Crisis (2017), 38(1), 36â&#x20AC;&#x201C;43

Almost all suicide victims likely experienced a mental process before completing suicide, beginning with suicidal ideation and progressing to the decision to complete suicide, and finally to carrying out suicide. This mental process probably revealed signs of suicide, such as expressing suicidal ideation (e.g., expressing despair and helplessness in words, asking for or discussing suicide-related information, and showing abnormal behavior), engaging in suicide preparation (getting oneâ&#x20AC;&#x2122;s affairs in order, making suicide plans), and writing notes. In this study, before completing suicide, all of the suicide victims (134/134, 100%) revealed suicidal ideation, of whom 93.3% (125/134) worked on suicide preparation, and 30 (22.4%) left suicide notes; however, the suicide signs of the remaining 25 suicide victims were unknown.

Psychiatric Disorders and Deliberate Self-Harm In this study, 23 suicide victims (14.5%; 13 males and 10 females) had mental disorders at the time of their death according to the DSM-5. Depressive disorders predominated (20/23, 87.0%). Two suicide victims had anxiety disorders (8.7%) and one (4.3%) had substance-related and addictive disorders. In all, 16 suicide victims (10.1%) had deliberately harmed themselves before (8 males and 8 females), of whom 10 (62.5%) had a history of two or more suicide attempts. ÂŠ 2016 Hogrefe Publishing

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

Discussion Main Findings A study by Ozawa-de Silva (2008, p. 530) suggested that an important reason for Internet suicide pacts is that individuals are “afraid to die alone.” Group suicide can contribute to eliminating the fear of death and strengthening suicidal ideation. In the present study, about 32.2% of all media-reported Internet suicide pacts were triple pacts, and double pacts accounted for 56.5%, which was inconsistent with previous findings in Japan. Curtin (2004) stated that Internet suicide pacts in Japan mainly involved three or more suicide victims, because double pacts may easily be misinterpreted as spouse pacts or lovers’ pacts. There was an obvious seasonal concentrated trend in the summer (40.5%) in the present study, which is similar to the trend observed for traditional suicide in China. As shown previously (Li, Sun, Qu, & Li, 1995; Xu, 2009), suicide in China is concentrated in the summer. We believe the seasonality of suicide in mainland China may be related to the seasonality of the Internet suicide pacts. The reasons for the seasonality of Internet suicide pacts remain unclear and further research is expected. It is recognized that suicide is related to depression, schizophrenia, and some types of substance addiction (e.g., drug addiction or Internet addiction; Dragisic, Dickov, Dickov, & Mijatovic, 2015; Gallego et al., 2015; Jia, Wang, Xu, Dai, & Qin, 2014). A previous study showed that more than 60% of all suicide victims had psychiatric disorders (Gunnell & Frankel, 1994). A survey by Ikunaga, Nath, and Skinner (2013) reported that 23% of 201 discussants in suicide sites have mental health problems or a specific diagnosis of a mental health disorder. In our study, 14.5% of suicide victims suffered from mental illness; therefore, people with mental health disorders should receive particular attention. Gunnell held that the suicide risk of suicide attempters is 20–30 times that of the general population (Gunnell & Frankel, 1994). In our study, 16 suicide victims (10.1%) had a history of suicide attempts, of whom 10 had a history of two or more attempts. The erroneous notion that suicide attempters would not try suicide again inhibited many people from identifying danger signs and seizing the best time for suicide prevention. In this study, 84.1% of all suicide victims were 20–30 years old, which is consistent with previous investigations in Japan. Ozawa-De Silva (2010) reported that 52.6% of Internet-related suicide victims are 20–30 years old. Numerous previous studies showed that the suicide rate in females is higher than in males in China (Qin, & Mortensen, 2008; Yang, Zhou, Huang, & Chen, 2004). By contrast, the number of male Internet suicide victims was 2.3 times © 2016 Hogrefe Publishing

higher than that of females, perhaps because males spent longer hours online than did females (Hagihara, Tarumi, & Abe, 2007). Young people are the most active group of Internet users. Studies have shown that 30% of all adolescents have suicidal thoughts and approximately 50% of them use the Internet. Adolescents who display dependency, insecurity, fear, and evasiveness may be especially at risk, since they may not resort to their friends, relatives, or psychologists for help and thus seek advice on the Internet instead (Becker, Mayer, Nagenborg, El-Faddagh, & Schmidt, 2004). Young people who live on their own can communicate with like-minded people about suicide-related issues in QQ groups and Internet forums, which may lead them to believe that suicide is the “right” option (Starcevic & Abouja, 2015) and to join like-minded suicidal persons to complete suicide together. In addition, some suicide websites and improper media reports on successful Internet suicide pacts that describe the suicide methods – including pictures of the scene, details of doses of lethal medication, details on necessary arrangements, graphic construction of the processes, especially about celebrity suicidal behavior – may facilitate copycat acts or the Werther effect in vulnerable populations (Yip & Lee, 2007). Furthermore, in some cases, people who had not initially desired suicide are also drawn into an Internet suicide pact (Naito, 2007). Another main finding of our study is that charcoal burning (nonviolent) was the commonest method of suicide in the media-reported Internet suicide pacts in mainland China. Rajagopal (2004) showed the commonest ways of dying in suicide pacts is poisoning by motor vehicle exhaust fumes (Brown & Barraclough, 1997). In recent years, the number of charcoal-burning suicides has been increasing rapidly in Asia, North America, Europe, and other regions (Chan et al., 2009; Liu, 2007), and charcoal-burning suicide has become the most popular method in Internet suicide pacts (Naito, 2007). There are several reasons for the popularity of charcoal burning as a method of suicide. First, charcoal-burning suicide is easier to share with partners than jumping off buildings, hanging, or other methods (Ikunaga, et al., 2013; Yip & Lee, 2007). Second, charcoal briquettes are affordable and easily accessible (Ikunaga et al., 2013). Finally, charcoal-burning suicide was romanticized as a quick, nonviolent, nondisfiguring, and painless way of dying in many suicide-related websites and in media coverage, and this potentially contagious impact of media reporting leads desperate individuals to imitate Internet suicidal behavior (Chan, Lee, & Yip, 2003; Yip & Lee, 2007). In addition, almost all of the Internet suicide pacts due to charcoal-burning were carried out in sealed hotels or rental housing. We think the preference of suicide attempters was influenced by other attempters and especially by reports of famous Internet suicide cases. Crisis (2017), 38(1), 36–43

In conclusion, the Internet, as an important tool of communication, was also utilized by young people to facilitate suicide. Its influence on suicide was much stronger than that of television, newspapers, and other traditional media because of the concealment and universality of online information on suicide (Tam, Tang, & Fernando, 2007).

Strengths and Limitations Many studies have shown the influence of the Internet on suicide, but Internet suicide pacts in mainland China are scarcely studied. To our knowledge, this is the first study to explore the characteristics of media-reported Internet suicide pacts in mainland China using a content analysis method, compared with previous works of a small number of qualitative case studies or purely theoretical reports. This study can provide a clue for understanding the characteristics of Internet suicide pacts in mainland China and offer scientific bases for suicide prevention and intervention. However, this study also has a few limitations. On the one hand, the present study did not include data on all Internet suicide pacts in mainland China because of the difficultly in accessing official data on Internet suicide pacts, which has legal restrictions. Only media-reported Internet suicide pacts were analyzed. However, these can provide a clue for further studies on Internet suicide pacts. On the other hand, although we conducted an extensive search for the main information of each pact through various network channels, some data are not complete, such as marital status, occupation, education level, and family background of the suicide victims. We expect that more detailed data on the process of communicating suicide-related issues and making plans for Internet suicide pacts, on the precipitating factors for Internet suicidal behaviors, and on what the Internet suicide attempters were thinking will help us to explore this knowledge gap in future studies.

Conclusion The main characteristics of media-reported Internet suicide pacts in mainland China can be summarized as:(1) People who complete suicide in Internet suicide pacts are more likely to be males, aged 20–30 years; and (2) carbon monoxide poisoning by charcoal burning in sealed hotels or rental housing was the commonest way of dying. Acknowledgments We would like to acknowledge Lin Chen’s guidance of statistical method and Cai-Qian He for language support. The authors have no conflict of interest to declare. Crisis (2017), 38(1), 36–43

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

References American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders, ﬁfth edition. DSM-5 (5th ed.). Arlington, VA: Author. Retrieved from http://www.psychiatry.org/ Becker, K., Mayer, M., Nagenborg, M., El-Faddagh, M., & Schmidt, M. H. (2004). Parasuicide online: Can suicide websites trigger suicidal behaviour in predisposed adolescents? Nordic Journal of Psychiatry, 58(2), 111–114. doi:10.1080/08039480410005602 Brown, M., & Barraclough, B. (1997). Epidemiology of suicide pacts in England and Wales, 1988-92. British Medical Journal, 315, 286–287. doi:org/10.1136/bmj.315.7103.286 Chan, S. S., Chiu, H. F., Chen, E. Y., Chan, W. S., Wong, P. W., Chan, C. L., … Yip, P. S. (2009). What does psychological autopsy study tell us about charcoal burning suicide – a new and contagious method in Asia? Suicide and Life-Threatening Behavior, 39(6), 633–638. doi:10.1521/suli.2009.39.6.633 Chan, K. P. M., Lee, D. T. S., & Yip, P. S. F. ( 2003). Media influence on suicide. Media’s role is double edged. British Medical Journal, 326(7387), 499. Retrieved from http://hdl.handle. net/10722/43676 Curtin, J. S. ( 2004). Suicide in Japan: Part seventeen – profile of internet suicide groups. Global Communications Platform: Japanese Institute of Global Communication. Retrieved from http:// www.glocom.org/special_topics/social_trends/20041104_ trends_s91 /index.html Dragisic, T., Dickov, A., Dickov, V., & Mijatovic, V. (2015, June). Drug addiction as risk for suicide attempts. Materia Socio Medica, 27(3), 188–191. doi:10.5455/msm.2015.27.188-191 Gallego, J. A., Rachamallu, V., Yuen, E. Y., Fink, S., Duque, L. M., & Kane, J. M. (2015). Predictors of suicide attempts in 3322 patients with affective disorders and schizophrenia spectrum disorders. Psychiatry Research, 228(3), 791–796. doi:10.1016/j. psychres.2015.05.024 Gunnell, D., & Frankel, S. (1994). Prevention of suicide: Aspirations and evidence. British Medical Journal, 308, 1227–1233. doi:org/10.1136/bmj.308.6938.1227 Hagihara, A., Tarumi, K., & Abe, T. (2007). Media suicide-reports, internet use and the occurrence of suicides between 1987 and 2005 in Japan. BioMed Central Public Health, 7, 321. doi:10.1186/1471-2458-7-321 Ikunaga, A., Nath, S. R., & Skinner, K. A. (2013). Internet suicide in Japan: A qualitative content analysis of a suicide bulletin board. Transcultural Psychiatry, 50(2), 280–302. doi:10.1177/1363461513487308 Jia, C. X., Wang, L. L., Xu, A. Q., Dai, A. Y., & Qin, P. (2014). Physical illness and suicide risk in rural residents of contemporary China: A psychological autopsy case-control study. Crisis, 35(5), 330–337. doi:10.1027/0227-5910/a000271 Li, W. S., Sun, X. Y., Qu, L. M., & Li, S. S. (1995). Analysis on 185 suicide cases [in Chinese]. Chinese Journal of Psychiatry, 28(1), 1921. Retrieved from http://www.cnki.net Liu, K. Y. (2007). Charcoal burning suicides in Hong Kong and urban Taiwan: An illustration of the impact of a novel suicide method on overall regional rates. Journal of Epidemiology & Community Health, 61(3), 248–253. doi:10.1136/jech.2006.048553 Naito, A. (2007). Internet suicide in Japan: Implications for child and adolescent mental health. Clinical Child Psychology and Psychiatry, 12(4), 583–597. doi:10.1177/1359104507080990 Ozawa-de Silva, C. (2008). Too lonely to die alone: Internet suicide pacts and existential suffering in Japan. Culture, Medicine, and Psychiatry, 32, 516–551. doi:10.1007/s11013-008-9108-0 Ozawa-De Silva, C. (2010). Shared death: Self, sociality and internet group suicide in Japan. Transcultural Psychiatry, 47(3), 392– 418. doi:10.1177/1363461510370239392-418 © 2016 Hogrefe Publishing

F.-F. Jiang et al.: Media-Reported Internet Suicide Pacts

Qin, P., & Mortensen, B. P. (2008). Specific characteristics of suicide in China. Acta Psychiatrica Scandinavica, 103(2), 117–121. doi:10.1034/j.1600-0447.2001.00008.x Rajagopal, S. (2004). Suicide pacts and the internet. British Medical Journal, 329, 1298–1299. doi:10.1136/bmj.329.7478.1298 Starcevic, V., & Abouja, E. (2015). Cyberchondria, cyberbullying, cybersuicide, cybersex: “New” psychopathologies for the 21st century? World Psychiatry, 14(1), 97–100. doi:10.1002/wps.20195 Tam, J., Tang, W. S., & Fernando, D. J. (2007). The internet and suicide: A double-edged tool. European Journal of Internal Medicine, 18(6), 453–455. doi:10.1016/j.ejim.2007.04.009 World Health Organization. (1992). The tenth revision of the International Classification of Diseases and related health problems (10th ed.). Geneva, Switzerland: Author. Retrieved from http:// www.who.int/en/ Xu, H., & Li, J. (2015). The study on the situation of Internet suicide pacts. Journal of Psychiatry, 28(2), 153–155. doi:10.3969 /j.issn.2095-9346.2015.02.023 Xu, H. L. (2009). Psychosocial autopsy study of suicides by people aged 15-35 in rural Hunan (Doctoral dissertation) [in Chinese]. Retrieved from China National Knowledge Infrastructure and Central South University Repository (UMI No. 2009208466). Yang, G. H., Zhou, L. N., Huang, Z. J., & Chen, A. P. (2004). The trend and geographic distribution of suicide in Chinese population. Chinese Journal of Epidemiology, 25(4), 280–284. Retrieved from http: //www.cnki.net Yip, P. S. F., & Lee, D. T. S. (2007). Charcoal-burning suicides and strategies for prevention. Crisis, 28(S1), 21–27. doi:10.1027/02275910.28.S1.21 You, Z. Q., Han, X. P., Lü, L., & Yeung, C. H. (2015). Empirical studies on the network of social groups: The case of Tencent QQ. PLoS One, 10(7), e0130538. doi:10.1371/journal.pone.0130538

Received September 18, 2015 Revision received February 16, 2016 Accepted February 20, 2016 Published online June 9, 2016 About the authors Fang-Fan Jiang is a master’s degree student at XiangYa School of Public Health, Central South University, China. Her research focuses on suicide prevention and crisis intervention, social epidemiology, and chronic diseases prevention. Hui-Lan Xu, MD, is Professor and Director of the Department of Social Medicine and Heath Management at XiangYa School of Public Health, Central South University, China, and Vice-Director of the Institute of Suicide Prevention, Central South University. Her research interests include suicide prevention and crisis intervention, social epidemiology, and chronic diseases prevention. Hui-Ying Liao is a master’s degree student at XiangYa School of Public Health, Central South University, China. Her research interest is health law and health policy. Ting Zhang is a master’s degree student at XiangYa School of Public Health, Central South University, China. Her research interest is social epidemiology. Hui-Lan Xu XiangYa School of Public Health Central South University 110 Xiangya Road Kaifu District, Changsha, Hunan, 410078 China Tel. +86 731 8480-5459 Fax +86 731 8480-5454 E-mail xhl6363@126.com

Crisis (2017), 38(1), 36–43

Research Trends

An Analysis of Depression, Self-Harm, and Suicidal Ideation Content on Tumblr Patricia A. Cavazos-Rehg, Melissa J. Krauss, Shaina J. Sowles, Sarah Connolly, Carlos Rosas, Meghana Bharadwaj, Richard Grucza, and Laura J. Bierut Department of Psychiatry, Washington University School of Medicine, St. Louis, MO, USA Abstract. Background: Social networking about depression can be indicative of self-reported depression and/or can normalize risk behaviors such as self-harm and suicidal ideation. Aim: To gain a better understanding of the depression, self-harm, and suicidal content that is being shared on Tumblr. Method: From April 16 to May 10, 2014, 17 popular depression-related Tumblr accounts were monitored for new posts and engagement with other Tumblr users. A total of 3,360 posts were randomly selected from all historical posts from these accounts and coded based on themes ascertained by the research team. Results: The 17 Tumblr accounts posted a median number of 185 posts (range = 0–2,954). Content was engaged with (i.e., re-blogged or liked) a median number of 1,677,362 times (range = 0–122,186,504). Of the 3,360 randomly selected posts, 2,739 (82%) were related to depression, suicide, or self-harm. Common themes were self-loathing (412, 15%), loneliness/feeling unloved (405, 15%), self-harm (407, 15%), and suicide (372, 14%). Conclusion: This study takes an important first step at better understanding the displayed depression-related references on Tumblr. The findings signal a need for suicide prevention efforts to intervene on Tumblr and use this platform in a strategic way, given the depression and suicidal content that was readily observed on Tumblr. Keywords: social media, depression, self-injurious behavior, suicidal ideation, adolescent

Social networking sites (SNS) are online communities whereby individual users virtually connect to each other, resulting in their own personal networks (Lenhart & Madden, 2007). These networks can consist of family, friends, coworkers, classmates, and even strangers with shared interests. SNS have transformed the way in which people communicate with each other. Over the past decade, SNS users have skyrocketed from just 7% of all US adults in 2005 to 65% in 2015 (Perrin, 2015). Youth and young adults make up the largest groups of SNS users, with approximately 90% of 13–29-year-old Internet users using at least one SNS (Brenner, 2014). SNS have presented young people with new opportunities for self-expression, and the possibility to interact with peers without adult supervision (Boyd, 2014). In particular, the option to create accounts under a pseudonym or obscure username that separates users from their true identity may facilitate a sense of anonymity that allows for the disclosure of sensitive information on these sites (Marwick & Boyd, 2011; Robinson et al., 2015; Tufekci, 2008). Given the widespread use of SNS, especially among young people, there has been emerging interest in studying posts about mental health problems (i.e., expressions of depression, suicide, and/or self-harm behaviors) on social media platforms and online. For instance, Moreno et Crisis (2017), 38(1), 44–52 DOI: 10.1027/0227-5910/a000409

al. (2011) found that 25% of a sample of 200 college-aged SNS users socially networked about feeling depressed on Facebook, and related studies have found that individuals who post depression-related content on SNS likewise self-report symptoms of depression (Moreno et al., 2012; Park, Lee, Kwak, Cha, & Jeong, 2013). For 2 months in 2009, Microsoft researchers monitored the keyword depression on Twitter and obtained 20,000 depression- related tweets; their analysis of these tweets suggested that tweeters were posting about their depression and mental health treatment (Park, Cha, & Cha, 2012). In a recent review of suicide and social media, Robinson, Rodrigues, Fisher, and Herrman (2014) reported that in the six studies that examined the relationship between suicide and social media, people commonly used social media as a space to interact with individuals who shared similar problems. In one such study, Baker and Fortune (2008) conducted interviews with young adults that utilized suicide and self-harm websites. Their participants viewed these websites as supportive communities that assisted them in coping with their problems (Baker & Fortune, 2008). Additionally, in an analysis of responses to YouTube videos containing content relating to self-harm, the viewers often socially networked with the individual posting the content and subsequently shared their own experiences with © 2016 Hogrefe Publishing

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

self-harm. Some viewers also expressed admiration and/ or encouragement for the individual posting the self-harm content (Lewis, Heath, Sornberger, & Arbuthnott, 2012). However, the comments rarely mentioned recovery and many expressed continued engagement in the self-harm behavior (Lewis et al., 2012). These studies indicate that online social networking about depression and related behaviors occurs on SNS and may normalize risk behaviors such as self-harm and suicidal ideation (Whitlock, Powers, & Eckenrode, 2006; Ybarra, Alexander, & Mitchell, 2005). Foundational research on this topic exists for many social media platforms; however, one site is curiously absent despite its increasing popularity. Tumblr is a mostly anonymous, content-centered microblogging site that allows users to blog with text, pictures, and video clips about topics that are of interest to them, creating a community of people with shared interests. Tumblr has approximately 268 million users (i.e., blogs) and 125.7 billion posts (http:// tumblr.com/about), and 42% of its user traffic originates from the US (http://tumblr.com/press). Tumblr is widely used among teens and young adults, with almost one seventh (14%) of teens aged 13–17 using Tumblr and 40% of Tumblr users being 13–25 years old (Brenner, 2014; Zain, 2013). Tumblr’s lenient user terms (e.g., allowing users to sign up/interact anonymously), its role as a largely public site, and its content-centered nature helps fuel its popularity among youth and young adults (Mander, 2014). The growing popularity of Facebook among older adults is another factor that may be driving teens and young adults to the community of Tumblr users (Brenner, 2014). With its multiple types of media options, Tumblr offers itself as an alternative to more restrictive social media platforms such as picture-based Instagram and character-limiting Twitter. Tumblr is also unique because, instead of using the platform to connect with family, friends, and coworkers, many users choose to connect with members of the Tumblr community whom they may not know personally (i.e., offline), but with whom they share common interests (Attenberg, 2012). Because of this, it is possible that a community of Tumblr users will share depression as their focus. This exploratory study examines the expression of depression-related content on Tumblr, specifically focusing on the most popular posts (as designated by the Tumblr search algorithm) relating to depression, self-harm, and suicidal ideation. Self-harm is defined as deliberate self-injury or poisoning, regardless of suicidal intent (Hawton et al., 2003; Hawton, Saunders, & O’Connor, 2012). The content and engagement of these posts are also characterized. Gaining a better understanding of the types of depression, self-harm, and suicidal content that is being shared on Tumblr could be an important step toward informing suicide prevention efforts on social media.

Method The Tumblr data in the current study are public. The University’s Institutional Review Board (IRB) reviewed the project and determined that it does not involve activities that are subject to Institutional Review Board oversight.

Identifying Tumblr Accounts Related to Depression and Self-Harm On April 16, 2014, the research team searched Tumblr for posts related to the terms depressed and suicide. In addition, because many young people who engage in self-harm behaviors suffer from depression (Andover et al., 2005; Moran et al., 2012), the research team also examined Tumblr for posts related to the terms self-mutilation and cutting. When searching for these terms, Tumblr first provides a screen with suggestions for seeking help or finding more inspirational content, and then provides the option of continuing to the next screen to view the results of the search (see Figure 1). The Tumblr search landing page defaults to the Most Popular posts displayed in a grid view. The research team searched Tumblr for the most popular posts related to each of the four search terms. Accounts with posts appearing on the top of the results page were scanned by the coders, and five were selected for each search term (total n = 20 Tumblr accounts for analysis). Only the accounts

Figure 1. First Tumblr screen after searching for the depression-related terms. Crisis (2017), 38(1), 44–52

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

Figure 2. Example of the search results page on Tumblr (search was for self-mutilation, on July 7, 2014).

that explicitly posted about depression-related content (i.e., posted images, video clips, and/or text that expressed feelings of depression, suicide, and/or self-harm behaviors) in about 50% of the posts were selected for analysis. Screenshots were not captured from the initial searches on April 16, but as an example Figure 2 shows a screen shot of results from a search for self-mutilation performed on July 7, 2014.

Characteristics of Tumblr Accounts and Engagement With Other Users Similar to other SNS, Tumblr allows its users to provide basic information such as gender and age. Thus, in order to characterize the selected Tumblr accounts, each account was reviewed individually for demographic information provided on their profile page. Simply Measured, a company that provides social media analytics and measurement (http://www.simplymeausured.com), was used to collect all available posts (all historical posts that were available) for the 20 Tumblr accounts chosen and to monitor the posts and engagement of each account with other Tumblr users over several weeks (April 16 to May 10, 2014). Because of the exploratory nature of the study, this shorter period of time for monitoring posts was chosen to offer a snapshot of the acCrisis (2017), 38(1), 44â&#x20AC;&#x201C;52

countsâ&#x20AC;&#x2122; activity. Before data collection and analysis could be completed, three Tumblr accounts became inactive, leaving 17 accounts for analysis. To describe the engagement of each account with other Tumblr accounts, the number of posts from the account during the reporting period was documented. The number of times someone else re-blogged their content during the reporting period and the total number of notes a post had from its creation (original post date) to the end of the report time period (lifetime notes) were also recorded. On Tumblr, a post gets a note when someone either likes or re-blogs the post.

Coding Tumblr Posts for Common Depression-Related Themes Using the SAS Proc SURVEYSELECT procedure, 200 posts were randomly selected from all posts available for each Tumblr account, including historical posts (or all posts if less than 200 posts were available), resulting in 3,360 posts for analysis. This number of posts was manageable in enabling manual coding and permitting discussion between team members as needed. Two members of the research team with expertise in mental health research studied a sample of these posts and developed a list of relevant themes based on repeatedly observed topics (Ryan & ÂŠ 2016 Hogrefe Publishing

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

Bernard, 2003; listed here). These senior team members trained five research assistants and those five viewed the content of the posts online and coded the posts appropriately. Each post was coded by two research assistants individually. Any discrepancies were then discussed and resolved. Coders first determined whether the post contained images, text, or video relating to depression, suicide, and/ or self-harm behaviors. Those that were not related to the subject of interest were excluded from further analysis. For posts that were related to the subject of interest, the type of content of the post was then coded as a photo/image, text (including images of text), and/or short video clip (i.e., .gif), and it was noted whether a picture or video clip from a popular movie/television show was used. Coders then examined each post and coded for the following five themes: (1) self-loathing, (2) loneliness/feeling unloved, (3) self-harm and/or scars from self-harm, (4) suicide/suicidal thoughts/death, (5) posts of graphic images/video clips related to suicide or self-harm. Last, it was noted whether the post involved interacting with another Tumblr user (e.g., asking/answering a question or correspondence between users), including seeking/giving advice, recommending professional help/therapy, or supporting each other. Each post could have more than one of the above themes. In addition to examining themes across posts in the 17 accounts in our sample, the most prominent themes within each individual Tumblr account were also determined in order to identify the primary topic that was most emphasized by the owners of each account.

A random sample of 200 Tumblr posts was also coded by a senior research team member for reliability of the research assistants. Percent agreement for whether the post was related to depression was 91% (κ = .69). Median percent agreement for all other themes was 94% (range = 82–100%) and median κ value was .79 (range = .57–1.0).

Results Tumblr Account Demographics and Engagement Of the 17 Tumblr accounts, nine did not provide demographic information in their profile. Of the eight Tumblr accounts that did provide some demographic information, six were female and two were male users; ages provided ranged from 14 to 20 years. The number of posts and engagement of this study’s sample of Tumblr users with other Tumblr users is shown in Table 1. The median number of total posts across accounts was 874 (range = 160– 11,378). About half of the accounts posted mostly original content, while the rest relied mostly on re-blogs from others’ accounts. Engagement of posts with other Tumblr users was tracked over approximately 3.5 weeks in April and May of 2014. The median number of posts during that period was 185 (range = 0–2,954). The median number of times their content was re-blogged was 6,893 (range = 0–407,796). Finally, the median number of total lifetime

Crisis (2017), 38(1), 44–52

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

Table 1. Engagement of the sample of Tumblr accounts that focus on depression, suicide, or self-harm Account number

No. of available posts published anytime

Percentage of all available posts that are original posts (vs. re-blogs)

No. of posts during 4/16–5/10/2014

No. of reblogs during 4/16–5/10/2014

No. of lifetime notes on posts published during 4/16–5/10/2014a

Most popular accounts to post about depressed 1.

1,565

99%

508

260,369

383,467

1,337

86%

992

343,507

13,615,908

874

32%

3,389

1,677,362

425

71%

754,277

160

100%

6893

30,705

Most popular accounts to post about suicide 6.

11,378

68%

2,954

407,796

83,962,003

3,339

12%

412

49,913

40,538,023

394

78%

111

20,479

1,143,416

13%

2,904

286,493

122,186,504

Most popular accounts to post about self-mutilation 9.

6,154

10.

1,734

318

243

19,964,880

11.

769

13%

254

1,868

15,055,666

12.

538

78%

498

3,479,473

13.

475

13%

Most popular accounts to post about cutting 14.

1,149

15%

185

1,974

6,365,903

15.

1,009

100%

268

101,041

148,432

16.

307

100%

18,003

29,844

17.

203

30%

4,986

258,487

Note. aThe number of lifetime posts also includes notes (i.e., someone likes or re-blogs its content) on re-blogs that were originally posted before the reporting period.

notes (i.e., someone either likes or re-blogs the original post) on posts published during this period was 1,677,362 (range = 0–122,186,504).

Themes Across Accounts Up to 200 posts from each account were randomly selected for qualitative analysis. This selection was from all posts available for each Tumblr account. Of the 3,360 total posts that were randomly selected, 350 (10%) did not appear to be related to depression, suicide, or self-harm and were excluded from content analysis. In addition, the links for 271 (8%) of the posts did not work and thus could not be analyzed. This left 2,739 posts for qualitative analysis. The most common themes found in the 2,739 depression-related posts are shown in Figure 3, along with example posts. Posts about self-loathing, loneliness/feeling unloved, and self-harm and/or scars from self-harm were the most popular observed themes. Specifically, self-loathing was apparent in 15% (412) of the posts, and 15% (405) of Crisis (2017), 38(1), 44–52

the posts expressed loneliness or feeling unloved. Approximately 15% (407) of the posts were about self-harm and/ or scars from self-harm; 14% (372) were about suicide or death. Among the 717 posts that were about self-harm or suicide/death, 18% (127) were graphic images/video clips. Approximately 8% (220) of the posts were comforting, supportive, or prevention posts. Nine percent (249) of the posts involved directly interacting with another Tumblr user, among which 47% (117/249) provided emotional support or reassuring messages to each other. Among these interactive posts, over half (127/249, 51%) represented seeking or providing advice, 41% of which (52/127) provided positive/supportive advice (e.g., encouragement in stopping self-harm or fighting depressing thoughts), 25% of which (32/127) provided potentially harmful advice (e.g., advising how to secretly engage in self-harm or maladaptive behaviors), and 34% of which (43/127) could not be distinguished as positive/supportive or potentially harmful advice. Among posts providing advice, only 13% (17/127) suggested professional help or therapy to cope with their mental health problems. © 2016 Hogrefe Publishing

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

The media used in the posts varied, including text only (some of which were images of text, 1,024, 37%), text and photos/images together (556, 20%), text and short video clips (i.e. gifs; 545, 20%), photos/images only (438, 16%), and video clips only (176, 6%). Approximately 15% (407) of the posts used images or video clips of celebrities from popular movies or television shows.

Prominent Themes Within Individual Accounts Over 20% of the posts from three accounts focused on suicide/death. Nearly 30% or more of the posts from four accounts focused on self-harm. Two accounts expressed self-loathing in nearly 30% of their posts. Some of the selected accounts also had a relatively high number of posts that were direct interactions (i.e., engaging in online correspondence) with other Tumblr users, including one account where the majority of posts (63%) were interactions with other Tumblr users. Among the interactive posts found within this account, most (59/90, 66%) represented seeking or providing advice, 37% of which (22/59) provided positive/supportive advice (e.g., tips on coping with/overcoming eating disorders, finding resources to help fight depression), 46% of which (27/59) provided potentially harmful advice (e.g., advising how to fast to lose weight or how to skip meals without others knowing), and 17% of which (10/59) could not be distinguished as positive/supportive or potentially harmful advice.

Discussion Social media use has become increasingly popular, and there is emerging research signaling that social networking about depression occurs online (Moreno et al., 2011; Moreno et al., 2012; Park et al.,2012; Park et al., 2013). In response, this exploratory study examines the most popular posts relating to depression and suicide on Tumblr, a popular social media platform where individuals can network with other members via text, pictures, and video clips about topics that are of interest to them. In addition, because many young people who engage in self-harm behaviors suffer from depression (Laye-Gindhu & Schonert-Reichl, 2005; Lloyd-Richardson, 2007), Tumblr accounts related to the terms self-mutilation and cutting were also examined. The most common themes from these accounts were posts about self-hatred, loneliness, and suicide/ death. Posts about self-harm, including graphic pictures of cutting, were also readily shared. ÂŠ 2016 Hogrefe Publishing

In the present study, an extensive amount of depression-related content came from popular Tumblr posts that displayed harmful behaviors including self-harm behaviors and suicide. In recent years, researchers have found that SNS can facilitate social networking about health-related topics, even concerns about challenging conditions such as a cancer diagnosis and mental health topics (Himelboim & Han, 2014; Naslund, Grande, Aschbrenner, & Elwyn, 2014), and there is some indication that online communities can provide a supportive, nurturing environment that helps individuals cope with their struggles (Baker & Fortune, 2008; Barak & Dolev-Cohen, 2006; Hawton, Rodham, Evans, & Weatherall, 2002). However, there is concern that the posts observed in this study are glorifying these dangerous behaviors, especially when considering that suicide and suicidal behaviors can cluster among youth and young adults (Becker, Mayer, Nagenborg, El-Faddagh, & Schmidt, 2004; Duggan, Heath, Lewis, & Baxter, 2012; Lewis, Heath, St. Denis, & Noble, 2011; Mitchell & Ybarra, 2007; Whitlock, Powers, & Eckenrode, 2006). Because the vast majority of social media users are young people who may be easily influenced by their peers and the media, it is important to work toward understanding why this content is being posted and if there is an impact on the individuals who post it as well as those who view it. The posts studied in this sample connected with a sizeable audience, as evidenced by the large number of responses from other Tumblr users (i.e., re-blogs and lifetime notes) accumulated by these blogs. Given the spread of the posts observed, it is inferred that many Tumblr users have an interest in and likely seek out this type of content. In addition, direct interactions between users about depression, suicide, and/or self-harm behaviors on Tumblr was observed in 9% of the posts, and most of these extended conversations were online exchanges where individuals offered emotional support and/or sought out suggestions on how to cope. Advice exchanges between Tumblr users tended to be positive and/or supportive and thus show promise for benefiting individuals in need of support. Still, a sizeable amount of discussions was observed to be potentially harmful advice, and in these cases, oversight from parents and/or mental health professionals appears warranted in order to offset interactions that seem to endorse unhealthy behaviors. Given the content of these accounts, it is not surprising that few of them provided identifying information in their profiles. It is likely that anonymous accounts help to facilitate the disclosure of intimate thoughts and feelings about depression observed in this sample, and it is also probable that providing demographic information would hinder this type of expression. Nonetheless, of those accounts that contained age details, there were a range of youth and young adults who were between 14 and 20 years old. It is Crisis (2017), 38(1), 44â&#x20AC;&#x201C;52

concerning that individuals who by self-report are quite young are posting this type of content, especially when it is about engaging in self-injury and/or struggling with suicidal ideation and there were few prevention messages that countered this content. A better understanding is needed on how to respect the freedom of expression and support that these social media platforms facilitate while using social media data in order to identify and help individuals who may benefit from intervention. This study is exploratory and the first of its kind to examine the depression, suicide, and self-harm related content on Tumblr; therefore, a number of limitations are worth noting. Seventeen accounts were examined, based on posts designated by Tumblr to be among the most popular resulting from selected search terms on a specific day. Including more accounts, and perhaps a more comprehensive list of search terms and at different time points, would create a more complete picture of what is being shared on this platform. The biographical and demographic information provided on Tumblr varies. For the accounts in this study, only eight provided some form of demographic information, and its absence hinders the research application of understanding the population of interest. In addition, three of the accounts that were selected for analysis became inactive prior to analysis. The reason for these accounts becoming inactive is unknown, but could be related to Tumblr’s policy that permits discussion and support about sensitive topics like self-injury but does not allow blogs that “actively promote self-harm” (Tumblr, 2015). Despite these limitations, coding over 3,000 posts allowed a baseline picture to be established of the depression, self-harm, and suicidal ideation content on Tumblr and can serve as a starting point for discussions of this topic in the mental health community. The results here show that blogging about self-hatred, loneliness, and suicidal thoughts occurs on Tumblr. Furthermore, the online dialog indicates that people who socially network about these topics may be forming relationships to chat about their shared interests in depression, self-harm, and/ or suicidal ideation. The accounts/posts in this study are considered to be popular and are potentially connecting with a sizeable audience. Peer-led efforts to aid in suicide and self-harm prevention have started to appear across the Tumblr platform (Cuen, 2015). The American Foundation for Suicide Prevention supports this form of initial contact by peers with a subsequent referral to a skilled professional (Cuen, 2015; American Foundation for Suicide Prevention, 2015). This research presents an opportunity for mental health professionals to work toward further understanding how people socially network about these topics, assess the content of the conversations including how it materializes, and consider ways to access those at high risk in order to offer appropriate avenues for Crisis (2017), 38(1), 44–52

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

prevention and treatment (Han et al., 2009; Himelboim & Han, 2014). Acknowledgments This work was supported by the National Institutes of Health, grant numbers R01DA032843 (PCR), R01DA 039455 (PCR). One of the authors, Dr. Bierut, is listed as an inventor on Issued U.S. Patent 8, 080, 371, “Markers for Addiction,” covering the use of certain SNPs in determining the diagnosis, prognosis, and treatment of addiction. All other authors declare they have no conflicts of interest.

References American Foundation for Suicide Prevention. (2015). Worried about someone. Retrieved from https://www.afsp.org/find-support/ worried-about-someone Andover, M. S., Pepper, C. M., Ryabchenko, K. A., Orrico, E. G., & Gibb, B. E. (2005). Self-mutilation and symptoms of depression, anxiety, and borderline personality disorder. Suicide and Life-Threatening Behavior, 35(5), 581–591. doi:10.1521/suli.2005.35.5.581 Attenberg, J. (2012, June 30). Network: How to use Tumblr to connect with readers. Retrieved from http://www.pw.org/content/ network_how_to_use_tumblr_to_connect_with_readers_0?cmnt_all=1 Baker, D., & Fortune, S. (2008). Understanding self-harm and suicide websites. Crisis, 29, 118–122. doi:10.1027/0227-5910.29.3.118 Barak, A., & Dolev-Cohen, M. (2006). Does activity level in online support groups for distressed adolescents determine emotional relief. Counselling and Psychotherapy Research, 6, 186–190. doi:10.1080/14733140600848203 Becker, K., Mayer, M., Nagenborg, M., El-Faddagh, M., & Schmidt, M. H. (2004). Parasuicide online: Can suicide websites trigger suicidal behaviour in predisposed adolescents? Nordic Journal of Psychiatry, 58, 111–114. doi:10.1080/08039480410005602 Boyd, D. (2014). It’s complicated: The social lives of networked teens. New Haven, CT: Yale University Press. Brenner, J. (2014). Pew Internet: Social networking. Retrieved from http://www.pewinternet.org/fact-sheets/social-networking-fact-sheet/ Cuen, L. (2015). Tumblr is becoming a resource for teen suicide prevention. Retrieved from http://www.vocativ.com/news/207970/ tumblr-suicide/ Duggan, J. M., Heath, N., Lewis, S. P., & Baxter, A. L. (2012). An examination of the scope and nature of non-suicidal self-injury online activities: Implications for school mental health professionals. School Mental Health, 4, 56–67. doi:10.1007/s12310011-9065-6 Han, J. Y., Hawkins, R. P., Shaw, B. R., Pingree, S., McTavish, F., & Gustafson, D. H. (2009). Unraveling uses and effects of an interactive health communication system. Journal of Business Economics and Management, 53, 112–133. doi:10.1080/ 08838150802643787 Hawton, K., Hall, S., Simkin, S., Bale, L., Bond, A., Codd, S., & Stewart, A. (2003). Deliberate self-harm in adolescents: A study of characteristics and trends in Oxford, 1990–2000. Journal of Child Psychology and Psychiatry, 44(8), 1191–1198. doi:10.1111/ 1469-7610.00200 © 2016 Hogrefe Publishing

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

Hawton, K., Rodham, K., Evans, E., & Weatherall, R. (2002). Deliberate self harm in adolescents: Self report survey in schools in England. BMJ, 325, 1207–1211. Hawton, K., Saunders, K. E. A., & O’Connor, R. C. (2012). Self-harm and suicide in adolescents. The Lancet, 379(9834), 2373–2382. doi:10.1016/S0140-6736(12)60322-5 Himelboim, I., & Han, J. Y. (2014). Cancer talk on Twitter: Community structure and information sources in breast and prostate cancer social networks. Journal of Health Communication, 19, 210–225. doi:10.1080/10810730.2013.811321 Laye-Gindhu, A., & Schonert-Reichl, K. (2005). Nonsuicidal selfharm among community adolescents: Understanding the “whats” and “whys” of self-harm. Journal of Youth and Adolescence, 34, 447–457. doi:10.1007/s10964-005-7262-z Lenhart, A., & Madden, M. (2007, January 7). Social networking websites and teens: An overview. Retrieved from http:// www.pewinternet.org/2007/01/07/social-networking-websites-and-teens/ Lewis, S. P., Heath, N. L., St. Denis, J. M., & Noble, R. (2011). The scope of nonsuicidal self-injury on YouTube. Pediatrics, 127, e552–e557. doi:1542/peds.2010-2317 Lewis, S. P., Heath, N. L., Sornberger, M. J., & Arbuthnott, A. E. (2012). Helpful or harmful? An examination of viewers’ responses to nonsuicidal self-injury videos on YouTube. Journal of Adolescent Health, 51(4), 380–385. doi:10.1016/j.jadohealth.2012.01.013 Lloyd-Richardson, E. E. (2007). Characteristics and functions of non-suicidal self-injury in a community sample of adolescents. Psychological Medicine, 37, 1183. doi:10.1017/S003329 170700027X Mander, J. (2014). GWI social summary Q3 2014. Retrieved from http://www.globalwebindex.net/blog/tumblr-instagram-audiences Marwick, A. E., & Boyd, D. (2011). I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience. New Media & Society, 13(1), 114–133. doi:10.1177/ 1461444810365313 Mitchell, K. J., & Ybarra, M. L. (2007). Online behavior of youth who engage in self-harm provides clues for preventive intervention. Preventive Medicine, 45, 392–396. doi:10.1016/j. ypmed.2007.05.008 Moran, P., Coffey, C., Romaniuk, H., Olsson, C., Borschmann, R., Carlin, J. B., & Patton, G. C. (2012). The natural history of self-harm from adolescence to young adulthood: A population-based cohort study. The Lancet, 379(9812), 236–243. doi:10.1016/ S0140-6736(11)61141-0 Moreno, M. A., Christakis, D. A., Egan, K. G., Jelenchick, L. A., Cox, E., Young, H., … Becker, T. (2012). A pilot evaluation of associations between displayed depression references on Facebook and self-reported depression using a clinical scale. The Jounal of Behavioral Health Services & Research, 39, 295–304. doi:10.1007/s11414-011-9258-7 Moreno, M. A., Jelenchick, L. A., Egan, K. G., Cox, E., Young, H., Gannon, K. E., & Becker, T. (2011). Feeling bad on Facebook: Depression disclosures by college students on a social networking site. Depression and Anxiety, 28, 447–455. doi:10.1002/da.20805 Naslund, J. A., Grande, S. W., Aschbrenner, K. A., & Elwyn, G. (2014). Naturally occurring peer support through social media: The experiences of individuals with severe mental illness using YouTube. PLOS one, 9(10), e110171. doi:10.1371/journal. pone.0110171 Park, M., Cha, C., & Cha, M. (2012). Depressive moods of users portrayed in Twitter. Proceedings of the ACM SIGKDD Workshop on Healthcare Informatics (HI-KDD) 2012, 1–8. Park, S., Lee, S. W., Kwak, J., Cha, M., & Jeong, B. (2013). Activities on Facebook reveal the depressive state of users. Journal of Medical Internet Research, 15, e217. doi:10.2196/jmir.2718 © 2016 Hogrefe Publishing

Perrin, A. (2015). Social media usage: 2005–2015. Washington, DC: Pew Internet & American Life Project. Retrieved from http:// www.pewinternet.org/2015/10/08/social-networking-usage2005-2015/ Robinson, J., Cox, G., Bailey, E., Hetrick, S., Rodrigues, M., Fisher, S., & Herrman, H. (2015). Social media and suicide prevention: A systematic review. Early Intervention in Psychiatry. Advance online publication. doi:10.1111/eip.12229 Robinson, J., Rodrigues, M., Fisher, S., & Herrman, H. (2014). Suicide and social media: Findings from the literature review. Melbourne, Australia:Young and Well Cooperative Research Centre.Retrieved from http://www.youngandwellcrc.org.au/wp-content/uploads/ 2014 / 07/Robinson_2014_Suicide-and-Social-Media_ Findings-from-the-Literature-Review.pdf Ryan, G. W., & Bernard, H. R. (2003). Techniques to identify themes. Field Methods, 15(1), 85–109. doi:10.1177/1525822X02239569 Tufekci, Z. (2008). Can you see me now? Audience and disclosure regulation in online social network sites. Bulletin of Science, Technology & Society, 28(1), 20–36. doi:10.1177/0270467607311484 Tumblr. (2015). Community guidelines. Retrieved from https://www. tumblr.com/policy/en/community Whitlock, J. L., Powers, J. L., & Eckenrode, J. (2006). The virtual cutting edge: The internet and adolescent self-injury. Developmental Psychology, 42, 407–417. doi:10.1037/0012-1649.42.3.407 Ybarra, M. L., Alexander, C., & Mitchell, K. J. (2005). Depressive symptomatology, youth Internet use, and online interactions: A national survey. Journal of Adolescent Health, 36(1), 9–18. doi:10.1016/j.jadohealth.2003.10.012 Zain, A. L. (2013). Will Tumblr bring a younger audience to Yahoo! sites? Retrieved from http://www.comscore.com/Insights/ Data-Mine/will-tumblr-bring-a-younger-audience-to-yahoosites

Received September 3, 2015 Revision received February 25, 2016 Accepted February 25, 2016 Published online July 22, 2016

About the authors Patricia A. Cavazos-Rehg, PhD, is Associate Professor of Psychiatry at the Washington University in St. Louis School of Medicine, MO, as well as a clinically trained licensed psychologist. She specializes in mental health epidemiology, policy, and social media research. Melissa J. Krauss, MPH, is Research Statistician in the Department of Psychiatry at the Washington University in St. Louis School of Medicine, MO, USA. She has 15 years of experience analyzing and interpreting data from public health studies on substance use and mental health. Shaina J. Sowles, MPH, is Clinical Research Coordinator in the Department of Psychiatry at the Washington University in St. Louis School of Medicine, MO, USA. She is experienced in managing studies related to tobacco, marijuana, and mental health. Sarah Connolly is an undergraduate at Truman State University, Kirksville, MO, studying chemistry. Carlos Rosas is a PhD candidate at the University of Illinois at Chicago, IL, USA, obtaining a degree in psychology. Meghana Bharadwaj is a senior at Lafayette High School, Wildwood, MO. She plans to study political science and psychology. Crisis (2017), 38(1), 44–52

Richard Grucza, PhD, is Associate Professor of Psychiatry at the Washington University in St. Louis School of Medicine, MO, USA. His research is focused on policy influences on substance abuse, mental health, and health behaviors. Laura J. Bierut, MD, is a physician scientist and Alumni Endowed Professor of Psychiatry at the Washington University in St. Louis School of Medicine, MO, USA. She has significant experience in genetic studies of smoking behaviors, addiction, and other psychiatric and medical illnesses.

Crisis (2017), 38(1), 44â&#x20AC;&#x201C;52

P. A. Cavazos-Rehg et al.: Depression-Related Content on Tumblr

Patricia A. Cavazos-Rehg Department of Psychiatry Washington University School of Medicine 660 South Euclid Avenue, Box 8134 St. Louis, MO 63110 USA Tel. +1 314 362-2152 Fax +1 314 362-4247 E-mail rehgp@psychiatry.wustl.edu

ÂŠ 2016 Hogrefe Publishing

Research Trends

Promoting Help Seeking to Veterans A Comparison of Public Messaging Strategies to Enhance the Use of the Veterans Crisis Line Elizabeth Karras1,2, Naiji Lu3, Heather Elder1,4, Xin Tu2,3, Caitlin Thompson5, Wendy Tenhula6, Sonja V. Batten6,7, and Robert M. Bossarte1,5,8,9 VISN 2 Center of Excellence for Suicide Prevention, Department of Veterans Affairs, Canandaigua VA Medical Center, Canandaigua, NY, USA Department of Psychiatry, University of Rochester, NY, USA 3 Department of Biostatistics and Computational Biology, University of Rochester, NY, USA 4 Department of Public Health Sciences, University of Rochester, NY, USA 5 Suicide Prevention Office, Department of Veterans Affairs, Washington, DC, USA 6 Mental Health Services, Department of Veterans Affairs, Washington, DC, USA 7 Booz Allen Hamilton, Washington, DC, USA 8 Injury Control Research Center, West Virginia University, Morgantown, WV, USA 1 2

Department of Behavioral Medicine, West Virginia University, Morgantown, WV, USA Abstract. Background: Few studies have considered different messaging strategies that may augment campaign efficacy to generate help-seeking behaviors among populations at increased risk for suicide, mainly US military veterans. Aims: Findings are presented from the pilot evaluation of the It’s Your Call campaign implemented by the Department of Veterans Affairs (VA). Three messaging strategies (with varying intensity and mix of messages) were compared to explore which best promote use of the Veterans Crisis Line (VCL) among veteran populations. Method: Daily VCL call data were obtained for 10 US cities during 2011–2012 where the campaign was active, and modeled using Poisson regression to identify changes in utilization patterns associated with the implementation of different messaging strategies. Results: Significant increases in call rates were only evident during the campaign in communities where mixed messages were disseminated. Further, use of mixed messages yielded greater increases in call rates when compared with the other tested strategies. This was an observational study where identification of causal relationships between variables was limited. Conclusion: Findings are encouraging as messaging was associated with help seeking, and they provide insights into strategies that may rapidly promote crisis line use. Results also underscore the need for further research on suicide prevention campaigns and dissemination practices. Keywords: campaigns, veterans, military, crisis line, public messaging

Communication campaigns play an important role in shaping public perceptions of health issues and motivating behavior change across large populations (Noar, 2006; Snyder et al., 2004; Wakefield, Loken, & Hornik, 2010). There is compelling evidence that these efforts are an effective strategy to promote public health (Noar, 2006; Snyder et al., 2004; Wakefield et al., 2010; Hornik, 2002b; Huhman et al., 2007) and have recently drawn attention in expert panel reports for their promise as interventions for suicide prevention (US Department of Health and Human Services [HHS], 2012). Suicide is the leading cause of injury-related death in the US (Rockett et al., 2012), and claimed more than 42,000 American lives in 2014 (US National Center for Health Statistics, 2016). While suicide represents a © 2016 Hogrefe Publishing

major public health problem it is also largely preventable, underscoring a critical opportunity for campaigns to help to address risk and facilitate timely help seeking. Research on the effects of suicide prevention campaigns is still limited (Chambers et al., 2005); however, several studies have demonstrated their ability to improve antecedents to help seeking for at-risk populations. For example, significant associations were identified between public messaging and an increased awareness and likelihood to use crisis lines among US veterans (Bossarte et al., 2014; Karras, Stephens, Kemp, & Bossarte, 2014). Crisis lines are publicly available resources that play an important role in mitigating suicidal behaviors and connecting individuals to local treatment and support (Wilson, Bushnell, Crisis (2017), 38(1), 53–62 DOI: 10.1027/0227-5910/a000418

& Caputi, 2011). Their effective promotion may be especially beneficial to veterans, who account for approximately 20% of US suicide deaths each year (Kemp & Bossarte, 2013) and possess multiple risk factors for poor mental health outcomes (Brancu, Straits-Tröster, & Kudler, 2011; Zivin et al., 2007). While little data have been published on behavioral changes associated with suicide prevention campaigns, there is some preliminary evidence to suggest that these efforts may generate crisis line use (Karras et al., 2016; Till, Sonneck, Baldauf, Steiner, & Niederkrotenthaler, 2013). As such, further investigation is needed to identify messaging strategies that effectively promote this type of help seeking to veteran populations. In order for campaigns to encourage such behavior it is critical that messages reach their intended audience. However, effective health promotion can be difficult in today’s media markets as consumers are inundated with pervasive messaging from numerous competing sources (Noar, 2006; Randolph & Viswanath, 2004). As a result, campaign designers often adopt widespread message dissemination to increase the likelihood of exposure and desired outcomes (high-dose messaging; Hornik, 2002). This strategy has proven moderately successful, as high-intensity campaigns have been associated with reductions in risk behaviors (e.g., Hersey et al., 2005; Farrelly et al., 2005) and engagement in healthy practices (e.g., Snyder et al., 2004). However, messages are subject to wear with repetition, which can result in diminished or adverse effects (Harris, Pierce, & Bargh, 2014; Koch & Zerback, 2013; LaVail, Anker, Reinhart, & Feeley, 2010). Alternatively, there is growing evidence that delivering a strategic mix of messages may produce greater results as the use of various appeals and behavioral incentives may help to avoid the potential pitfalls of overexposure (mixed messaging; Hornik, 2002b; Ma, Dollar, & Kibler, 2011; Scheufele & Tewksbury, 2007). While the development of safe messages has been considered when designing suicide prevention campaigns (Chambers et al., 2005; Daigle et al., 2006; Langford, Litts, & Pearson, 2013), no study to date has explicitly examined dissemination strategies (i.e., intensity or mix of messages) that may augment their efficacy to endorse help seeking. We seek to address this research gap by examining the use of several messaging strategies that may enhance help-seeking behaviors (crisis line use). This study presents findings from the pilot evaluation of the It’s Your Call campaign implemented by the Department of Veterans Affairs (VA). The It’s Your Call campaign was launched in 2011 with the primary goal of increasing awareness and use of the newly rebranded Veterans Crisis Line (VCL) broadly among US veterans (formerly the National Veterans Suicide Prevention Hotline). Several messaging strategies were implemented as part of this pilot work (described further in the methods), and participatCrisis (2017), 38(1), 53–62

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

ing communities were assigned to one of three exposure groups: (a) a low campaign dose of It’s Your Call messages; (b) high doses of It’s Your Call advertising; or (c) mixed messages where both high doses of It’s Your Call messaging and the VA’s Make the Connection public awareness campaign were disseminated. The Make the Connection campaign was also implemented in 2011 by VA to provide veterans and their families with information on mental health issues and treatment as well as to connect them to available resources and support. The primary aims of this study were twofold. First, we sought to identify changes in VCL use associated with the implementation of each messaging strategy during the It’s Your Call campaign (Aim 1). To do so, we modeled average daily call volume within each exposure group and compared calls during and after the campaign with those received before the campaign (reference category). Interaction terms were included in all study models a priori to examine changes in call volume over time (i.e., rates). Rates were then used in subsequent analyses for Aim 2 to allow for direct comparison of messaging strategies and their impact on VCL use. Based on prior research cited earlier on the promotion of crisis lines, we hypothesized significant increases in daily calls and rates (i.e., interaction between time and daily calls received during the campaign) would be found during the campaign for all of the messaging strategies when compared with precampaign call volume. Next, we examined whether enhanced (additional) public messaging is associated with increased VCL use during the campaign (Aim 2). More specifically, we modeled daily call rates for each exposure group and compared the high-dose and mixed-message strategies with the use of low-dose marketing (reference category) during the campaign. The extant literature suggests that higher amounts of messaging are associated with greater outcomes (e.g., Hersey et al., 2005). As such, we hypothesized that higher call rates and significant interactions (i.e., time and daily call rates during the campaign) would be found during the campaign for both high-dose and mixed-message strategies when compared with the low-dose group.

Method The primary outcome in this study is call volume to the VCL, which is a toll-free confidential service that provides care to veterans and active duty military personnel in crisis and connects their families and friends to resources. The VCL shares a common phone number with the National Suicide Prevention Lifeline (1-800-273-8255). Callers are prompted to “press 1” at the beginning of their call to be routed to this specialized service. © 2016 Hogrefe Publishing

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

Selection of Pilot Communities Table 1 lists the 10 US cities included in this pilot evaluation. Sites were selected based on demographic composition (e.g., estimated veteran population), geographic location, opportunities for promotion (available media outlets), and the capacity for local VA facilities to respond to an increased need that may result from the public messaging. Communities were then randomly assigned to exposure groups (listed in previous section) that received varying doses of VA campaigns to test the impact of different messaging strategies on crisis line use. The characteristics of study sites were also examined and compared to identify potential baseline differences. In the year preceding the study period (2010) the overall VCL call rate for all pilot cities was 4.81/1,000 veterans (Table 1). When rates were comparable across exposure groups (low dose: 4.62/1,000; mixed message: 4.78/1,000) they were slightly higher in areas that were selected to receive high doses of the It’s Your Call campaign (5.08/1,000). When community demographics were compared across exposure groups, several significant differences were found among residents including: veteran status (low dose: 7.97% vs. high dose: 11.07 vs. mixed message: 5.95, p < .001); race (White; low dose: 49.45% vs. high dose: 77.96 vs. mixed message: 67.66, p < .001); marital status (married; low dose: 30.22% vs. high dose: 40.18 vs. mixed

Table 1. Call rates to the Veterans Crisis Line for pilot cities in the year preceding the It’s Your Call campaign: US, 2010 Cities

Call ratesa

Overall (all pilot cities)

4.81

Cities with low-dose It’s Your Call campaign

4.62

Sacramento (CA)

4.82

Fargo (ND)

3.12

Fayetteville (NC)

5.78

St. Louis (MO)

4.44

Cities with high-dose It’s Your Call campaign

5.08

Virginia Beach (VA)

5.09

Boise (ID)

3.69

Sioux Falls (SD)

6.47

Johnson City (TN)

5.57

Mixed-message markets (It’s Your Call and Make the Connection campaigns)

4.78

Seattle (WA)

4.79

Tampa (FL)

4.75

Note. a Call rates are per 1,000 veteran population during the calendar year. Population estimates are from the US Census Bureau, 2010 American Community Survey.

message: 32.83, p < .001); age (65 and older; low dose: 10.42% vs. high dose: 10.96 vs. mixed message: 10.76, p < .001); and unemployment (low dose: 9.86% vs. high dose: 5.02 vs. mixed message: 7.11, p < .001; US Census Bureau, 2010a). Finally, household Internet access for study regions was comparable across exposure groups (low dose: 77.16% vs. high dose: 74.09 vs. mixed message: 78.44, p < .001; US Census Bureau, 2010b). While not perfectly balanced, the composition of study sites suggests that important similarities across exposure groups exist (e.g., call rates, prevalence of veterans, unemployment, age, Internet access).

Campaign Materials This evaluation focused on outcomes associated with It’s Your Call campaign; however, the VA’s Make the Connection campaign was also disseminated in select cities to assess for the potential impact of the use of various messages on calls to the crisis line. Both campaigns targeted the broad veteran population with efforts that promote mental health-related help seeking and messages used are described here. It’s Your Call Campaign The primary goal of this campaign was to promote awareness and use of the VCL to the broad veteran population. Messages displayed images of veterans and their loved ones as well as the campaign slogan, “It’s Your Call,” and used specific language to target the intended audience (“confidential help for veterans…”). Instructions for accessing the VCL were included, and the toll-free number (1-800-273-8255, “press 1”) was prominently displayed on campaign materials. Strategic design efforts such as patriotic imagery (e.g., American flag; the use of the colors red, white, and blue) were also employed to increase the saliency of VCL messages to veteran populations. It’s Your Call campaign materials are accessible on the VCL website (http://www.veteranscrisisline.net). Make the Connection Campaign A central component of this campaign is its website (http:// www.maketheconnection.net) where individuals are introduced to narrative videos of personal stories from real veterans and their families recognizing mental health issues, overcoming challenges, and succeeding with treatment and recovery. The public messages utilized in the current study featured quotes by veterans that encouraged others to seek help (“I’m a veteran, I know what it’s like... There’s a whole community of veterans out there who just want to help”) and promoted the website by providing the link. While these messages did not directly advertise the VCL, Crisis (2017), 38(1), 53–62

each page of the Make the Connection website provides information on how to access it.

Messaging Strategies As mentioned earlier, study sites were randomly assigned to one of three exposure groups that differed based on the type and/or amount of messaging received, and included: (a) low-dose It’s Your Call campaign; (b) high-dose It’s Your Call campaign; and (c) mixed messages, with dissemination of both It’s Your Call and Make the Connection campaigns. For the purpose of this pilot study, messages were disseminated for 16 weeks in participating cities (September 1, 2011, to December 31, 2011). Low-Dose It’s Your Call Campaign The campaign consisted of low-intensity messaging efforts implemented online. Advertisements (i.e., banner ads) were placed on websites that target veteran populations and their lifestyle or interests as well as disseminated through popular social media. Relevant keyword search queries were also targeted to reach help seekers with campaign messages, and positioned the VCL website as a readily available resource for immediate care. While explicitly assessed in four study sites, this low-dose strategy was also implemented nationally. High-Dose It’s Your Call Campaign This campaign also received the online strategy described in the previous section; however, a series of additional media were purchased and disseminated in these markets to assess the effects of augmented or high-intensity messaging. This included campaign roadside billboards positioned along popular driving routes in each city, public transportation ads (e.g., placed on mass transit and inside bus shelters), print advertisements in local newspapers, and the radio broadcast of the campaign’s public service announcement (PSA), which also played in local movie theaters during previews. This brief PSA featured a veteran facing personal struggles, making the decision to call the crisis line for help, and provided VCL contact information. Media buys were tailored to broadly reach the veteran population through popular outlets. Mixed-Message Markets This included the distribution of Make the Connection campaign messages in several cities along with high doses of It’s Your Call (described in the previous section) to examine the impact of diversified messaging efforts on VCL use. Similar to the strategy employed for the high-dose campaign, print Make the Connection media were purchased including roadside billboards, public transportaCrisis (2017), 38(1), 53–62

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

tion ads, local newspaper advertisements, as well as online placements promoting the use of the website. The campaign’s PSA, which featured a short compilation of veteran stories of seeking help and advertised the website, was also broadcast through radio outlets and during previews in local movie theaters.

Data Analysis De-identified daily VCL call data were obtained from the VA for area codes within the 10 pilot cities for the year surrounding the campaign period (May 1, 2011, to April 30, 2012), and were grouped according to exposure (those assigned to low-dose It’s Your Call campaign vs. high-dose vs. mixed-message markets). Several steps were then taken to identify changes in VCL call volume associated with the implementation of the different messaging strategies. First, the average weekly calls to the VCL were calculated and the 8-week moving average was plotted for the study period for each exposure group. These plots allow for the identification of trends in calls before, during, and after the campaign, which could be used to explain potential differences that are observed. Next, the average daily number of VCL calls was tabulated for each exposure group, and grouped into three equal time periods: (a) precampaign, 5/1/2011–8/31/2011; (b) during the campaign, 9/1/2011–12/31/2011; and (c) postcampaign, 1/1/2012–4/30/2012. Data were then modeled with Poisson regression using a generalized estimating equation (GEE; Kowalski & Tu, 2007) approach that also provided robust estimates of standard errors. Interaction terms were included in each model a priori to assess change in call volume over time as indicated in the aims section. The working independence correlation model was used for all analyses, which were conducted in SAS version 9.3 (Cary, NC). Estimates of associations between messaging strategies (i.e., exposure group) and changes in call volume are detailed in the results section. Data on the characteristics of callers were not available to the research team and therefore not included in the models.

Results Trends in VCL Calls Figure 1 displays plots of the 8-week moving average for calls to the VCL from May 2011 through April 2012. A gradual increase in calls over time is evident for all exposure categories. Trends also suggest that weekly calls were generally lower for those areas that received the high-dose © 2016 Hogrefe Publishing

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

Figure 1. Eight-week moving average of calls to the Veterans Crisis Line (VCL) for study exposure groups: US, May 1, 2011, to April 30, 2012

It’s Your Call campaign than for the other exposure categories. Spikes in call volume are most evident in mixed-message markets during the campaign period.

Aim 1: Changes in VCL Use Associated With Each Messaging Strategy Table 2 summarizes the results from regression analyses of the daily call volume to the VCL for each exposure group. We first sought to identify changes in the average number of daily calls during the time periods assessed in the study (pre-, during, and postcampaign). When compared with the precampaign period (reference category), VCL calls were higher during and following (post) the campaign for all three exposure groups. While significant increases in the number of daily calls were identified, this may reflect the overall upward trend in call volume to the VCL reported in the plots. Next, rates (slope) were examined to measure changes in calls over the study period. Interactions of (a) time and (b) average number of daily calls made during each time period (pre-, during, and postcampaign) were examined to identify significant changes in the daily call rates to the VCL for each exposure group. Call rates significantly decreased during the campaign for the high-dose It’s Your Call exposure group (–0.003, SE = 0.001, p = .01) and slightly increased after the campaign (0.002, SE = 0.001, p = .05). Conversely, a small yet statistically significant increase in daily call rates was observed during the campaign in the mixed-message markets (0.003, SE = 0.001, p = .001), which was followed by a decrease after the campaign (–0.002, SE = 0.001, p = .02). No statistically significant interactions were found for the © 2016 Hogrefe Publishing

low-dose group. These findings partially support our hypothesis, as increased daily calls were observed during the campaign for each messaging strategy; however, greater call rates were only identified for mixed-message markets.

Aim 2: Comparisons of Messaging Strategies and Associated VCL Use Additional regression analyses were conducted to compare the use of different messaging strategies, and examine whether enhanced public messaging was associated with increased VCL use. Analyses for Aim 2 were restricted to VCL calls received during the campaign (9/1/2011– 12/31/2011). Findings in Table 3 reveals partial support for our hypothesis, and show that significantly higher daily call rates were found only in the mixed-message markets (0.77, SE = 0.23, p = .001) as compared with those received in the low-dose It’s Your Call campaign (reference category). Interactions were also included in the model to assess changes over the study period, and examined (a) time (over the campaign) and (b) daily call rates for exposure groups. Results revealed a significant interaction for only the mixed-message markets (0.002, SE = 0.001, p = .002) indicating greater increases in call rates in these areas over the campaign period when compared with the low-dose exposure group. No differences in daily call rates or significant interactions were identified for the high-dose It’s Your Call campaign when compared with the low-dose strategy.

Crisis (2017), 38(1), 53–62

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

Table 2. Regression analyses for average daily calls and daily call rates to the Veterans Crisis Line (VCL) before, during, and after the It’s Your Call campaign for study exposure groups: US, May 1, 2011, to April 30, 2012 Cities with low-dose It’s Your Call campaignc,e Coefficient

Mixed-message marketsc,d,g (It’s Your Call and Make the Connection campaigns)

Cities with high-dose It’s Your Call campaignc,f

SEb

Coefficient

SEb

Coefficient

SEb

0.77

0.03

< .0001

0.77

0.03

< .0001

0.99

0.02

< .0001

Calls during the campaigna

0.13

0.03

.0002

0.10

0.04

.007

0.09

0.03

.009

Calls postcampaign

0.17

0.03

< .0001

0.24

0.04

< .0001

0.08

0.03

.02

0.0003

0.001

.69

0.003

0.001

.0002

0.001

.06

0.001

.40

–0.003

0.001

.01

0.003

0.001

.001

0.0003

0.001

.76

0.002

0.001

.05

–0.002

0.001

.02

Intercept

Timea Time*During the campaign

Time*Postcampaigna

Note. aThe reference category for analyses was calls during the pre-campaign period. bSE = Standard Error. cIt’s Your Call campaign = A suicide prevention campaign to promote awareness and use of the Veterans Crisis Line that was implemented by the Department of Veterans Affairs and assessed d in pilot cities from 9/1/2011 to 12/31/2011. Make the Connection campaign = A multimedia mental health promotion campaign implemented by the Department of Veterans Affairs to provide veterans and their families with information on mental health issues and treatment, and connect them to e resources and support. This campaign was assessed in select pilot cities from 9/1/2011 to 12/31/2011. Low-dose It’s Your Call Campaign cities include f Sacramento (CA), Fargo (ND), Fayetteville (NC), St. Louis (MO). High-dose It’s Your Call Campaign cities include Virginia Beach (VA), Boise (ID), Sioux Falls g (SD), Johnson City (TN). Mixed-message markets with both It’s Your Call and Make the Connection messages include Seattle (WA), Tampa (FL). Statistically significant values are bold.

Table 3. Regression analyses comparing daily call rates to the Veterans Crisis Line during the VA It’s Your Call campaign for study exposure groups: US, September 1, 2011, to December 31, 2011 Coefficient

SEb

Intercept

0.79

0.17

< .0001

Cities with high-dose It’s Your Call campaigna,c,f

0.41

0.24

.09

Mixed-message markets (It’s Your Call and Make the Connection campaigns)

0.77

0.23

.001

Timea

0.001

0.0007

.10

–0.001

0.001

.212

0.002

0.001

.002

a,c,d,g

a,c,f

Time*High-Dose It’s You Call Campaign Time*Mixed-Message Marketsa,c,d,g

Note. aThe reference category for analyses was call rates in cities with the low-dose It’s Your Call campaign. bSE = Standard Error cIt’s Your Call campaign = A suicide prevention campaign to promote awareness and use of the Veterans Crisis Line that was implemented by the Department of Veterans Affairs d and assessed in pilot cities from 9/1/2011 to 12/31/2011. Make the Connection campaign = A multimedia mental health promotion campaign implemented by the Department of Veterans Affairs to provide veterans and their families with information on mental health issues and treatment, and cone nect them to resources and support. This campaign was assessed in select pilot cities from 9/1/2011 to 12/31/2011. Low-dose It’s Your Call campaign f cities include Sacramento (CA), Fargo (ND), Fayetteville (NC), St. Louis (MO). High-dose It’s Your Call campaign cities include Virginia Beach (VA), Boise (ID), g Sioux Falls (SD), Johnson City (TN). Mixed-message markets with both It’s Your Call and Make the Connection messages include Seattle (WA), Tampa (FL). Statistically significant values are bold.

Post Hoc Analyses Given the limited research conducted on the use of mixed messages to promote crisis support services, a post hoc sensitivity analysis was performed to assess whether the findings from Aim 2 would hold true when the daily call rates for the mixed-message markets were directly compared with those for the high-dose It’s Your Call campaign areas. The same modeling approach from Aim 2 was used, and daily call rates during the campaign were compared between the mixed-message group and the high-dose group (reference category). Results were similar to those reported in Aim 2, as significantly greater rates were found for mixed-message markets versus the high-dose group Crisis (2017), 38(1), 53–62

during the campaign (not shown). Findings lend some support for the use of a combination of public messages to facilitate health behavior change, and suggest this strategy may be especially valuable for promoting crisis line use to veteran populations.

Discussion Several messaging strategies were piloted as part of the VA’s It’s Your Call Campaign, and compared to explore their potential to promote crisis line (VCL) use among veteran populations. Findings are encouraging as campaigns © 2016 Hogrefe Publishing

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

were associated with some detectable increases in call volume, and extend prior research to provide initial evidence for the use of multiple branded messages for suicide prevention. However, audience responses to campaigns are multifaceted and results underscore the need for further investigation to identify optimal message variation and dose that best facilitate help seeking. Increases in call rates were only evident during the campaign in areas where both the It’s Your Call and Make the Connection messages were disseminated. These findings are consistent with a body of research that suggests the strategic use of multiple messages may effectively generate awareness of available resources while potentially avoiding unintended negative effects (e.g., message burnout; Hornik, 2002a, 2002b; Ma et al., 2011). The distribution of several types of campaigns may have primed veterans to be more receptive to related subsequent VA messages including information on the VCL. Further, the use of both suicide prevention (It’s Your Call) and mental health promotion (Make the Connection) messages together may have communicated help seeking as viable and acceptable to veterans, and provided an available tool (VCL) to engage in behavior change. While this messaging strategy is novel to suicide prevention, there is evidence to support its use with veteran populations. Recent research has shown that exposure to various public communication efforts is positively associated with an increased likelihood to utilize telehealth services for a range of health issues among veteran households (Elder, Karras, & Bossarte, in press). Such results should be considered when designing suicide prevention initiatives as the dissemination of a combination of campaigns may help to develop larger environments that reinforce crisis line use (help seeking) as a normative behavior. The use of multiple campaigns also yielded greater results when compared with the other messaging strategies examined in this study. More specifically, call rates were higher (and increased) during the campaign for areas with both Make the Connection and It’s Your Call messages as compared with communities that received the high or low doses of only It’s Your Call. The short-term increases in crisis line use found in mixed-message markets (vs. for other strategies) may suggest that a diversified set of messages is a valuable public communication tactic to rapidly promote help seeking among veteran populations. Analyses conducted for Aim 1 of call data following the campaign lend support to this as calls significantly decreased in these areas once messaging stopped. These findings may have important implications for prevention efforts as this public messaging strategy could potentially be used to target communities with increased suicide risk (e.g., concerns over areas with potential for contagion) to quickly diffuse information on available resources and motivate help seeking. © 2016 Hogrefe Publishing

Despite prior research suggesting that high-intensity messaging may improve health behaviors (Berkowitz, Huhman, & Nolin, 2008; Niederdeppe, Farrelly, & Haviland, 2004; Southwell, Barmada, Hornik, & Maklan, 2002), augmented doses of It’s Your Call were not associated with increased call rates to the VCL during the campaign. Instead, small yet significant decreases in daily call rates were reported in Aim 1 for pilot communities that received high doses. Further, no significant differences were found when call rates were compared with those for low-dose messaging areas during the campaign. Taken together, no additive effect was produced by the use of supplemental VCL advertising on calls. While underlying factors such as the demographic composition of communities assigned to this group may have contributed to these outcomes, another possible explanation is that the It’s Your Call campaign experienced issues with market saturation in pilot cities (e.g., message overexposure vs. insufficient message diffusion), which is a common concern for public communication campaigns (Hornik, 2002b). As such, measures of direct message exposure (including frequency) should be integrated into the next steps of this research to further explore the impact of campaign saturation on related behavior change. Alternatively, while shifts in behavioral outcomes were not identified during the campaign for this strategy, increased It’s Your Call messaging may have generated changes in precursors to behavior not measured in the current evaluation. For example, improvements in awareness and attitudinal measures have been associated with the use of higher levels of campaign messaging in other studies (Berkowitz et al., 2008; McVey & Stapleton, 2000; Niederdeppe et al., 2004; Snyder et al., 2004; Southwell et al., 2002). Such changes in behavioral determinants may strengthen intentions for future help seeking including the use of crisis support services. There may be some evidence to support this in our data as small significant increases in call rates were reported following the campaign. Conversely, campaign messages simply may not have resonated with the audience and consequently did not impact crisis line use. Additional research investigating perceptions of messages and campaign effects on proximal cognitive outcomes is warranted to assess their potential to drive help-seeking behavior over time (e.g., short- vs. long-term changes). When examining patterns of crisis line use, low-dose It’s Your Call efforts did not increase call rates in pilot cities during the campaign either. These results may reflect the implementation of this messaging strategy, which consisted solely of online advertisements. Again, messages may not have reached or were not salient to the intended audiences. However, when considering the ways in which individuals may seek information, those already online may Crisis (2017), 38(1), 53–62

be likely to continue to utilize the same channel to obtain help (e.g., Chisolm, 2010; Dutta-Bergman, 2004; Karras & Rintamaki, 2012). Therefore, messaging may still have generated increased awareness of the VCL, but potentially encouraged complementary (online) help-seeking behaviors. These findings underscore the need for research that explores the impact of Internet-based outreach on service use online (e.g., VCL confidential chat; Make the Connection website).

Limitations and Next Steps There are several limitations to this research that should be noted. While these analyses provide important initial data on the use of public messaging strategies to promote help seeking, this was an observational study where direct measures of campaign exposure and crisis line use were not collected. As such, the identification of causal relationships between variables was not permitted. Additionally, changes in VCL call volume may be ascribed to community differences or other unobserved factors such as exposure to unrelated local or regional efforts to promote help seeking or crisis line use, particularly surrounding September (suicide prevention month). Yet, it is unlikely that promotional activities were disseminated consistently across cities over the study period. Future research should tease out campaign effects from those produced by other community intervention or outreach. A number of baseline differences were also identified between exposure groups that were not accounted for in our models, and may have attributed to observed changes in VCL use. The demographic composition of cities randomly assigned to exposure groups could have influenced call volume. For example, a greater number of veterans lived in communities that received high doses of messaging, and baseline call rates were also slightly higher in these areas when compared with the other exposure groups. However, no significant increases in call rates were identified for this group even when compared with other messaging strategies (e.g., call rates were not greater in high-dose vs. low-dose campaign areas despite higher baseline call rates and a larger veteran population in high-dose communities). Further, fewer residents living in areas where increased call rates were identified (i.e., mixed-message markets) reported characteristics that may influence VCL use (i.e., veteran status) when compared with the other exposure groups. Additionally, demographics of callers were not available and thus not considered in the current study. While this limits our ability to control for potential confounding, the primary goal of the It’s Your Call campaign was to broadly promote crisis line use to veterans and did not target specific subpopulations. The analyses presented provide important inCrisis (2017), 38(1), 53–62

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

itial insights into understanding the use of distinct public messaging strategies to promote this crisis support service. Future analyses could supplement this current work to explore campaign effects among different types of callers and subpopulations of veterans (e.g., females, younger veterans). While the campaign broadly promoted help seeking, data on motives for calling were also unavailable and not included in analyses. Research focused on such factors should be considered in future studies to determine reasons for crisis line use associated with campaigns. Study analyses were conducted with aggregate VCL call data in which the veteran status of callers was not confirmed. As such, it is possible that family or friends may have contributed to observed increases in call volume. However, the VCL may have still played an important role in facilitating help seeking by providing information to those close to a veteran in distress (e.g., how to access local resources). Finally, the generalizability of results from this study is limited. Analyses were focused on select pilot communities that are not necessarily representative of veterans living in the US, and the placement of advertisements (e.g., mass transit) may have restricted exposure to select subpopulations.

Conclusion The use of campaigns for suicide prevention is growing yet little is known of their impact on health behavior, particularly crisis line use. This study makes an important contribution to the small extant literature by providing initial insights into messaging strategies that may increase the likelihood of help seeking among veteran populations. Findings begin to address a priority area for suicide prevention (e.g., evaluation of communication efforts) and reinforce the need for further research on campaign and dissemination practices.

References Berkowitz, J. M., Huhman, M., & Nolin, M. J. (2008). Did augmenting the VERB campaign advertising in select communities have an effect on awareness, attitudes, and physical activity? American Journal of Preventive Medicine, 34(6), S256–266. Bossarte, R. M., Karras, E., Lu, N., Tu, X., Stephens, B., Draper, J., & Kemp, J. (2014). Associations between the VA’s suicide prevention campaign and calls to related crisis lines. Public Health Reports, 129, 428–438. Brancu, M., Straits-Tröster, K., & Kudler, H. (2011). Behavioral health conditions among military personnel and veterans: Prevalence and best practices for treatment. North Carolina Medical Journal, 72(1), 54–60. Chambers, D. A., Pearson, J. L., Lubell, K., Brandon, S., O’Brien, K., & Zinn, J. (2005). The science of public messages for suicide © 2016 Hogrefe Publishing

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

prevention: A workshop summary. Suicide and Life Threatening Behavior, 35, 134–46. doi:10.1521/suli.35.2.134.62871 Chisolm, D. J. (2010). Does online health information seeking act like a health behavior? A test of the behavioral novel. Telemedicine and e-Health, 16(2), 154–60. Daigle, M., Beausoleil, L., Brisoux, J., Raymond, S., Charbonneau, L., & Desaulniers, J. (2006). Reaching suicidal people with media campaigns: New challenges for a new century. Crisis, 27, 172–180. doi:10.1027/0227-5910.27.4.172 Dutta-Bergman, M. J. (2004). Complementarity in consumption of news types across traditional and new media. Journal of Broadcasting & Electronic Media, 48, 41–60. Elder, H., Karras, E., & Bossarte, R. M. (in press). Promoting help seeking to veteran households: Associations between exposure to multiple health messages and intentions to utilize related hotlines. Military Medicine. Farrelly, M. C., Davis, K. C., Haviland, M. L., Messeri, P., & Healton, C. G. (2005). Evidence of a dose-response relationship between “truth” antismoking ads and youth smoking prevalence. American Journal of Public Health, 95(3), 425-431. Harris, J. L., Pierce, M., & Bargh, J. A. (2014). Priming effect of antismoking PSAs on smoking behaviour: A pilot study. Tobacco Control, 23, 285–290. Hersey, J. C., Niederdeppe, J., Evans, W. D., Nonnemaker, J., Blahut, S., Holden, D., … Haviland, M. L. (2005). The theory of “truth”: How counterindustry media campaigns affect smoking behavior among teens. Health Psychology, 24, 22–31. Hornik, R. C. (2002a). Exposure: Theory and evidence abut all the ways that matter. Social Marketing Quarterly, 8(3), 30–67. Hornik, R. (Ed.). (2002b). Public health communication: Evidence for behavior change. Mahwah, NJ: Lawrence Erlbaum. Huhman, M. E., Potter, L. D., Duke, J. C., Judkins, D. R., Heitzler, C. D., & Wong, F. L. (2007). Evaluation of a national physical activity intervention for children: VERBTM campaign, 2002–2004. American Journal of Preventive Medicine, 32(1), 38–43. Karras, E., & Rintamaki, L. S. (2012). An examination of online health information seeking by deaf people. Health Communication, 27(2), 194–204. Karras, E., Lu, N., Stephens, B., Tu, X., Draper, J., Thompson, C., & Bossarte, R. M. (2016). Measuring associations of the VA’s suicide prevention campaign on the use of crisis support services. Suicide and Life-Threatening Behavior. doi:10.1111/sltb.12231 Karras, E., Stephens, B., Kemp, J., & Bossarte, R. M. (2014). Using media to promote suicide prevention hotlines to Veteran households. Injury Prevention, 20(1), 62–65. Kemp, J. E., & Bossarte, R. M. (2013). VA Suicide Data Report, 2012. Department of Veterans Affairs, Washington, DC. Koch, T., & Zerback, T. (2013). Helpful or harmful? How frequent repetition affects perceived statement credibility. Journal of Communication, 64(6), 993–1010. Kowalski, J., & Tu, X. (2007). Modern applied U statistics. New York, NY: Wiley. Langford, L., Litts, D., & Pearson, J. (2013). Using science to improve communications about suicide among military and veteran populations: Looking for a few good messages. American Journal of Public Health, 103(1), 31–8. LaVail, K. H., Anker, A. E., Reinhart, A., & Feeley, T. H. (2010). The persuasive effects of audiovisual PSAs to promote organ donation: The mediating role of psychological reactance. Communication Studies, 61(1), 46–69. Ma, M., Dollar, K. M., & Kibler, J. L. (2011). The effects of priming on a public health campaign targeting cardiovascular risks. Prevention Science, 12, 333–38. McVey, D., & Stapleton, J. (2000). Can anti-smoking television advertising affect smoking behaviour? Controlled trial of the Health Education Authority for England’s anti-smoking TV campaign. Tobacco Control, 9(3), 273–282. © 2016 Hogrefe Publishing

National Center for Health Statistics. (2016). Health, United States, 2015: With special feature on racial and ethnic health disparities. Hyattsville, MD: Author. Niederdeppe, J., Farrelly, M. C., & Haviland, L. (2004). Confirming “truth”: More evidence of a successful tobacco countermarketing campaign in Florida. American Journal of Public Health, 94(2), 255–257. Noar, S. M. (2006). A 10-year retrospective of research in health mass media campaigns: Where do we go from here? Journal of Health Communication, 11, 21–42. doi:10.1080/10810730500461059 Randolph, W., & Viswanath, K. (2004). Lessons learned from public health mass media campaigns: Marketing health in a crowded media world. Annual Review of Public Health, 25, 419–37. Rockett, I. R. H., Regier, M. D., Kapusta, N. D., Coben, J. H., Miller, T. R., Hanzlick, R. L., ... Smith, G. S. (2012). Leading causes of unintentional and intentional injury mortality: United States, 2000–2009. American Journal of Public Health, 102, e84–92. doi:10.2105/AJPH.2012.300960 Scheufele, D. A., & Tewksbury, D. (2007). Framing, agenda setting, and priming: The evolution of three media effects models. Journal of Communication, 57(1), 9–20. Snyder, L. B., Hamilton, M. A., Mitchell, E. W., Kiwanuka-Tondo, J., Fleming-Milici, F., & Proctor, D. (2004). A meta-analysis of the effect of mediated health communication campaigns on behavior change in the United States. Journal of Health Communication, 9, 71–96. doi:10.1080/10810730490271548 Southwell, B. G., Barmada, C. H., Hornik, R. C., & Maklan, D. M. (2002). Can we measure encoded exposure? Validation evidence from a national campaign. Journal of Health Communication, 7, 445–53. Till, B., Sonneck, G., Baldauf, G., Steiner, E., & Niederkrotenthaler, T. (2013). Reasons to love life. Effects of a suicide-awareness campaign on the utilization of a telephone emergency line in Austria. Crisis, 34(6), 382–389. United States Census Bureau. (2010a). 2010 American Community Survey. Retrieved from http://ftp2.census.gov/ United States Census Bureau. (2010b). 2010 Current Population Survey, Educational Attainment in the United States. Retrieved from https://www.census.gov/hhes/computer/ US Department of Health and Human Services (HHS) Office of the Surgeon General and National Action Alliance for Suicide Prevention. (2012). 2012 National Strategy for Suicide Prevention: Goals and objectives for action. Washington, DC: Author. Retrieved from http://www.surgeongeneral.gov/library/reports/ national-strategy-suicideprevention/ Wakefield, M. A., Loken, B., & Hornik, R. C. (2010). Use of mass media campaigns to change health behavior. Lancet, 376, 1261–71. doi:10.1016/S0140 6736(10)60809-4 Wilson, C. J., Bushnell, J. A., & Caputi, P. (2011). Early access and help seeking: Practice implications and new initiatives. Early Intervention Psychiatry, 5, 34–9. doi:10.1111/j.17517893.2010.00238.x Zivin, K., Kim, H. M., McCarthy, J. F., Austin, K. L., Hoggatt, K. J., Walters, H., & Valenstein, M. (2007). Suicide mortality among individuals receiving treatment for depression in the Veterans Affairs health system: Associations with patient and treatment setting characteristics. American Journal of Public Health, 97(12), 2193–2198.

Received October 30, 2015 Revision received March 29, 2016 Accepted April 8, 2016 Published online August 26, 2016 Crisis (2017), 38(1), 53–62

About the authors Dr. Karras is a health communication researcher at the VISN 2 Center of Excellence for Suicide Prevention, Canandaigua, NY, with the Department of Veterans Affairs, and holds an appointment with the Department of Psychiatry at the University of Rochester in Rochester, NY. Dr. Lu is a biostatistician with the Department of Biostatistics and Computational Biology at the University of Rochester in Rochester, NY. Heather Elder is a research assistant at the VISN 2 Center of Excellence for Suicide Prevention with the Department of Veterans Affairs, and a doctoral student in the epidemiology program in Public Health Sciences at the University of Rochester, Rochester, NY. Dr. Tu is Professor of Biostatistics and Psychiatry in the Department of Biostatistics and Computational Biology and Department of Psychiatry, University of Rochester, NY. He is the Director of the Statistical Consulting Center and the Director of the Psychiatric Statistics Division within the Department of Biostatistics and Computational Biology. Dr. Thompson is the Director of the Suicide Prevention Office in the Department of Veterans Affairs, Washington, DC.

Crisis (2017), 38(1), 53–62

E. Karras et al.: Messaging Strategies to Enhance Crisis Line Use

Dr. Tenhula is the Deputy Chief Consultant for Specialty Mental Health in Mental Health Services, Department of Veterans Affairs, Washington, DC. Dr. Batten is a senior associate with Booz Allen Hamilton, with a focus on behavioral health, health communications, and veterans health. She formerly served as the Deputy Chief Consultant for Specialty Mental Health with Mental Health Services in the Department of Veterans Affairs, Washington, DC. Dr. Bossarte is the Director of West Virginia University’s Injury Control Research Center, Morgantown, WV. He serves as an associate professor in the Department of Behavioral Medicine and is a member of the West Virginia Clinical and Translational Science Institute. He also works with the Suicide Prevention Office in the Department of Veterans Affairs. Elizabeth Karras VISN 2 Center of Excellence for Suicide Prevention Department of Veterans Affairs Canandaigua VA Medical Center 400 Fort Hill Avenue Canandaigua, NY 14424 USA elizabeth.karras@va.gov

Erratum Correction to Till et al., 2016 Article “Relationship Satisfaction and Risk Factors for Sui cide" by Benedikt Till, Ulrich S. Tran, and Thomas Nie derkrotenthaler (Crisis, 2016, advance online publication, doi: 10.1027/0227-5910/a000407) contained an error in the abstract. The corrected sentence reads as follows: Results: Risk factors for suicide were higher among s ingles than among individuals in happy relationships, but highest among those with low relationship satisfaction.

The editors regret any inconvenience or confusion this error may have caused.

Reference Till, B., Tran, U. S., & Niederkrotenthaler, T. (2016). Relationship satisfaction and risk factors for suicide. Crisis. Advance online publication. doi:10.1027/0227-5910/a000407

Crisis (2017), 38(1), 63 DOI: 10.1027/0227-5910/a000437

News, Announcements, and IASP Announcements The 50th Annual Conference of the American Association of Suicidology (AAS) is taking place April 26–29, 2017, in Phoenix, AZ, USA. For more information go to http://www.suicidology.org/annual-conference/50th-annualconference The 29th Annual Healing After Suicide Loss Conference of the American Association of Suicidology (AAS) is taking place April 29, 2017, in Phoenix, AZ, USA. For more information go to http://www.suicidology.org/annualconference/29th-annual-healing-conference The XXIX World Congress of the International Association for Suicide Prevention (IASP), entitled “Preventing Suicide: A Global Commitment, From Communities to

Continents,” is taking place July 18–22, 2017, in Kuching, Malaysia. For more information go to https://www.iasp. info/ or see next page The International Summit on Suicide Research of the International Academy of Suicide Research (IASR) and the American Foundation for Suicide Prevention (AFSP) is taking place November 5–8, 2017, in San Juan, Puerto Rico. For more information go to http://suicideresearchsummit.org/ The 17th European Symposium on Suicide and Suicidal Behaviour (ESSSB) is taking place September 5–8, 2018, in Ghent, Belgium. For more information go to http://www.esssb17.org

Awards The Stengel Research Award The Stengel Research Award, instituted in 1977, is named in honour of the late Professor Erwin Stengel, one of the founders of IASP. This award is for outstanding active research with at least 10 years of scientific activity in the field, as evidenced by number and quality of publications in internationally acknowledged journals and indicators of esteem and reputation. National representatives as well as members of the International Association for Suicide Prevention (IASP) are invited to nominate suitable persons for the Stengel Research Award. Nominees do not necessarily have to be IASP members. Self-nomination is permitted. Award committee members are not permitted to make nominations. Candidates should not be nominees for any other IASP Award in the current round. The criteria for selection are: The award is open to currently active academics with at least 10 years of scientific activity in the field, who have a deserved reputation for outstanding research in the field of suicidology, as evidenced by: Crisis (2017), 38(1), 64–67 DOI: 10.1027/0227-5910/a000462

a. The number and quality of publications in leading peer-reviewed journals that significantly contribute to our understanding/knowledge of suicidal behaviour and/or to suicide prevention; and b. Indicators of esteem and reputation (e.g., invitation to give keynote lecture at international conference, invitation to serve on editorial board of international journal). A researcher who continues to be active in the field despite being officially retired could be considered for this award. A retired researcher who is no longer active in the field cannot be considered for this award. Stengel Research Chairperson: Prof. Paul Yip Director Centre for Suicide Research and Prevention Social Work and Social Administration University of Hong Kong Tel. +852 2831-5190 Fax +852 2549-7161 E-mail sfpyip@hku.hk

News, Announcements, and IASP

Please send your nomination directly to Prof. Paul Yip, Chairperson of the Stengel Research Award, and attach a brief summary (CV) of why your nominee is deserving of the award. Deadline: April 15th, 2017

The Ringel Service Award The Ringel Service Award, instituted in 1995, honours the late Professor Erwin Ringel, the founding President of the Association. This award is for distinguished service in the field of suicidology. National representatives as well as members of the International Association for Suicide Prevention (IASP) are invited to nominate suitable persons for the Ringel Service Award. Nominees do not necessarily have to be IASP members. Self-nomination is permitted. Award committee members are not permitted to make nominations. Candidates should not be nominees for any other IASP Award in the current round. The criteria for selection are: The award is open to currently active practitioners who: a. Have been involved in the development and implementation of evidence-informed and best practice suicide prevention programs/projects over a period of at least 10 years; and (b) are acknowledged as a leader in the field, as confirmed by opinion leaders (in academic, practitioner, voluntary, and/or community sectors) nationally or internationally. A practitioner who continues to be active in the field despite being officially retired could be considered for this award. A retired practitioner who is no longer active in the field cannot be considered for this award. Ringel Service Award Chairperson: Prof. Lars Mehlum University of Oslo National Center for Suicide Research & Prevention Unit Sognsvannsveien 21 Bygning 12, Oslo No-0327 Norway Tel. +47 22 923-473 Fax +47 22 923-958 E-mail lars.mehlum@medisin.uio.no

The Farberow Award The Farberow Award was introduced in 1997 in recognition of Professor Norman Farberow, a founding member and driving force behind the IASP. This award is for a person who has contributed significantly in the field of work with survivors of suicide. National representatives as well as members of the International Association for Suicide Prevention (IASP) are invited to nominate suitable persons for the Stengel Research Award. Nominees do not necessarily have to be IASP members. Self-nomination is permitted. Award committee members are not permitted to make nominations. Candidates should not be nominees for any other IASP Award in the current round. The award is open to currently active practitioners or academics who: a. Have been involved in the development and implementation of evidence-informed and best practice postvention programmes/projects over a period of at least 10 years; and b. Are acknowledged as a leader in the field, as confirmed by opinion leaders (in academic, practitioner, voluntary and/or community sectors) in bereavement/ postvention, nationally or internationally. A potential candidate who continues to be active in the field despite being officially retired could be considered for this award. A potential candidate who is retired and no longer active in the field cannot be considered for this award. Farberow Award Chairperson: Ms. Jill Fisher 8 Kundart Street Coes Creek, QLD 4560 Australia Tel. +61 448 19347 E-mail tarome@bigpond.com.au Please send your nomination directly to Ms. Jill Fisher, Chairperson of the Farberow Award Committee, and attach a brief summary (CV) of why you feel your nominee is deserving of the award. Deadline: April 15th, 2017

Please send your nomination directly to Prof. Lars MehÂ lum, Chairperson of the Ringel Service Award Committee, and attach a brief summary (CV) of why your nominee is deserving of this award. Deadline: April 15th, 2017 ÂŠ 2017 Hogrefe Publishing

Crisis (2017), 38(1), 64â&#x20AC;&#x201C;67

News, Announcements, and IASP

De Leo Fund Award

Andrej Marušič Award

The De Leo Fund Award honours the memory of Nicola and Vittorio, the beloved children of Professor Diego De Leo, IASP Past President. The award is offered to distinguished scholars in recognition of their outstanding research on suicidal behaviors carried out in developing countries. National representatives as well as members of the International Association for Suicide Prevention (IASP) are invited to nominate suitable persons for the De Leo Fund Award. Nominees do not necessarily have to be IASP members. Self-nomination is permitted. Award committee members are not permitted to make nominations. Candidate should not be nominees for any other IASP Award in the current round. The award is open to mid-career researchers (no more than 20 years from graduation) who: a. Have performed (most of) their research in a low- or middle-income country (as defined by the World Bank) and b. Who have a deserved reputation for outstanding research in the field of suicidology, as evidenced by the quality of publications in leading peer-reviewed journals that significantly contribute to our understanding/ knowledge of suicidal behavior and/or to suicide prevention.

The Andrej Marušič Award is dedicated to young researchers who are invited to submit a research proposal relating to suicidal behavior and prevention for consideration. National representatives as well as members of the International Association for Suicide Prevention (IASP) are invited to nominate suitable persons for the Andrej Marušič Award. Nominees do not necessarily have to be IASP members. Self-nomination is permitted. Award committee members are not permitted to make nominations. Candidates should not be nominees for any other IASP Award in the current round. The award is open to young researchers (under 40 years old or with no more than five years of experience in the field of suicidology) who submit a research proposal which: a. Identifies a topic/question which is important for suicidology and/or suicide prevention; b. Demonstrates methodological rigour; c. Can be feasibly implemented (shows the necessary mix of skill, experience, project management, and resources for success); d. Meets appropriate ethical standards; and e. Is likely to produce findings which add to the existing body of knowledge and/or lead to changes in suicide prevention policy or practice.

All potential candidates must be currently active in the field. De Leo Fund Award Chairperson: Prof. Diego De Leo Australian Institute for Suicide Research and Prevention Griffith University 176 Messines Ridge Rd. Mt. Gravatt Campus Mt. Gravatt, QLD 4122 Australia Tel. + 61 7 3735-3377 Fax + 61 7 3735-3450 E-mail d.deleo@griffith.edu.au Please send your nomination directly to Prof. Diego De Leo, Chairperson of the De Leo Fund Award Committee, and attach a brief summary (CV) of why your nominee is deserving of the award.

All potential candidates must be currently active in the field. Andrej Marušič Award Chair: Dragan Marušič Administrator Dr. Vita Poštuvan Univerza na Primorskem Inštitut Andrej Marušič Muzejski trg 2 6000 Koper Slovenia Tel. +386 4049-0580 E-mail vita.postuvan@upr.si Please send your nomination directly to Dr. Vita Poštuvan, Administator of the Andrej Marušič Award Committee, and attach a brief summary (CV) of why your nominee is deserving of the award. Deadline: April 15th, 2017

Deadline: April 15th, 2017

Crisis (2017), 38(1), 64–67

News, Announcements, and IASP

International Association for Suicide Prevention Would you like to join IASP?

−− to carry out programs of research, especially those which can be pursued through international joint cooperation.

IASP is a nonprofit organization for those interested and working in suicide prevention. The membership consists of individuals (clinicians, scientists, crisis workers, volunteers and persons who have lost a family member by suicide) and community, national and international organizations. At this moment the membership extends over 50 countries and is affiliated with the World Health Organization as the key organization concerned with suicide prevention.

The activities of the association focus on the importance of sharing information, research and knowledge in order to address the issues of suicide and its prevention. The biennial congress, the bimonthly journal Crisis, biannual newsletter and the website serves as a medium for international exchange and a directory of members involved in suicidology is made available. Candidates for IASP membership may apply directly to the Central Administration Office membership@iasp.info or join IASP online at http://www.iasp.info Membership consists of: −− Regular membership for individuals regularly engaged in suicide studies, clinical management of suicidal patients, and/or suicide prevention activities, and who are interested in furthering the work of the Association. −− Organizational membership is for national or international voluntary or other nonprofit organizations (incorporated or unincorporated) working in the field of suicide studies, clinical management of suicidal patients and/or suicide prevention activities under the terms of their governing document. −− Associate membership is for individuals and organizations who are interested in working toward the goals of the Association, but who are not qualified for regular membership. Please contact the IASP Central Administration Office at membership@iasp.info.; on the web at http://www.iasp.info

The main objectives of IASP are: −− to provide a common platform for all representatives of the many professions who are engaged in the field of suicide prevention and crisis intervention; −− to allow interchange of acquired experience in this area in various countries, especially through the exchange of literature; −− to promote the establishment of national organizations for suicide prevention; −− to facilitate the wider dissemination of the fundamentals of effective suicide prevention to professional groups and to the general public; −− to arrange for specialized training of selected persons in the area of suicide prevention in selected training centers;

IASP membership fees (include a subscription to Crisis) Zone

Crisis

Individual Membership

Organizational Membership

Hard copy

US $190.00 (early bird US $180.50, 3 years US $515.00)

US $235.00 (early bird US $225.00, 3 years US $635.00)

2 3 4 All zones: students, volunteers, & retirees

Online only

US $162.00 (early bird US $153.00, 3 years US $438.00)

US $200.00 (early bird US $190.00, 3 years US $543.00)

Hard copy

US $160.00 (early bird US $152.00, 3 years US $430.00)

US $180.00 (early bird US $170.00, 3 years US $485.00)

Online only

US $136.00 (early bird US $129.00, 3 years US $368.00)

US $153.00 (early bird US $145.00, 3 years US $414.00)

Hard copy

US $135.00 (early bird US $128.00, 3 years US $365.00)

US $160.00 (early bird US $150.00, 3 years US $430.00)

Online only

US $115.00 (early bird US $109.00, 3 years US $310.00)

US $136.00 (early bird US $129.00, 3 years US $368.00)

Hard copy

US $115.00 (early bird US $109.00, 3 years US $310.00)

US $125.00 (early bird US $120.00, 3 years US $340.00)

Online only

US $98.00 (early bird US $93.00, 3 years US $239.00)

US $106.00 (early bird US $101.00, 3 years US $288.00)

Online only

US $115.00 (early bird US $109.00, 3 years US $310.00)

For Advanced Organization Membership fees go to http://www.iasp.info/application.php

IASP Executive Committee 2015–2019 President: Vice President 1: Vice President 2: Vice President 3:

Prof. Ella Arensman Prof. Murad M. Khan Prof. Steve Platt Prof. Maurizio Pompili

General Secretary: Prof. Jane Pirkis Treasurer: Dr. Mort M. Silverman Representative of the Council of Presidents: Prof. Diego de Leo

IASP Council of National Representatives 2015–2019 Co-Chairs: Dr. Loraine Barnaby and Prof. Thomas Niederkrothentaler For a current listing of the IASP Council of National Representatives please refer to https://www.iasp.info/council_of_national_representatives. php#natrep © 2017 Hogrefe Publishing

Crisis (2017), 38(1), 64–67

Instructions to Authors – Crisis: The Journal of Crisis Intervention and Suicide Prevention Crisis – The Journal of Crisis Intervention and Suicide Prevention is an international periodical that publishes original articles on suicidology and crisis intervention. Papers presenting basic research as well as practical experience in the field are welcome. Crisis also publishes potentially life-saving information for all those involved in crisis intervention and suicide prevention, making it important reading for clinicians, counselors, hotlines, and crisis intervention centers. Crisis: The Journal of Crisis Intervention and Suicide Prevention publishes the following types of articles: Research Trends, Short Reports, and Clinical Insights. Manuscript Submission: All manuscripts should be submitted electronically at http://www.editorialmanager.com/cri Detailed instructions to authors are provided at http://www. hogrefe.com/j/cri Copyright Agreement: By submitting an article, the author confirms and guarantees on behalf of him-/herself and any coauthors that he or she holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches and tables, and that the article and its contents do not infringe in any way on the rights of third parties. The author indemnifies and holds harmless the publisher from any third-party claims. The author agrees, upon acceptance of the article for publication, to transfer to the publisher on behalf of him-/herself and any coauthors the exclusive right to reproduce and distribute the article and its contents, both physically and in nonphysical, electronic, and other form, in the journal to which it has been submitted and in other independent publications, with no limits on the number of copies or on the form or the extent of the distribution. These rights are transferred for the duration of copyright as defined by international law. Further-

more, the author transfers to the publisher the following exclusive rights to the article and its contents: 1. The rights to produce advance copies, reprints, or offprints of the article, in full or in part, to undertake or allow translations into other languages, to distribute other forms or modified versions of the article, and to produce and distribute summaries or abstracts. 2. The rights to microfilm and microfiche editions or similar, to the use of the article and its contents in videotext, teletext, and similar systems, to recordings or reproduction using other media, digital or analog, including electronic, magnetic, and optical media, and in multimedia form, as well as for public broadcasting in radio, television, or other forms of broadcast. 3. The rights to store the article and its content in machine-readable or electronic form on all media (such as computer disks, compact disks, magnetic tape), to store the article and its contents in online databases belonging to the publisher or third parties for viewing or downloading by third parties, and to present or reproduce the article or its contents on visual display screens, monitors, and similar devices, either directly or via data transmission. 4. The rights to reproduce and distribute the article and its contents by all other means, including photomechanical and similar processes (such as photocopying or facsimile), and as part of so-called document delivery services. 5. The right to transfer any or all rights mentioned in this agreement, as well as rights retained by the relevant copyright clearing centers, including royalty rights to third parties. Online Rights for Journal Articles: Guidelines on authors’ rights to archive electronic versions of their manuscripts online are given in the document “Guidelines on sharing and use of articles in Hogrefe journals” on the journal’s web page at www.hogrefe. com/j/cri. February 2016

29th WORLD CONGRESS OF THE INTERNATIONAL ASSOCIATION FOR SUICIDE PREVENTION 18 â&#x20AC;&#x201C; 22 JULY, 2017 KUCHING (SARAWAK) MALAYSIA The 29th IASP World Congress will take place in Kuching, the capital city of the Malaysian state of Sarawak, situated on the island of Borneo. The Congress is held immediately after the World Rainforest Music Festival 2017, so that delegates who love music and culture may have a taste of the East. Sarawak is known for its natural beauty and friendly people. The orang utan and the hornbill are native to this island state.

Attractive very Early bird registration fees and hotel rates are available. For more information please go to the website: www.iasp2017.org

Keep up-to-date with recent research and practice in suicide

Diego De Leo / Vita PoĹĄtuvan (Editors)

Resources for Suicide Prevention Bridging Research and Practice 2017, x + 158 pp. US $36.80 / â&#x201A;Ź 29.95 ISBN 978-0-88937-454-6 Also available as eBook This book is based on the TRIPLE i in Suicidology international conferences, which are organised annually by the Slovene Centre for Suicide Research in memory of the late Prof. Andrej MaruĹĄic with the aim of promoting intuition, imagination, and innovation in the research and prevention of suicide and suicidal behaviour. The carefully selected chapters provide food for thought to practitioners, researchers, students, and all those who come into contact with the tragedy of suicide, with the hope of stimulating new ideas and interventions in the difficult fight against suicidal behaviours. In four parts, the interna-

www.hogrefe.com

tionally renowned team of authors summarise the achievements of suicidology so far (both in quantitative and qualitative research), present effective interventions in suicide prevention (including for youths and older people) and knowledge gained in bereavement and postvention studies (such as in different cultures and those bereaved by suicide), and highlight future directions for suicide research and prevention. The volume is thus a useful resource for all those interested in keeping up-to-date with recent research and practice in suicide.