Proefschrift fpc waterschoot by Franka Waterschoot

Nice to have or need to have? Unraveling Dosage of Pain Rehabilitation

Franka Waterschoot

The research project “Nice to have or need to have” was funded by: Pijnrevalidatie - Centrum voor revalidatie UMCG

The publication of this thesis was ﬁnancially supported by: University of Groningen - RUG Graduate school for Health Research - SHARE Centrum voor Revalidatie - Universitair Medisch Centrum Groningen (UMCG) Pijnrevalidatie - Centrum voor revalidatie UMCG Stichting Beatrixoord Noord - Nederland Menzis Zorgverzekering

Cover design:

Isontwerp - ‘s Hertogenbosch

Layout:

Gildeprint - Enschede

Printed by:

Gildeprint - Enschede

Waterschoot, Franka. Nice to have or need to have. Unraveling dosage of pain rehabilitation. Thesis University of Groningen, the Netherlands – with references – with summary in Dutch.

Nice to have or need to have? Unraveling Dosage of Pain Rehabilitation

Proefschrift

ter verkrijging van de graad van doctor aan de Rijksuniversiteit Groningen op gezag van de rector magniďŹ cus prof. dr. E. Sterken en volgens besluit van het College voor Promoties. De openbare verdediging zal plaatsvinden op maandag 15 februari 2016 om 14.30 uur

door

Franka Petronella Catharina Waterschoot geboren op 20 juni 1978 te Hoogeloon

Promotores Prof. dr. M.F. Reneman Prof. dr. J.H.B. Geertzen Prof. dr. P.U. Dijkstra

Beoordelingscommissie Prof. dr. J.G.M. Rosmalen Prof. dr. K.M.G. Schreurs Prof. dr. C. van Bennekom

Paranimfen:

Annemieke de Jong Berry van Holland

Table of content Chapter 1 Chapter 2

General introduction. Dose or content? Effectiveness of pain rehabilitation programs for

9 19

patients with chronic low back pain: A systematic review. Chapter 3

Course of disability reduction during a pain rehabilitation program:

A prospective clinical study. Chapter 4

Dosage of pain rehabilitation programs for patients with chronic

musculoskeletal pain: A non- inferiority randomized controlled trial. Chapter 5

Dosage of pain rehabilitation programs: A qualitative study from

patient and professionalsâ&#x20AC;&#x2122; perspectives. Chapter 6

Towards a case complexity index of patients with chronic

non-speciďŹ c musculoskeletal pain: A Delphi and feasibility study. Chapter 7

General discussion.

109

Summary

123

Samenvatting

129

Dankwoord

137

Over de auteur

143

Research institute SHARE

147

Extremities, Pain and Disability (Expand)

153

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 10 | Chapter 1

Background of thesis

In Europe, 19% of the adult population suffers from chronic pain of moderate to severe intensity 1

. Chronic pain is associated with lower quality of life, less participation in major life domains,

higher healthcare utilization, lower work productivity, and higher direct and indirect costs

1,2

Among the many treatment options for chronic pain are multidisciplinary pain rehabilitation programs (PRPs). There is robust evidence that PRPs are effective

3-5

. They can help to improve

quality of life, improve participation in major life domains, and decrease healthcare utilization. There is conflicting evidence on the benefits for work-related outcomes, however 5,6. The positive effects are robust, although the average size of the effects is modest and there is a great need for further improvements. Optimizing the outcomes of PRPs should benefit patients suffering from chronic pain, and the benefits should extend to the patients’ families, in addition to other stakeholders such as rehabilitation centers, healthcare providers, healthcare insurers, employers, coworkers and disability insurers (temporary and permanent). Up to now, all studies that have aimed at optimizing outcomes of PRP have focused on content, composition and delivery form, such as individually or multidisciplinary therapy

3-10

. However, optimal PRP may

also relate to dosage of PRP. Dosage of PRP appears to be a neglected, but potentially relevant, topic in PRP and PRP research. Dose variables can be expressed as the total duration and amount of contact hours (Text box). Additionally, the number of healthcare professionals involved in the program, regardless of whether an inpatient or an outpatient program is offered, and dosage per contact hour are related to the total dosage of PRP. There are several motives for arguing why a closer look at the dosage of PRP might be necessary: -

In clinical practice, large variations of PRP-dosage are observed, in the Netherlands as well as internationally. There are also large variations observed in dosages that have been described in PRP research 6,7. The rationale for dosage variations is unknown.

Scientiﬁcally, a consequence of research focused on content only is that inﬂuence of dosage on the effect of PRP is neglected. For example, a study concluding inferiority of PRP, comparing multidisciplinary PRP of 100 hours with mono-disciplinary physiotherapy of 50 hours could be confounded because not only does the content differ, but also dosage does differ. Theoretically, both content and dose variables can contribute to the effectiveness of PRP. There is a lack of studies comparing multidisciplinary PRPs with only dosage as a comparing variable: randomized controlled trials comparing differences in dosage between multidisciplinary treatments with the same content have not yet been performed.

Viewed from the process of care, optimizing the entire logistic process of PRP, from registration to discharge, is important. It can reduce waiting lists and improve cost effectiveness, beneﬁtting both patients and rehabilitation centers. To manage this process, it is necessary

General introduction | 11

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

to determine the total number of patients treated, as well as the in- and outflow of patients. Consequently, to influence the outflow to optimize inflow, it is necessary to determine the offered dosage of PRP. -

Ethically, not knowing the optimal dosage of PRP, the question is raised as to whether overor undertreatment will occur.

Developments in the Dutch healthcare system require more transparency regarding quality and financial sustainability of specialist medical care. This transparency will lead to a better understanding of rationale of care, which will benefit patients, health insurers and healthcare providers. These developments call for comparability of treatments with regard to content, composition, dosage, and finances.

Therefore, the time has come to include dose variables in the question of “what works for whom”, and therefore to analyze dosage at the level of the individual patient related to accomplishing their goals, and analyzing dosage related to effect from the perspective of different stakeholders to optimize the cost-effectiveness of multidisciplinary PRPs. Textbox Deﬁnitions related to dosage of pain rehabilitation programs Pain rehabilitation programs

Deﬁnitions Multidisciplinary treatments for patients suffering from chronic musculoskeletal pain, existing of a combination of physical and psychosocial components, with ﬁnal responsibility of the rehabilitation physician to coach patients on managing their pain and understanding the consequences of pain in daily life 11

Chronic pain

Pain without apparent biological value that has persisted beyond the normal tissue healing time (usually taken to be three months) 12

Disability

Disability is an umbrella term for impairments, activity limitations and participation restrictions referring to the negative aspects of the interaction between a person’s health condition(s) and that individual’s contextual factors (environmental and personal factors) 13

Dosage variables: • Duration of program

Total length of program expressed in weeks or months, from ﬁrst to last contact

• •

Intensity of program The number of contact hours per week or month Intensity of content/exercise The extent of exercise or other content per unit of time, frequency of exercises per session, intensity of the forces applied or exertion required

•

Contact hours

Total number of hours that a patient spends with his rehabilitation professional during the program

•

Number of healthcare professionals

The number of different disciplines (psychologists, physiotherapists, occupational therapists) involved in the program

Generally, patients with chronic pain improve signiﬁcantly in several domains (pain-related disability, return to work, quality of life) after PRP

5,6

. But it is presently unknown how much

treatment is needed to achieve the beneﬁcial effects or enhance the effect size. Multidisciplinary

12 | Chapter 1

PRPs are based on the bio-psycho-social model that aims to reduce pain-related disability. The cognitive-behavioral components assume that patients change their pain management skills during PRP. These behavioral changes support the reduction of pain-related disability without the aim of pain reduction. Consequently, aiming for behavioral changes during PRP, the dosage of PRP could be based on the dosage of other therapies that use cognitive-behavioral therapy for chronic pain. Unfortunately, the optimum dosage of those cognitive-behavioral interventions for patients with chronic pain is also unknown 14. In conclusion, there is currently a gap in knowledge regarding dosage of PRP based on cognitive-behavioral principles. In order to ﬁnd the optimum dosage of PRP, underlying constructs of dosage of PRP should be explored. When unraveling this topic in pain rehabilitation research, multiple questions were raised. Some questions are as follows: •

What is the rationale for the dosage of the present programs in today’s daily practice in Dutch PRP?

•

Is therapy needed each week or is a time-out also possible?

•

Which healthcare professionals should be involved in PRP and what is their added value in PRP?

•

Which patients require higher dosage of PRP and for whom would a lower dosage be sufﬁcient?

•

What are reasons to stop PRP and how do these reasons relate to the desired improvements?

•

How does reduction of disability occur during weeks of PRP?

•

Can the dosage of PRP be reduced without a reduction in the effects?

•

Should there be different dosages for patients that exhibit differences in case complexity?

Aim and research questions The overall aim of the thesis is to acquire insight into the effects of dosage of multidisciplinary pain rehabilitation programs for patients with chronic pain. This overall aim is speciﬁed in the following research questions: •

What is the reported dosage of treatment for patients with chronic low back pain and what is the inﬂuence of dose variables on the outcome of PRPs for patients with chronic low back pain? (Chapter 2).

•

Which course does disability reduction during PRP follow and which factors inﬂuence this course over time? (Chapter 3).

•

Is a shorter form of PRP non-inferior to the PRP in care as usual? Do effects differ signiﬁcantly? (Chapter 4).

General introduction | 13

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

•

What are the differences and similarities in the experiences and perspectives of patients who underwent PRP and rehabilitation professionals working in PRPs regarding dosage of PRP between three centers in the Netherlands? (Chapter 5).

•

Which factors determine case complexity of patients with CMP that are eligible for PRP? (Chapter 6).

Outline of thesis Since, up to now, dosage has been a neglected topic in pain research, different research methods and perspectives have been applied in this thesis. The systematic review and meta-analysis described in Chapter 2 aimed to analyze the influence of dose variables on the outcome of PRPs. There were no studies found that primarily analyzed the relationship between dose and effects of PRP. Therefore, this review focused on randomized controlled trials (RCTs) aimed at assessing effectiveness of PRPs that reported dose variables of the studied programs. The reported dose variables were used to analyze the influence of dose variables on the effects. The course of disability reduction during PRP and factors influencing this course were analyzed in Chapter 3. The majority of studies analyzed pre- and post-outcomes of disability reduction. However to acquire insight into the dosage of PRP, it is a prerequisite to analyze the course of the primary outcome during PRP. Chapter 4 describes the first randomized controlled trial in PRP research that compared PRPs with similar content but different dosages. The control arm was PRP performed in care as usual dosage (8, 12, 16 or 20 weeks) and the experimental arm was PRP performed at a duration four weeks shorter than the care as usual. To gain in-depth information on the experiences and perspectives of patients and rehabilitation professionals regarding dosage of PRP, a qualitative study was performed, which is described in Chapter 5. Interviews with patients and focus group interviews with rehabilitation professionals were conducted in three centers in the Netherlands offering PRP with differences in dosage of PRP, while patient characteristics were similar. From clinical practice, a relationship was hypothesized between case complexity and the required dosage of PRP. However, little research is available regarding the interpretation of case complexity within multidisciplinary PRP. Therefore, it was not possible to test this relationship. In Chapter

14 | Chapter 1

6, the operationalization of case complexity was studied using a Delphi technique among rehabilitation professionals working in pain rehabilitation centers throughout the Netherlands. In Chapter 7, the general discussion, all ďŹ ndings from the previous chapters are integrated and discussed. Implications for clinical practice and recommendations for future research are presented.

General introduction | 15

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1.

Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain in Europe: Prevalence, impact on daily life, and treatment. Eur J Pain. 2006;10(4):287-333.

Lambeek LC, van Tulder MW, Swinkels IC, Koppes LL, Anema JR, van Mechelen W. The trend in total cost of back pain in the netherlands in the period 2002 to 2007. Spine (Phila Pa 1976). 2011;36(13):10501058.

Flor H, Fydrich T, Turk DC. Efﬁcacy of multidisciplinary pain treatment centers: A meta-analytic review. Pain. 1992;49(2):221-230.

Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: A systematic review of interventions and outcomes. Rheumatology (Oxford). 2008;47(5):670-678.

Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350:h444.

van Geen JW, Edelaar MJ, Janssen M, van Eijk JT. The long-term effect of multidisciplinary back training: A systematic review. Spine (Phila Pa 1976). 2007;32(2):249-255.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

Smeets RJ, Vlaeyen JW, Hidding A, et al. Active rehabilitation for chronic low back pain: Cognitivebehavioral, physical, or both? ﬁrst direct post-treatment results from a randomized controlled trial [ISRCTN22714229]. BMC Musculoskelet Disord. 2006;7:5.

Dufour N, Thamsborg G, Oefeldt A, Lundsgaard C, Stender S. Treatment of chronic low back pain: A randomized, clinical trial comparing group-based multidisciplinary biopsychosocial rehabilitation and intensive individual therapist-assisted back muscle strengthening exercises. Spine (Phila Pa 1976). 2010;35(5):469-476.

10.

Kainz B, Gulich M, Engel EM, Jackel WH. Comparison of three outpatient therapy forms for treatment of chronic low back pain-- ﬁndings of a multicentre, cluster randomized study. Rehabilitation (Stuttg). 2006;45(2):65-77.

11.

Köke A, Brouwers M, Heuts P, et al. Consensus rapport pijnrevalidatie Nederland. 2005.

12.

International Association for the Study of Pain. Task Force on Taxonomy. Classiﬁcation of chronic pain: Descriptions of chronic pain syndromes and deﬁnitions of pain terms. 2nd ed. Seattle: IASP Press; 1994.

13.

World HO, Organisation Mondiale de lS, World HO. How to use the ICF: A practical manual for using the international classiﬁcation of functioning, disability and health (ICF): Exposure draft for comment. Genf: WHO; 2013.

14.

Ehde DM, Dillworth TM, Turner JA. Cognitive-behavioral therapy for individuals with chronic pain: Efﬁcacy, innovations, and directions for research. Am Psychol. 2014;69(2):153-166.

16 | Chapter 1

General introduction | 17

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Fran Fr anka nka P.C C. Wa W te t rs rsch choo ch oot, oo t, Pie ete terr U. Dijijks kstr ks trra Ph P D, Nie ek Ho Holllak MSc Sc, Hait Ha i ze J. de Vri it r ess PhD hD, Jaan H. H.B. B. Gee e rt r ze z n MD D, Ph PhD, D, Micchi hiel e F. Re el Rene neema an Ph P D. D

P in Pa in, 20 2014 1 ,1 14 155 (1) 1):: 17 1799-- 89 89..

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Abstract We sought to systematically analyze the influence of dose of Pain Rehabilitation Programs (PRPs) for patients with chronic low back Pain (CLBP) on disability, work participation and quality of life (QoL). Literature searches were performed in PUBMED, Cochrane library, Cinahl and EMBASE up to October 2012, using MeSH terms, other relevant terms and free-text words. Randomized controlled trials in English, Dutch and German, analyzing the effect of PRPs, were included. One of the analyzed interventions had to be a PRP. Outcomes should be reported regarding disability, work participation or QoL. To analyze dose, the number of contact hours should be reported. Two reviewers independently selected titles, abstract and full text articles on the basis of inclusion and exclusion criteria. Data were extracted and risk of bias was assessed. Effect sizes (ES) were calculated for each intervention, and influence of dose variables was analyzed by a mixed model analysis. Eighteen studies were identified, reporting a wide variety of dose variables and contents of PRPs. Analyses showed that evaluation moment, number of disciplines, type of intervention, duration of intervention in weeks, percentage of women, and age, influenced the outcomes of PRPs. The independent effect of dose variables could not be distinguished from content because these variables were strongly associated. Because dose variables were never studied separately or reported independently, we were not able to disentangle the relationship between dose, content, and effects of PRPs on disability, work participation and QoL. Keywords Chronic low back pain, dose, effectiveness, pain rehabilitation programs

20 | Chapter 2

Introduction Multidisciplinary pain rehabilitation programs (PRPs) are effective in improving daily functioning of patients with chronic low back pain (CLBP)

1-3

. Most studies investigating the effects of PRPs

focused on the relationship between therapy content and effect. Guidelines for managing CLBP are based on evidence obtained from these studies 4,5. However, this evidence might be biased. In a literature search, we could not identify any study analyzing the relationship between dose and effect as a primary objective. Two systematic reviews

2,3

were identiﬁed that analyzed dose

and effect as a secondary objective. These reviews presented conﬂicting conclusions on dose effects. In a review of Guzman et al 2, 10 randomized controlled trials (RCTs) were included reporting on 12 PRPs. PRPs were divided into 2 categories: daily intensive programs with more than 100 h of therapy, and once- or twice-weekly programs with less than 30 h of therapy. Based the basis of that distinction, it appeared that multidisciplinary PRPs of more than 100 h were superior to monodisciplinary treatment, and PRPs of less than 30 h were not. The authors concluded that intensive multidisciplinary PRPs are superior to less intensive multidisciplinary PRPs 2. These results have been used in guidelines and clinical practice

4,6

. However, it may be debated whether the

conclusion regarding dose is valid to support its clinical implications. First, there is an absence of knowledge about PRPs with doses between 30 to 100 h. Second, no distinction was made within intensive and less intensive PRPs, although there was a wide variety within the groups. Less intensive PRPs ranged from 17.5 to 30 h, while intensive PRP ranged from 100 to 280 h. Third, the review 2 was designed to assess the effect of different PRPs on clinically relevant outcomes. All PRPs differed not only in dose, but also in content, setting and number of disciplines involved, which may have confounded the dose aspect of the conclusion. Additionally, the authors discussed whether the improvements gained with intensive PRP are worth the expense. The review of Van Geen et al.

also had some limitations. It assessed the long-term effect of

PRPs with different contents and doses. Ten studies were included, and a distinction was made between 30 h of training a week or more (intensive therapy) vs less than 30 h of training a week (low-intensive therapy). Both the intensive and low-intensive PRPs showed positive effects. The conclusion in that review was that the dose of the intervention had no substantial inﬂuence on the effectiveness of the intervention. In summary, while many studies provide evidence for effectiveness of multidisciplinary PRP, dose of multidisciplinary PRP to achieve these effects remains unclear. The objective of the current study was to analyze the inﬂuence of dose variables on the outcome of PRPs for patients with CLBP.

Dose or content? Effectiveness of PRP | 21

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Methods Publications were retrieved by computer-aided search on PUBMED, the Cochrane library, Cinahl, and EMBASE up to October 2012. A specific search was developed using MeSH terms and other relevant terms for each database. The PubMed search is described in the Supplement. Refworks was used to store the results of the searches and to remove duplicates. Selection of studies Selection criteria were applied independently by two reviewers (FW and NH). The retrieved studies were first selected by title and abstract. Doubtful cases were discussed by the reviewers and included or excluded for full text analysis by consensus. Full-text reports of studies eligible for inclusion were analyzed. Disagreements were resolved by consensus or when necessary by a third reviewer (HdV). Studies were selected on the basis of the following inclusion criteria: (1) RCTs written in English, Dutch or German; (2) the objective was to assess effectiveness of a multidisciplinary PRP for patients with CLBP; PRP was defined as a rehabilitation program on the basis of the biopsychosocial model 7

with 3 or more disciplines providing the program (with or without a medical doctor); (3) total

number of contact hours of PRP was described; (4) participants were between 18 and 65 years with disabling nonspeciﬁc CLBP for at least 3 months; (5) outcome variables were described in the domain of disability, work participation or quality of life (QoL). Studies were excluded if: (1) the multidisciplinary PRP was given in primary care; (2) the objective was to assess effectiveness of only a biomedical intervention; (3) the study included participants who were diagnosed with speciﬁc disorders or severe comorbidities interfering with PRP, such as heart failure, rheumatoid arthritis, or psychiatric disorders; and (4) only total duration of PRP was reported. Data management Risk of bias of included studies was assessed according to the Cochrane Back Review Group

by 2 reviewers independently (FW and HdV). The criteria are presented in Table 1. Each criterion was scored as positive (Y), negative (N) or unclear (U). The total score was computed by counting the number of criteria scored as positive. Studies with a score of 6 or higher were defined as low risk of bias; a score lower than 6 was defined as high risk of bias. In case of unclear scores, corresponding authors of the studies were contacted by e-mail. Performance bias was analysed with different items of blinding. The item regarding blinding of care providers for intervention is frequently impossible in nondrug trials. To analyze the influence of lack of blinding of care providers on the judgment of methodological quality, we performed a sensitivity analyses by excluding item 4 of the risk of bias analyses.

22 | Chapter 2

Table 1 Criteria risk of bias analyses 1 2 3 4 5 6 7 8 9 10 11 12

Criteria risk of bias analyses Was the method of randomization adequate? Was the treatment allocation concealed? Was knowledge of the allocated interventions adequately prevented during the study? Was the patient blinded to the intervention? Was the care provider blinded to the intervention? Was the outcome assessor blinded to the intervention? Were incomplete outcome data adequately addressed? Was the drop-out rate described and acceptable? Were all randomized participants analyzed in the group to which they were allocated? Are reports of the study free of suggestion of selective outcome reporting? Other sources of potential bias: Were the groups similar at baseline regarding the most important prognostic indicators? Were co-interventions avoided or similar? Was the compliance acceptable in all groups? Was the timing of the outcome assessment similar in all groups?

A data extraction form was developed and piloted before data extraction. Data were extracted by 2 reviewers independently (FW and HdV). Disagreement was resolved by consensus or, if necessary, by a third reviewer (MR). Data were extracted on general study, participant, and dose characteristics (including total number of contact hours and total duration of the treatment in weeks) and treatment content (including description of treatment, treatment components, and number of disciplines). Reported outcome measures were categorized into disability, work participation and QoL. The following interventions were distinguished: PRP, no treatment, care as usual (CAU), individual physical treatment, individual psychological treatment, surgery, and multidisciplinary treatment not deďŹ ned as PRP. For each intervention the effect size (ES) was calculated by subtracting the posttreatment mean from the pretreatment mean (for each evaluation moment separately), divided by the pretreatment standard deviation (SD). If the mean change (pretreatment - posttreatment) was reported, ES was calculated by dividing the mean change by the SD of the change. If outcome was reported as a proportion (eg, work participation), ES was calculated according to Hojat et al. 9. If medians and ranges were reported 8,10-12 means and SDs were estimated 13. Because of the small sample sizes of the different interventions, the ES were corrected with Hedges J as described by Borenstein et al. 14. To analyze the inďŹ&#x201A;uence of dose variables on ES, a linear mixed effect model was applied by SPSS version 18.0 software. Analyses were applied for all interventions on each outcome category separately. The study of Hellum et al

was the only one comparing effects of PRP

with those of surgery. In that study, patients with CLBP were included, but it was not clear

Dose or content? Effectiveness of PRP | 23

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

whether they had a specific or nonspecific diagnosis. Because of the unique character of that study, the analyses were done twice: with and without this study for outcome categories disability and QoL. ES per intervention was the response variable. Predictor variables were PRP (yes/no), number of disciplines involved in the program, number of contact hours, treatment duration in weeks, type of intervention (no treatment, CAU, surgery, individual physical therapy [exercise], psychological treatment, other multidisciplinary treatment, PRP), evaluation moment (in months), age, percentage of women and risk of bias (high/low). These variables were entered one by one into the model. The variables remained in the model if the regression coefficient was significant (<0.05) or the model fit increased significantly (-2log-likelihood criterion). Random intercepts were modeled because of a better fit of the model. Patient characteristics such as duration of pain and work status were reported insufficiently to be included in the statistical model. Because augmentation of the content of the program can lead to deployment of more disciplines and more contact hours or total treatment duration, we assumed that content and dose variables could be correlated. Spearman correlations were used to analyze this assumption.

Results Results of the search A total of 2000 records was obtained from the electronic search, and 528 duplicates were removed. In total, 1472 articles were assessed on title and abstract, and 127 articles were eligible for full-text assessment. Grounds for exclusion after full-text assessment were mostly because studies were not randomized, did not fit the definition of multidisciplinary PRP, or contact hours were not described. Follow-up studies of included RCTs were not regarded as a separate study, but follow-up data were included in the analysis. Eighteen unique studies were included for risk of bias assessment. (Figure 1) Assessment of risk of bias Agreement between reviewers expressed in Cohen’s Kappa was 0.65. All disagreements between reviewers were solved by consensus. After first assessment, 12 studies were categorized as low risk of bias (score > 6)

. However, many items were answered with “unclear” because

information was not clearly reported in the articles. Two articles had no unclear answers 17,18. The e-mail address of 1 author 19 could not be retrieved. The authors of 14 articles were contacted for clariﬁcation, and answers were received from 11 of the 14 authors. After clariﬁcation with the author, 16 studies

8,10-12,15,17,18,20-28

had a score of 6 or higher (low risk of bias) and 2 studies

had a lower score and therefore a high risk of bias (Table 2).

24 | Chapter 2

19,29

Articles retrieved by database search n=2000 Duplicates n=528

Total after duplicates removed n=1472 Excluded on title/abstract selection n=1345

Total eligible for full text evaluation n=127

Excluded on full text evaluation n=109

Studies included n=18

Figure 1 Flow diagram of retrieval of studies

Dose or content? Effectiveness of PRP | 25

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

In the sensitivity analysis, item 4 of the risk of bias analysis was omitted. Because no study scored positive on this item, the total scores remained the same. Two studies had a score of 4 19,29. If the cutoff point for low risk of bias would be decreased to 5 positive items, the risk of bias of these 2 studies would still be regarded as high. Table 2 Results of risk of bias analyses Author Alaranta H Bendix AF Bendix AF Bendix AF Bendix T Jensen IB Skouen JS Vollenbroek MMR Kaapa EH Kainz B Smeets RJ Leeuw M Linton SJ Mangels M Dufour N Hellum C Meng K Roche Leboucher G

Year 1994 1995 1996 1997 2000 2001 2002 2004 2006 2006 2006 2008 2008 2009 2010 2011 2011 2011

1 ? + + + + + + + + ? + + + + + ? + +

2 ? + + + + + + + + + + + + + + + + +

3 ? ? ? ? ? + ? + ?

4 ? ? ? ? ?

5 ? + + + + + + + ? ? + + + + -

6 + + + + + + + + + + + + + + +

7 + + + ? + + + + + + -

8 + + + + + + + + + + + + + + + +

9 + + + + + + + + + + + + + + + +

10 ? ? ? ? ? + ? ? + ? ? ? + ? ? +

11 ? ? ? ? ? + + + ? ? ? ? ? ? ? ? ?

12 + + + + + + + + + + + + + + + + + +

Total + 4 6 6 7 7 9 8 10 7 4 9 7 6 7 9 6 7 7

RoB high low low low low low low low low high low low low low low low low low

RoB= Risk of Bias, +=Yes, -=No, ?=Unclear

Description of studies The included studies were published between 1995 and 2011. In total, 45 interventions were compared (no treatment/CAU: 6; surgery: 1; individual physical therapy: 11; individual psychological treatment: 1; other multidisciplinary treatment: 6; PRP: 20), and 27 comparisons between interventions were made (Table 3). The total number of contact hours for the control interventions (no treatment and CAU not included) ranged from 4.3 to 223.2 h (median=26 h, IQR=15.0 - 46.0) and for PRPs from 6.4 to 196.8 h (median=120 h, IQR=51.8 - 135.0). Disability was measured with 12 different measures (eg, RMDQ, PDI, subscales of SF-36). Six different measures were included in category of work participation and consisted percentages of people who were able to work (part time or full-time). QoL was measured with 4 different measures (Euroqol, self-constructed questions regarding general well-being).

26 | Chapter 2

Dose or content? Effectiveness of PRP | 27

1994

1995

1996

1997

2000

2001

Author

Alaranta H

Bendix AF

Bendix T

Jensen IB

Behavioral medicine

Behavior oriented physical therapy

Outpatient intensive physical training program

Active combined psycho-physical program 59

Active physical training program

Intensive multidisciplinary treatment program + functional restoration

No treatment or treatment elsewhere

Active combined psycho-physical program 45

Active physical training program

Intensive multidisciplinary treatment program + functional restoration

152

AKSELI program

Intensive multidisciplinary treatment program +functional restoration

141

Current National Type (CNT) of rehabilitation

Interventions

Table 3 Overview of characteristics and effect sizes of included studies

4.8

43.3

9.4

68.5

47.6

64.7

66.1

58.1

53.5

58.7

69.4

71.1

77.1

74.2

75.0

55.9

53.2

36.9

35.6

x 3 weeks

3 weeks

Total Wks/ duration mths

6 weeks

0 x

6 weeks

M 67.1

4 weeks

8 weeks

x 3 + 3 FU weeks

M 68.0

SD % Wom- Duration mths/yrs en pain (M/Y)

42.5 11.8

41*

40*

42*

43*

40*

41*

42*

44*

40*

40.5

40.4

Age

1.413 0.170 0.371 -0.330 0.382 1.960 0.356 1.307 0.539 0.514 0.849 0.388 1.535 0.349 0.926 0.281 1.365

12 12 12 12 12 12 pre 12 pre 12 pre 12 pre post 6 18 18 pre post 6 18 18

Workm 36 Disabilitya Workm 134 Disabilityc

pre pre pre pre pre pre

Workm 24 Disabilitya Workm 24 Disabilitya Workm 138 Disabilitya

0.747 x 0.286 1.087

4 4 4 12

pre pre pre pre

Workm 0 Disabilitya Workm 135 Disabilitya

1.072 x 0.117 x 0.362 x

4 4 4 4 4 4

pre pre pre pre pre pre

Workm 24 Disabilitya Workm 24 Disabilitya Workm 135 Disabilitya

Workp

Workp 80 Disabilityc

135 Disabilitya

126 Disabilityk

0.498 0.487 0.900 0.837 x

3 12 3 12 4

pre pre pre pre pre

Evaluation moment (mths)

52.5 Disabilityk

Total Outcome hrs domainâ&#x20AC;

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

28 | Chapter 2

2002

Skouen JS

Kaapa EH

2006

Vollenbroek MMR 2004

Author

Multidisciplinary rehabilitation

Usual Care

Controls

Light multidisciplinary treatment

Roessingh Back Rehabilitation Program

Control group/treatment as usual

Extensive multidisciplinary treatment

Cognitive behavioral therapy

Interventions 9.6

39.5

38.5

7.9

9.9

9.8

44 11.7

43.7 11.5

42.9 10.5

100

64.0

59.6

70.2

58.3

44.9

48*

72*

27.3

22.7

M 39.2

M 38.4

SD % Wom- Duration mths/yrs en pain (M/Y)

43.9 10.8

43.8

Age

8 weeks

x x

7 weeks

0 x

x x

4 weeks

0 x

4 weeks

Total Wks/ duration mths

QoLt

Workq

70 Disabilitye

Qols

x Disabilityd

Qols

63 Disabilityd

0 Workn

4.25 Workn

Workp 120 Workn

Workp 0 Disabilityc

54 Disabilityc

Total Outcome hrs domainâ&#x20AC; pre post 6 18 18 pre post 6 18 18 12 18 24 12 18 24 12 18 24 pre post 4 pre post 4 pre post 4 pre post 4 pre post 6 12 24 pre post 6 12 24 post

Evaluation moment (mths)

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 0.136 0.379 0.112 1.411 0.037 0.406 0.166 1.298 0.802 0.522 0.558 0.807 0.511 0.481 0.326 0.385 0.462 0.495 0.743 0.147 0.367 0.000 0.495 -0.114 0.228 0.420 0.467 0.607 0.527 0.609 0.419 0.419 0.377 x

Dose or content? Effectiveness of PRP | 29

2006

Author

Kainz B

Smeets RJ

100

Outpatient Rehabilitation (AR)

Cognitive behavioral treatment

Combined treatment

Waiting list

100

Outpatient Physiotherapy (EAP)

Active physical treatment

100

Medical training therapy (MTT)

Individual Physiotherapy

Interventions

9.7

9.1

7.0

40.6 11.2

37.7

58.6

4105

59.9

100

43.8

56.1

68.3

56.9

M 70.8

M 67.5

M 74.2

M 75.9

SD % Wom- Duration mths/yrs en pain (M/Y)

40.7 10.1

42.5

42.7

47.9

46.5

Age

10 Weeks

5 weeks

8 weeks

15 weeks

6 to 8 weeks

Total Wks/ duration mths

0 Disabilityd

69.5 Disabilityd

26.5 Disabilityd

43.5 Disabilityd

76.4 Disabilityl Workr Qolv

46 Disabilityl Workr Qolv

QoLt 27.4 Disabilityl Workr Qolv

Workq

10 Disabilitye

Total Outcome hrs domainâ&#x20AC;

pre post 6 12 24 pre post 6 12 24 post pre 6 pre 6 pre post 6 pre 6 pre 6 pre post 6 pre 6 pre 6 pre post 6 pre post 6 12 pre post 6 12 pre post 6 12 pre post 6 12

Evaluation moment (mths)

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

0.186 0.491 0.448 0.377 0.283 0.354 0.318 0.035 x 0.347 0.417 0.574 0.485 0.255 0.543 0.560 0.480 0.335 0.293 0.610 0.460 0.714 0.330 0.334 0.791 0.357 0.360 0.533 0.354 0.341 0.020 x x

2008

2009

2010

2011

Author

Leeuw

Linton SJ

30 | Chapter 2

Mangels M

Dufour N

Hellum C

Rehabilitation

143

Intensive individually therapist-assisted back muscle strengthening exercises

119

Multidisciplinary rehabilitation + booster sessions

129

113

Multidisciplinary rehabilitation

Group based multidisciplinary biopsychosocial rehabilitation program

21 25 131

Expose TAU (treatment as usual) Waiting list TAU (treatment as usual) Traditional orthopedic rehabilitation treatment

9.0

40.8

49.7 50.3

46.6 53.4

7.1

48.3 15.8

49.5

46 9.9 49 7.3 48.7 14.7

43 44.21 9.54

58.6

55.9

56.6

75.6

78.8

62 48 78.6

44.2

52.4

83 76 x

9.00

8.13

% >Y % >Y x

x x x

Y 8.94

Y 9.95

SD % Wom- Duration mths/yrs en pain (M/Y)

Operant graded activity

Age

42 46.45 9.33

Exposure in vivo treatment

Interventions

QoLu

223.2 Disabilityg

QoLu

13 Disabilityi 0 Disabilityi 196.8 Disabilityg

26 Disabilityh,i

16 Disabilityh,i

Total Outcome hrs domainâ&#x20AC;

3 to 5 weeks

12 weeks

60 Disabilityb,e

Workm

46 Disabilityb,r

Workm

87 Disabilityb,r

QoLu

4 + 12 weeks + 223.2 Disabilityg FU FU mths

4 weeks

x x x x 3 weeks

x x

Total Wks/ duration mths

1.492 1.231 0.738 0.719 0.473 0.006 0.301 0.333 0.419 0.268 0.487 0.344 0.550 0.396 0.370 0.306 0.490 0.511 0.602 0.515 0.455 0.549 0.231 0.391 0.359 0.289 0.211 0.236 0.041 0.362 1.341 1.144 0.726 1.379

12 pre post 12 pre post 12 pre post 12 pre post 12 pre post 12 pre post 6 12 24 pre post 24 pre post 6 12 24 pre post 24 pre 1.5 3 6 12

pre post 6 pre post 6 pre post pre post pre post

Evaluation moment (mths)

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Dose or content? Effectiveness of PRP | 31 68 64

Active individual therapy

185

Traditional back school program (CAU)

Functional restoration program

197

New biopsychosocial back school program

Surgery

Interventions

38.7

40.8

49.5

50.2

41.1

Age

6.1

7.4

7.7

7.6

7.1

37.5

32.4

65.2

46.5

39.1

51.5

%>5 Y

SD % Wom- Duration mths/yrs en pain (M/Y)

5 weeks

24.9 mean (5.2) days

24.3 mean (4.8) days

x x

Total Wks/ duration mths

0.420 1.747 1.037 0.652 0.751

12 12 12 12 12 150 Disabilityf Worko 15 Disabilityf Worko

pre pre pre pre

0.377 0.569

12 pre 6

6.4 Disabilityj

QoLs

x Disabilityb,e

QoLs

5.5 Disabilityj

ES 1.527 0.894 0.671 0.766 0.862 1.054 1.121 1.787 1.870 2.076 2.131 0.957 1.320 1.254 1.221 1.254 0.702

Evaluation moment (mths) 24 1.5 3 6 12 24 pre 1.5 3 6 12 24 1.5 3 6 12 24 pre 6

Total Outcome hrs domain†

*=median, x=no data available. ES: Effect size. FU: Follow up, QoL: Quality of Life. † Measurement instrument used for operationalization of outcome domain. a: Function. b: SF36 physical component. c: SF36 physical functioning + SF36 role physical. d: RMDQ. e: Oswestry. f: Pain impact. g: PDI. h: PCS. i: QBPDS. j: Physical activity (freiburger). k: Million Index. l: Funktionsfahigkeit. m: work readiness. n: Return to work. o: percentage full time at work. p: work participation. q: subjective work capacity. r:arbeidsfähigkeit. s: Euroqol. t: general wellbeing. u: life satisfaction questionnaire. v: somatische gesundheit.

2011

Meng K

Roche Lebouche G

Author

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

ES for disability ranged from -0.33 to 2.13 (median 0.36, IQR 0.23 - 0.67) for control interventions and from 0.31 to 1.96 (median 0.54, IQR 0.43 - 1.04) for PRPs. ES for work participation ranged from 0.04 to 1.41 (median 0.38, IQR 0.32 - 0.54) for controls and for PRPs from 0.23 to 1.54 (median 0.54, IQR 0.38 - 0.98). QoL ranged from -0.11 to 1.32 (median 0.56, IQR 0.34 - 1.24) for controls and from 0.15 to 1.05 (median 0.55, IQR 0.43 - 0.81) for PRPs (Table 4). In one study 15

, the control intervention (surgery) was superior to PRP. That study was also responsible for the

highest ES for the control intervention for disability (2.13) and QoL (1.25). Table 4 Ranges of ES per outcome category for PRPs and control interventions Outcome Disability Work Participation Quality of life

PRP Controls Including study of Hellum et al. 0.306 to 1.960 -0.330 to 2.131 0.231 to 1.535 0.035 to 1.411 0.147 to 1.054 -0.114 to 1.320

PRP Controls Excluding study of Hellum et al. 0.306 to 1.960 -0.330 to 1.307 0.231 to 1.535 0.035 to 1.411 0.147 to 0.610 -0.114 to 0.574

ES: Effect size. PRP: Pain Rehabilitation Program.

Mixed effect modeling Disability Type of intervention, evaluation moment, and percentage of women significantly contributed to the regression equation. In this analysis, the reference category was effect of PRP immediately after treatment, with 0% women as participants. The negative coefficients indicate that control interventions had a smaller ES than PRP (β=0.501), except surgery, which had a significantly larger ES (β=0.748), indicating that the ES of surgery was 0.501 + 0.748= 1.249. The coefficient of 12 to 24 months for evaluation moment indicates that the ES is 0.407 points larger at 12 to 24 months than immediately after intervention (reference category) (Table 5). The coefficient of 0.002 for percentage of women indicates that for every 1% women extra as participants, the ES increases with 0.002. Performing the analyses without the study of Hellum et al

, the

reference category was effect of PRP with 6 weeks’ duration after treatment, with 0% women as participants. A significant contribution of type of intervention, evaluation moment, percentage of women, and duration in weeks was found (Table 5). All control interventions had a smaller ES than PRP (β=0.429) ranging from -0.714 to -0.154. Evaluation moment between 12 and 24 months after the study had a regression coefficient of 0.447, indicating that the ES at 12 to 24 months is 0.447 larger than immediately after treatment. The coefficient of 0.003 for percentage of women indicates that for every 1% women extra as participants, the ES increases by 0.003. The coefficient of -0.037 for duration in weeks indicates that for each additional week above 6 weeks, the ES decreases with 0.037.

32 | Chapter 2

Table 5 Results of mixed effect model analyses for the different outcome variables, in and excluding data from the study of Hellum et al (2011) Outcome Disability a

Variable Intercept r No treatment CAU Surgery Individual therapy/exercise Psychological treatment Other multidisciplinary treatment 1.5 – 6 months FU 12 months FU 12-24 months FU Percentage women Work Participation b Intercept r Number of disciplines Quality of life c PRP r CAU Surgery Individual therapy/exercise Other multidisciplinary treatment Analyses excluding study of Hellum et al.(2011) Disability d Intercept r Individual therapy/exercise Psychological treatment Other multidisciplinary treatment 1.5 – 6 months FU 12 months FU 12-24 months FU Percentage women Duration in weeks Quality of life e Intercept r CAU Individual therapy/exercise Other multidisciplinary treatment Age (centered at 45 years)

β 0.501 -0.617 -0.719 0.748 -0.142 -0.496 -0.294 0.173 0.182 0.407 0.002 0.475 0.088 0.538 -0.241 0.372 -0.011 -0.147

SE β 0.167 0.383 0.176 0.218 0.108 0.176 0.198 0.116 0.142 0.131 0.001 0.116 0.025 0.104 0.112 0.071 0.096 0.099

p 0.004 0.109 <0.001 0.001 0.188 0.006 0.140 0.139 0.201 0.002 0.086 0.001 0.001 0.006 0.041 <0.001 0.912 0.149

0.429 -0.154 -0.414 -0.714 0.174 0.188 0.447 0.003 -0.037 0.417 -0.231 0.040 -0.160 0.023

0.160 0.120 0.189 0.332 0.152 0.158 0.171 0.002 0.018 0.033 0.087 0.058 0.075 0.007

0.009 0.206 0.031 0.034 0.255 0.237 0.010 0.108 0.038 <0.001 0.018 0.496 0.050 0.006

a: Reference category: effect of PRP post treatment, with 0% women as participants. b: Reference category: effect of treatment with 0 disciplines. c: Reference category: PRP. d: Reference category: effect of PRP with 6 weeks duration, post treatment, with 0% women as participants. e: Reference category: effect of PRP for patients at age of 45 years. β: Regression coefficient. SE: standard error β. r: random. CAU: Care as usual. PRP: Pain rehabilitation Program. FU: follow up/evaluation moment. *Although β not significant model fit increased significantly.

Work Participation The number of disciplines signiﬁcantly contributed to the regression equation. All other variables entered into the model did not contribute signiﬁcantly. Table 5 shows that every discipline added

Dose or content? Effectiveness of PRP | 33

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

to the intervention, compared to the reference category of treatment with 0 disciplines, will increase the ES with 0.088. QoL To predict the ES for outcome QoL, only type of intervention contributed significantly to the regression equation. Table 5 shows an effect of 0.417 for PRP (reference category). The negative coefficients indicate a smaller effect for control interventions except for the intervention surgery (β=0.372), indicating an effect for surgery of 0.538 + 0.372= 0.910. Type of intervention and age contributed significantly to the regression equation when performing the analysis without the data of Hellum et al. 15 (Table 5). For QoL, the effect of PRP for patients at age of 45 years is 0.417; CAU and other multidisciplinary treatment for patients at age 45 years had smaller effect; only individual therapy for patients aged 45 years showed a 0.040 (not significant) larger effect compared to PRP. Contact hours and number of disciplines were correlated (r=0.57, p<.01). The number of disciplines involved and contact hours were larger in PRP (respectively, median 3 and 87.0) compared to other interventions (respectively, median 1 and 27.4). No other significant associations were found between contact hours, total duration of the treatment and number of disciplines involved.

Discussion To our knowledge, this is the first systematic review analyzing the relationship between dose and effect of multidisciplinary PRPs. Twelve out of the 18 studies were unique; they were not included in the reviews of Guzman et al. 2 and Van Geen et al. 3. Some studies were published after these 2 reviews; other studies were not included in the current review because of different inclusion and exclusion criteria. The current review shows a wide variety in contact hours and ES for all interventions treating patients with CLBP. Within the interventions defined as multidisciplinary PRP, substantial variation was observed in ES, contact hours, content, total duration and number of disciplines. Analyzing the influence of predictor variables on effect of PRP on disability demonstrated that the type of intervention, evaluation moment, percentage of women, and the total duration in weeks influenced that effect. The number of disciplines involved influenced effects on work participation, and QoL was additionally influenced by the type of intervention and age of participants. PRP and surgery had the strongest effects on the outcome disability. PRP had the strongest effect on the outcome work participation. Surgery, as described by Hellum et al. 15

, was the only intervention that was superior to PRP. Surprisingly, no other studies were obtained

from the literature search comparing surgery to PRP for patients with CLBP.

34 | Chapter 2

The review of Guzman et al. 2 provided evidence that intensive PRP of more than 100 h of therapy produced greater improvements in function than PRP of less than 30 h of therapy. The current review provides a wide variety of contact hours in PRPs, with a range of 6.4 to 196.8 h, with larger ES than most of the control interventions. This shows that there are also PRPs producing improvements within the range of 30 to 100 h of therapy. Therapies within this range of contact hours were not included in the review of Guzman et al.

15,17,18,22,26

. It is unknown whether these

PRPs are as effective as the PRPs of more than 100 h. This study focused on dose aspects of treatment. The results of the analyses of this review showed that type of intervention, evaluation moment, percentage of women, number of disciplines, duration in weeks, and the age of the participants inďŹ&#x201A;uenced the effect of interventions for patients with CLBP. However, because contact hours and number of disciplines were interrelated for all interventions, we are not able to draw conclusions about the relationship between independent components and effects. In most cases, the number of disciplines involved depends on the aim and content of the treatment. However, we demonstrated that it is currently not possible to disentangle dose and content, which supports the importance of paying attention to both components for analyzing the effectiveness of PRPs. Regarding the evaluation moment, it is not clear whether the effect increases over time as a result of the intervention or as a result of natural recovery. PRPs were on average more effective than the other interventions, but all included studies analyzed PRPs with different doses and contents. Because studies were not designed to investigate dose variables independently, we were not able to analyze the relationship between dose and effects in PRP alone. Consequently, research on the effectiveness of PRPs may have been biased because it has been focused solely on content and neglected dose of treatment. All included studies were similar with regard to participants and study design. Subgroups were made to analyze outcome domains and follow-up time separately. Therefore, we judged the studies to be sufďŹ ciently homogeneous to perform a meta-analysis. The heterogeneity of the content of PRPs and outcomes is a strength of the study when focusing on the clinical use of the information about dose aspects of PRPs, but is also a weakness regarding the methodological strength of this review. Study limitations RCTs included in this study were aimed to analyze the effect of PRP. No RCTs were obtained analyzing the relationship between dose and effect. Therefore, this study provides only indirect conclusions, which is a limitation of the study. Another limitation is the statistical heterogeneity, which was almost inevitable because of the wide variety of PRPs in terms of content, outcome measures, dose and so on. Fortunately, more and more guidelines and recommendations

are

being published to encourage the homogeneity in chronic pain research. Another shortcoming

Dose or content? Effectiveness of PRP | 35

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

of this study is that the conclusions are based on a small sample of RCTs concerning PRPs, resulting in a limited power. This small sample was the result of selection criteria for studies to provide similar interventions and study designs and from the lack of consistent reporting of dose variables. The small sample resulted in categories of evaluation moment with different intervals and ES from different interventions, which limits the interpretability of these results. The coefﬁcients of the categories seem to indicate a much larger effect at 12 to 24 months (β= 0.460, β= 0.485) compared to the other evaluation moments. However, this category contains 12 months compared to 1 or 4.5 months interval in the other categories. An increase of ES during the different evaluation moments exists, but it ﬂattens. Risk of bias was assessed with the tool recommended by the Cochrane Back Review Group. The choice to use this risk of bias scale may be debated because of the ongoing discussion about the blinding of patients and therapists in nondrug trials. Although blinding is recommended to assess performance bias in RCTs

patients were blinded in only 2 studies, and care providers were blinded in none of the studies 23,27

. Low risk of bias was deﬁned as scoring positively for 6 or more out of 12 items. Blinding of

patients and care providers covers 2 out of 12 items. Consequently, failure to blind care providers and/or patients does not necessarily lead to a low quality rating. To fulfill the criteria of risk of bias, there are methods regarding blinding of patients to interventions in nondrug trials, such as evaluating treatment credibility, patient expectations and patient treatment preference. Blinding of care providers is more difficult in nondrug trials; nevertheless, knowledge of which intervention a patient received can influence treatment effects. Because of this risk of bias, the extension of the CONSORT statement for nondrug trials suggests to report the interpretation of the results regarding lack of or partial blinding in RCTs 32. Another option to control for lack of blinding is measuring care providers’ allegiance and expectations towards dose and content of the treatment protocol. To analyze the effect of the risk of bias analysis on the outcome of this review, we performed a sensitivity analysis. Taking the item blinding of care providers out of the risk of bias analysis would not have made a difference in the overall judgment of high or low risk of bias, or of the final conclusions of this review. Other quality assessment tools to analyze the risk of bias are available for trials in which it is difficult to blind participants or therapists 33,34. Another limitation is the language restriction. It is possible that other high-quality RCTs exist written in other languages that could provide relevant information regarding aspects of dose. Conclusion This review supports the evidence that PRPs are effective and superior to most other interventions in treating patients with CLBP on disability and work participation. Within PRPs, a wide variety of contents and doses are observed. Dose aspects of PRPs are reported and analyzed insufficiently. On the basis of the current literature, it is unknown how many hours, months, or weeks are needed to achieve the best effects. It is also not possible to draw conclusions about the influences of dose on effect, which could support the clinical decisions for dose of PRP for individual patients. To

36 | Chapter 2

further improve PRPs, there is need to report dose aspects (total duration of treatment, number of contact hours, number of disciplines) in analyzing effects of PRPs and to study dosage (duration and intensity) of PRPs. Additionally, qualitative studies aiming to identify relevant aspects of dosage of PRPs could provide more insight in the importance and clinical relevance of dose.

Acknowledgement We certify that no party having a direct interest in the results of the research supporting this article has or will confer a beneďŹ t on us or on any organization with which we are associated.

Dose or content? Effectiveness of PRP | 37

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1.

Flor H, Fydrich T, Turk DC. Efﬁcacy of multidisciplinary pain treatment centers: A meta-analytic review. Pain. 1992;49(2):221-230.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

van Geen JW, Edelaar MJ, Janssen M, van Eijk JT. The long-term effect of multidisciplinary back training: A systematic review. Spine (Phila Pa 1976). 2007;32(2):249-255.

Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. european guidelines for the management of chronic nonspeciﬁc low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

Sanders SH, Harden RN, Vicente PJ. Evidence-based clinical practice guidelines for interdisciplinary rehabilitation of chronic nonmalignant pain syndrome patients. Pain Pract. 2005;5(4):303-315.

van Hooff ML, van der Merwe JD, O’Dowd J, et al. Daily functioning and self-management in patients with chronic low back pain after an intensive cognitive behavioral programme for pain management. Eur Spine J. 2010;19(9):1517-1526.

Waddell G. 1987 volvo award in clinical sciences. A new clinical model for the treatment of low-back pain. Spine (Phila Pa 1976). 1987;12(7):632-644.

Bendix AF, Bendix T, Vaegter K, Lund C, Frolund L, Holm L. Multidisciplinary intensive treatment for chronic low back pain: A randomized, prospective study. Cleve Clin J Med. 1996;63(1):62-69.

Hojat M, Xu G. A visitor’s guide to effect sizes: Statistical signiﬁcance versus practical (clinical) importance of research ﬁndings. Adv Health Sci Educ Theory Pract. 2004;9(3):241-249.

10.

Bendix AF, Bendix T, Lund C, Kirkbak S, Ostenfeld S. Comparison of three intensive programs for chronic low back pain patients: A prospective, randomized, observer-blinded study with one-year follow-up. Scand J Rehabil Med. 1997;29(2):81-89.

11.

Bendix AF, Bendix T, Ostenfeld S, Bush E, Andersen. Active treatment programs for patients with chronic low back pain: A prospective, randomized, observer-blinded study. Eur Spine J. 1995;4(3):148-152.

12.

Bendix T, Bendix A, Labriola M, Haestrup C, Ebbehoj N. Functional restoration versus outpatient physical training in chronic low back pain: A randomized comparative study. Spine (Phila Pa 1976). 2000;25(19):2494-2500.

13.

Hozo SP, Djulbego. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical research methodology. 2005;5:13.

14.

Borenstein M. Introduction to meta-analysis. Chichester, UK. Wiley; 2009:421.

15.

Hellum C, Johnsen LG, Storheim K, et al. Surgery with disc prosthesis versus rehabilitation in patients with low back pain and degenerative disc: Two year follow-up of randomised study. BMJ. 2011;342:d2786.

16.

Furlan AD, Pennick V, Bombardier C, van Tulder M, Editorial Board, Cochrane Back Review Group. 2009 updated method guidelines for systematic reviews in the cochrane back review group. Spine (Phila Pa 1976). 2009;34(18):1929-1941.

17.

38 | Chapter 2

18.

Vollenbroek-Hutten MM, Hermens HJ, Wever D, Gorter M, Rinket J, Ijzerman MJ. Differences in outcome of a multidisciplinary treatment between subgroups of chronic low back pain patients deﬁned using two multiaxial assessment instruments: The multidimensional pain inventory and lumbar dynamometry. Clin Rehabil. 2004;18(5):566-579.

19.

Alaranta H, Rytokoski U, Rissanen A, et al. Intensive physical and psychosocial training program for patients with chronic low back pain. A controlled clinical trial. Spine (Phila Pa 1976). 1994;19(12):13391349.

20.

Jensen IB, Bergstrom G, Ljungquist T, Bodin L, Nygren AL. A randomized controlled component analysis of a behavioral medicine rehabilitation program for chronic spinal pain: Are the effects dependent on gender? Pain. 2001;91(1-2):65-78.

21.

Skouen JS, Grasdal AL, Haldorsen EM, Ursin H. Relative cost-effectiveness of extensive and light multidisciplinary treatment programs versus treatment as usual for patients with chronic low back pain on long-term sick leave: Randomized controlled study. Spine (Phila Pa 1976). 2002;27(9):901-9; discussion 909-10.

22.

Kaapa EH, Frantsi K, Sarna S, Malmivaara A. Multidisciplinary group rehabilitation versus individual physiotherapy for chronic nonspeciﬁc low back pain: A randomized trial. Spine (Phila Pa 1976). 2006;31(4):371-376.

23.

Leeuw M, Goossens ME, van Breukelen GJ, et al. Exposure in vivo versus operant graded activity in chronic low back pain patients: Results of a randomized controlled trial. Pain. 2008;138(1):192-207.

24.

Linton SJ, Boersma K, Jansson M, Overmeer T, Lindblom K, Vlaeyen JW. A randomized controlled trial of exposure in vivo for patients with spinal pain reporting fear of work-related activities. Eur J Pain. 2008;12(6):722-730.

25.

Mangels M, Schwarz S, Worringen U, Holme M, Rief W. Evaluation of a behavioral-medical inpatient rehabilitation treatment including booster sessions: A randomized controlled study. Clin J Pain. 2009;25(5):356-364.

26.

27.

Meng K, Seekatz B, Roband H, Worringen U, Vogel H, Faller H. Intermediate and long-term effects of a standardized back school for inpatient orthopedic rehabilitation on illness knowledge and selfmanagement behaviors: A randomized controlled trial. Clin J Pain. 2011;27(3):248-257.

28.

Roche-Leboucher G, Petit-Lemanac’h A, Bontoux L, et al. Multidisciplinary intensive functional restoration versus outpatient active physiotherapy in chronic low back pain: A randomized controlled trial. Spine (Phila Pa 1976). 2011;36(26):2235-2242.

29.

30.

Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005;113(1-2):9-19.

31.

Higgins JP, Altman DG, Gotzsche PC, et al. The cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928.

32.

Boutron I, Moher D, Altman DG, Schulz KF, Ravaud P, CONSORT Group. Extending the CONSORT statement to randomized trials of nonpharmacologic treatment: Explanation and elaboration. Ann Intern Med. 2008;148(4):295-309. Dose or content? Effectiveness of PRP | 39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

33.

Yates SL, Morley S, Eccleston C, de C Williams AC. A scale for rating the quality of psychological trials for pain. Pain. 2005;117(3):314-325.

34.

Boutron I, Moher D, Tugwell P, et al. A checklist to evaluate a report of a nonpharmacological trial (CLEAR NPT) was developed using consensus. J Clin Epidemiol. 2005;58(12):1233-1240.

40 | Chapter 2

Dose or content? Effectiveness of PRP | 41

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Fran Fr anka ka PC Wa Wate ters te r ch rs choo oot, t Pie t, ete terr U Di D jk kstra sttra a PhD hD, Jan Ja n HB Gee eert rtze rt z n MD ze MD,, Ph PhD, D, Mic ichi hiel hi e F Ren el e em eman an n PhD hD.

Intt J Re In Reha habi ha bilil Re Ress 20 2015 15, 5 38 (1) 1): 34 34-9 -9. -9

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Abstract The aim of this study was to analyze the course of reduction of disability during a pain rehabilitation program (PRP) and factors influencing this course. A prospective cohort study was carried out. All patients with chronic musculoskeletal pain treated in a PRP between March 2010 and December 2010 were eligible for this study. All patients were treated at a University-based rehabilitation center and received an outpatient multidisciplinary PRP. Main outcome measures, Pain Disability Index (PDI), and average pain measured with a numeric rating scale, were measured every 2 weeks during the PRP. To analyze the course of disability, a linear mixed-effect model was applied. One hundred twenty-eight patients participated in the study, of which 20% dropped out during the PRP. Initial PDI (β= 0.8), treatment week (β= -0.2), treatment week squared (β=0.03), average pain (β= 2.3), and interaction between initial PDI and treatment week (β= -0.02) influenced the course of disability during PRP. Disability reduces during the PRP. Initial PDI, treatment week, average pain, and interaction between initial PDI and treatment week influence the course of disability reduction during the PRP. These results could aid in predicting the required duration of a PRP at the start. Keywords Chronic pain, disability, Pain Disability Index, Patient Care Team, Rehabilitation

44 | Chapter 3

Introduction On the basis of pre and post measurements, it is known that pain rehabilitation programs (PRPs) are effective in disability reduction in patients with chronic musculoskeletal pain (CMP)

1-7

However, the course of disability reduction is unknown. Understanding the course of disability reduction could aid in determining the optimum duration of PRPs. This understanding would beneﬁt patients, clinicians, insurance companies and health care providers because it prevents overtreatment or undertreatment and contributes towards better use of (public) recourses. A wide diversity of content, composition, and duration of PRPs exist 3,4. In one systematic review, differences in duration were unrelated to differences in outcomes 3. However, in another review it was reported that multidisciplinary PRP of more than 100 h was superior to monodisciplinary treatment, and that multidisciplinary PRP of less than 30 h was not superior to monodisciplinary treatment. Although the authors focused on the content of different PRPs and the review was designed to assess the effect of PRP, they concluded that PRPs of more than 100 h were more effective. The conclusions of these reviews were used to establish guidelines and clinical practice 8,9

. However, these conclusions were based only on pre-post assessments. To our knowledge,

no studies have analyzed the course of disability reduction during the programs to establish the optimum duration of a PRP. In addition, it is unknown whether the improvements gained with intensive PRPs are worth the expenses 4. Theoretically, if 90% of the improvements are observed in the first half of the PRP, it may be debated whether these 100 h or more, as proposed by Guzman et al 4, are needed. Because no previous research is available into this specific aspect of PRPs, we hypothesized that there would be several possible courses of reduction of disability (Fig. 1). Line A assumes that the course of disability has a steeper slope at the end of the PRP. This implies that disability decreases most in the last part of the PRP. In contrast, line B assumes a steeper slope at the first half of the PRP. This Implies that disability decreases most in the first part of the PRP. Line C implies a linear decrease of disability, and an alternating pattern of decrease and increase during the PRP is hypothesized in line D. The aim of this study was to analyze the course of reduction of disability during PRP and to analyze factors influencing this course.

Course of disability reduction during a PRP | 45

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Mean PDI

C D

X Treatment weeks

Figure 1 Hypothesized courses of reduction of disability during a hypothetical Pain Rehabilitation Program Line A: the decrease of disability has a steeper slope at the second half of PRP. Line B: the decrease of disability is steeper slope at the ďŹ rst half of PRP. Line C: a linear decrease of disability during PRP. Line D: an alternating pattern of decrease and increase of disability during PRP.

Patients and methods Study design and setting A prospective cohort study was carried out with patients with CMP admitted to the outpatient multidisciplinary PRP of the Center for Rehabilitation of the University Medical Center Groningen (UMCG). The primary aim of this PRP, on the basis of cognitive behavioral principles, was to decrease pain-related disability. The rehabilitation team consists of physicians, occupational therapists, physiotherapists and psychologists. PRP consists of education on differences between nociceptive pain and chronic pain. Patients are counseled to reďŹ&#x201A;ect on their own pain management strategies (avoiding pain) and how these strategies could be changed into more healthy management strategies (coping with pain, alternation between physical activity and rest, and to gradually increase functioning). Each patient sets individual treatment goals to create a meaningful life

46 | Chapter 3

despite the pain. To achieve these goals, each professional applies specific techniques. Physical therapists apply exercise programs and sports activities, improve patients’ confidence in movement, and reduce pain related fear. Occupational therapists assess current activities and patterns in daily living and educate patients on how these activities can be changed into healthy activity levels and patterns. The psychologists coaches patients in understanding and dealing with the social and emotional impact of pain in daily life, pain beliefs, and barriers for behavior change, and coach patients on how to cope with pain. The rehabilitation physicians are responsible for the medical diagnosis and interventions, and the overall treatment plan. During intake, the rehabilitation team decides whether the patient can be admitted for PRP. PRP consisted of occupational therapy sessions, 30 min, two times a week, physiotherapy sessions, 30-60 min, two times a week, and psychology sessions, 60 min, once a week. The session frequency of occupational therapy and physiotherapy reduced to once a week during the second half of PRP to encourage patients self management among patients. After intake, the rehabilitation physician proposed a duration of 8, 12, 16 or 20 weeks of PRP to the patient on the bases of assessment of the complexity of the physical, social and personal situation of the patient, motivation, and ability to change behavior. The duration in weeks of the PRP could be adapted (increased or reduced) depending on clinical progress. Duration could be increased when additional decrease in disability was expected by lengthening the program. Duration was reduced when treatment goals were achieved earlier than expected, when the patient showed continued lack of progress, or stops the treatment because it did not match patients’ expectations or for reasons not related to the program. Patients were included for PRP when they were 18 years or more of age, had CMP for more than 3 months, experienced disabilities because of CMP, were not attending any other type of treatment, were willing to participate in a PRP, and had signed an informed consent form. Patients with severe comorbidities, interfering with PRP, such as heart failure, rheumatoid arthritis, psychiatric disorders, and participants with an indication for surgery were not included for the PRP. All patients treated in the PRP of the UMCG from March 2010 until December 2010 were eligible for the study. At baseline, patient characteristics (age, sex, marital status and employment status) were recorded. The first visit to the PRP (T1) and every uneven treatment week during the program (T2= week 3, T3= week 5 etc), patients filled in the Pain Disability Index (PDI) and a numeric rating scale (NRS) for average pain in the previous week (0 = no pain, 10= unbearable pain). The last assessment was performed during the last visit. Patients were blinded for previous scores. If patients could not be measured at the proposed week, they filled in the questionnaire the next visit.

Course of disability reduction during a PRP | 47

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Main outcome The PDI measures self reported disability on seven domains: family/home responsibilities, recreation, social activity, occupation, sexual behavior, self-care, and life-support activity. The patient scores the amount of interference in daily life because of pain for each domain (0= no interference, 10= total interference). Total PDI is calculated by adding the scores of all domains [range 0 (no disability) to 70 (extreme disability)]. Internal consistency for the PDI is high (Cronbach’s α=0.86) 10. Construct validity is supported by a multiple regression analyses showing nine variables to predict the PDI scores (R=0.74) 10. Test-retest reliability (1-week interval) is good (intraclass correlation coefﬁcient=0.91) 11. The minimal clinically important change is 8.5 point 12. A maximum of two missing items was accepted for the PDI. If more than two items were missing, a sum score was not calculated. When one or two items of the PDI were missing, the mean score of the other items of that patient was calculated. The total score was calculated by replacing the missing item by this calculated mean. As patients provided written permission to use their clinical data and data were gathered during care as usual, in the Netherlands, approval from a Medical Ethics Committee is not needed. Statistical analysis Descriptive statistics were computed for patient characteristics and outcome variables. Means and SDs were used for PDI and pain scores. An independent sample t-test was used to compare means of initial PDI of completers (participants who completed the program) versus noncompleters (participants who dropped out before the program was ended). Q-Q plots were constructed to check the assumptions of normal distribution of initial PDI scores. The duration of PRP was categorized according to the total number of weeks that a patient received treatment. Categories were: ≤ 8, 9 ≤ 12, 13 to ≤ 16 and ≥17 weeks. To analyze the course of disability during PRP and factors inﬂuencing this course, a linear mixed-effect model was applied. A linear mixed-model analysis corrects for autocorrelation in data, which occurs in repeated measurements over time within individuals. All available data were taken into account in the analysis, including those of patients with incomplete datasets

13,14

. Outcomes on the PDI were entered as response variable.

Before this analysis, the change in PDI scores during PRP was assessed in a graph to check linearity for applying the appropriate regression model. Possible predictor variables of the course were patient characteristics (age, sex, marital status, and employment status), treatment week (=week of measurement), duration of PRP (in categories), initial PDI score, pain characteristics (average pain in the week before the PDI assessment, pain duration, and pain localization) and completing PRP (yes, no). These variables were entered stepwise in the regression model. Assessment of the graph showed that the change of PDI during PRP allowed a quadratic (treatment week x treatment week) or cubic model (treatment week x treatment week x treatment week). These models were explored by entering variable treatment week2 and treatment week3 in the regression

48 | Chapter 3

equation. Entered predictors remained in the model if the model fit (-log-likelihood) increased significantly or if the regression coefficient was significant. After the main effects were established the covariance structure best applicable to the data was explored, again using the model fit as criterion. After determining the covariance structure, random intercept and slopes were explored using the model fit. Interactions between predictors were explored if main effects were significant (P≤0.05). Data were analyzed using SPSS (version 18.0; SPSS Inc., Chicago, Illinois, USA).

3 Results A total of 128 patients participated in the study. Mean age of the patients was 41.7 (SD 11.8, range 19-68) years. CMP was categorized as chronic back pain (55%), chronic neck pain (19%) and other types of chronic pain (26%), such as widespread pain, non-specific pain of lower extremities (knee, foot) etc. Patients attended PRP for an average duration of 12.5 (SD 4.0, range 3-23) weeks and received on average 38 (SD 17.1, range 5.5-89.8) contact hours. Thirty-eight percent (n=49) of the patients received the proposed amount of treatment weeks (range ± 1 week). Thirty-seven percent (n=47) received less (including noncompleters) and 13% (n=17) received more than amount of treatment weeks proposed by the rehabilitation physician. The proposed duration was not reported in 12% (n=15) of the cases. Twenty-six (20%) of the 128 patients dropped out during the PRP; they did not complete the program for lack of effect (2%, n=2), unrealistic expectations of the program (3%, n=4), wanting additional diagnostic procedures (2%, n=2), because they believed that the benefits of the program did not weigh against efforts (3%, n=4), treatment was too demanding (3%, n=4) or other reasons (8%, n=10). The majority of noncompleters stopped during the first 8 weeks of PRP. Because of different treatment durations and dropout, the total number of participants decreased from the first to the 27th treatment week (Fig. 2). Baseline characteristics (n= 128) are shown in Table 1. Noncompleters and completers significantly differed on age, marital status, and employment status. The difference in duration of pain of completers versus noncompleters was not significant. The mean PDI for all patients at T1 (n = 125) was 34.2 (SD 12.0). Mean NRS

average pain

was 5.6

(SD 1.9). The mean initial PDI was signiﬁcantly (P=0.013) lower for completers compared to noncompleters [mean difference= 6.6, conﬁdence interval (CI) 1.4-11.7] (Table 2).

Course of disability reduction during a PRP | 49

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

50 | Chapter 3 22 (28)

10 (2)

28 (11)

50 (20) 23 (9) 27 (11) 3.0 (1.5-10.0) 23 (12)

60 (31) 17 (9) 23 (12) 3.0 (1.5-7.0) 20 (3)

53 (8) 27 (4) 20 (3) 2.8 (2.1-4.5)

26 (26)

54 (55) 21 (21) 26 (26) 3.0 (1.5-7.0)

5 (5) 8 (8) 76 (77) 9 (9) 3 (3)

53 (11) 14 (3) 33 (7) 5.0 (2.0-15.0)

6 (3) 6 (3) 75 (39) 9 (5) 4 (2)

7 (1) 7 (1) 79 (12) 7 (1) 0 (0)

55 (70) 19 (25) 26 (33) 3.0 (1.5-8.0)

2 (1) 10 (4) 73 (29) 13 (5) 2 (1)

8 (2)

58 (15) 15 (4) 27 (7) 4.5 (2.0-18.8)

19 (5) 0 (0) 62 (16) 19 (5) 0 (0)

0.208

0.836

0.019

0.045 0.148 0.020

P value*

24 (5) 0 (0) 62 (13) 14 (3) 0 (0)

Noncompleters (n= 26) 45.8(11.6) 85 (22) 92 (24) 0 (0)

8 (10) 6 (8) 73 (93) 11 (14) 2 (3)

≥17 (n= 15) 37.9 (11.7) 73 (11) 47(7) 0 (0)

Completers (n= 102) 40.6(11.7) 71 (72) 69 (70) 2 (2)

13 ≤ 16 (n= 52) 42.0 (10.2) 56 (29) 75(39) 4 (2)

≤8 (n= 21) 45.7 (12.3) 91 (19) 95 (20) 0 (0)

(n=128) 41.7 (11.8) 73 (94) 73 (94) 2 (2)

9 ≤ 12 (n= 40) 40.5 (13.3) 88 (35) 70(28) 0 (0)

Completers versus noncompleters

Sorted by duration PRP (weeks)

All patients/ durations

PRP, pain rehabilitation program. * P value of difference between completers and noncompleters.

Missing [n (%)]

Age [mean (SD)] Sex (female) [n (%)] Married [n (%)] Missing Employment status [n (%)] Houseman/-wife Attending education Working or sick leave Disability compensation Missing Localization of pain [n (%)] Low back pain Neck pain Other Duration of pain [median (IQR)] (years)

Table 1 Baseline characteristics (n=128)

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Course of outcome Pain Disability Index scores over time The mean PDI scores and the mean average pain scores on the previous week over time are shown in Fig. 2. The wide 95% CI at the even treatment weeks and at the tail of the curve (>17 weeks) can be attributed by the smaller number of patients measured in those weeks (n= shown under the x-axis). From week 17, the number of participating patients decreases to less than 1/3 of the total number of patients at week 1; therefore the 95% CI’s become wider and the scores are less precise.

Table 2 Mean PDI and average pain at T1. Variable

Total group n=125 Mean (SD)

Completers n= 99 Mean (SD)

Non-completers n=26 Mean (SD)

Differences in means (95% CI)

p value

PDI (0-70) Average pain (0-10)

34.2 (12.0) 5.6 (1.9)

32.8 (11.8) 5.5 (1.9)

39.4 (11.4) 5.7 (1.8)

6.6(1.4 to 11.7) 0.2 (-0.6 to 1.0)

0.013 0.627

8 60 Mean PDI and 95% CI

7 50

4 3

2 10 1 0 0

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 25 27

107 14

7UHDWPHQW ZHHN

-10 Q

119* 6

Mean of average pain in previous week

PDI

Pain

0 -1

Figure 2 Mean PDI and 95% CI and mean of average pain. n= Number of available measurements reported per week.

Mixed-effect modeling The covariance structure was autoregressive heterogeneous, which means that correlations between scores reduced over time

. Variables that were entered in the model but did not

improve the model signiﬁcantly were patient characteristics (age, sex, marital status, and employment status), duration of the PRP (≤8, 9 to ≤12, 13 to ≤16, and ≥17 weeks), pain duration, pain localization and interaction between duration of PRP and treatment week. This means that these variables did not inﬂuence the course of reduction on PDI over time. Predictor variables that

Course of disability reduction during a PRP | 51

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

significantly influenced the model were treatment week, initial PDI score, average pain, and the interaction between initial PDI and treatment week. This means that these variables did influence the course. The model improved significantly by allowing for each patient a random intercept and a random slope for average pain (Table 3). Treatment week2 contributed significantly toward the regression equation, treatment week3 did not. This means that the PDI reduces according to a quadratic model over time during PRP, and not according to a cubic model. Table 3 Results of a linear mixed-model analysis to predict change in Pain Disability Index score during a pain rehabilitation program. Predictor Intercept (Pb) Treatment week Initial PDI Average pain previous week (Pb) Time X Initial PDI Treatment week x Treatment week

βa -5.3 -0.2 0.8 2.3 -0.02 0.03

SE β 1.9 0.2 0.05 0.2 0.004 0.007

P value 0.005 0.174 < 0.001 < 0.001 < 0.001 < 0.001

PDI, Pain Disability Index. β: Regression coefﬁcient. b P: random for patient. a

Clinically, this means that treatment week, initial PDI score, average pain, and the interaction between initial PDI and treatment week influence the course of PDI scores during PRP. The coefficients of Table 3 indicate that, for instance, a patient with an initial PDI score of 40 and a score of 6 on average pain in the fourth treatment week will have an average PDI score of 36.3 in the fifth treatment week of the PRP. This can be calculated using the following formula: PDI= -5.3 + (-0.2 x treatment week) + (0.8 x initial PDI) + (2.3 x average pain) + (0.03 x treatment week2) + [-0.02 x (initial PDI x treatment week)].

Discussion At group level, disability reduction during PRP occurs following a quadratic model, quite similar to the hypothesized line B in Fig. 1. Reduction in average pain in the previous week precedes decrease of disability. The strength of this influence is 2.3 points on PDI per point pain reduction (Table 3). Hence, a reduction of 1 point on average pain will decrease disability with 2.3 points. The interaction between initial PDI and treatment weeks also influences the course, which means the strength of influence per treatment week on PDI (β=-0.02) depends on the initial PDI: a higher initial PDI will lead to a steeper decrease per week compared with a lower initial PDI. The influence of the quadratic factor of treatment week on the course is 0.03: the effect of treatment week on PDI (-0.2 per treatment week) is reduced by the quadratic term of treatment week

52 | Chapter 3

(0.03). This means that towards the end of the PRP, disability reduction per week becomes smaller (Fig. 1, line B). The results of this study have provided insight into the course of disability reduction over time. The clinical relevance of this study is two-fold. First, because most of the gain in disability reduction is obtained in the earlier weeks of PRP, the added value of additional PRP toward the end of the program may become trivial. Second, the relevance of knowing which factors influence the course of disability reduction may be that one can on average predict the duration of the PRP if one knows the target reduction of disability. For example, if the target reduction is 8.5 PDI points, the minimal clinically important change 12, the predicted mean duration of the PRP for an ‘average patient’, with an initial PDI of 40 and an average pain score of 6, would be 14 weeks. This is the first study in which the course of disability during a PRP was analyzed. Most studies have focused on the effect of PRPs based on pre, post, and follow-up measurements, focused on other types of treatments or other groups of participants 6,15,16. With this study, we have provided a new aspect of research of PRPs: dose of PRP. Within the rehabilitation literature, there is no robust evidence to substantiate dose aspects (duration and contact hours) of PRPs. It appears that dose aspects have been neglected in PRP research. Generalization of the findings of this study may be challenging and calls for replication studies within and outside the Netherlands. The general principle of the course of disability reduction might also apply to PRPs outside the Netherlands. However, dose choices are also influenced by cultural factors and differences between professional teams within and outside the Netherlands. In addition, financial reimbursement will also influence the choice of dose. Study limitations This study has some limitations. Some data were missing on PDI items. However, the effects of these missing data were limited because the total score of PDI is used, which represents general (average) disability. For the statistical analyses, the use of a mixed model limited the influence of missing data. In clinical practice, we could not arrange a rigidly fixed measurement time equal for each patient. We had to deal with absence of patients and care as usual was not ready for digitalized versions of the questionnaires to minimize missing data because of absence. Nevertheless, all available data were used in the mixed-effect model. Because of ethical regulations related to digital patient information safety, we could not follow up patients every 2 weeks after PRP. It is advised, however, that future studies focus on the follow up period also because it is assumed that further progress occurs after completion of the program. Twenty percent of the patients dropped out, but data of all patients, completers and noncompleters, were taken into account for the mixed-model analysis. Noncompleters may be considered a limitation from a statistical point of view. However, in daily clinical care, noncompletion is often

Course of disability reduction during a PRP | 53

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

observed. Noncompletion has not been the subject of study. In this study, it appeared challenging for the clinicians to estimate treatment duration before the PRP. As shown in the result section, the actual PRP duration differed from the pretreatment estimated duration. This may be regarded a limitation of this study because this provides variance in total treatment weeks and number of patients per treatment duration, but it underscores the clinical need to know the course of disability reduction and factors predicting this course that could contribute toward more rational estimations of treatment duration. The major strength of this study is the strong relation to clinical practice, which enhances the external validity: measurements were performed in care as usual and the results of the mixedmodel analyses can contribute toward estimate of the duration of PRPs. Also, the heterogeneous group strengthens the results of this study because differences in pain locations did not affect the results and therefore the outcome is generalizable to a broader group of patients with CMP. Conclusions regarding optimum duration of PRPs cannot be drawn. However, the results of this study show the need for further research on the course of disability and dose aspects of PRPs. Future research could focus on the effect of shortening a PRP on the course of disability reduction or the course of disability after completion of a PRP. It should also focus on the added value of additional PRP (h/weeks) to analyze at which point treatments are no longer beneďŹ cial for the patient or no longer cost-effective.

Conclusion From the results of this study, we can conclude that disability reduces according to a quadratic model during PRPs. Initial PDI, treatment week, average pain, and interaction between initial PDI and treatment week inďŹ&#x201A;uence the course of disability reduction during PRP. These data could aid in clinical practice to predict the duration of the PRP at start. Further research investigating dose aspects of PRPs is needed to aid in the effectiveness and cost-effectiveness of PRPs.

Acknowledgements Many thanks are due to all the therapists of the Pain Rehabilitation Department of the Center for Rehabilitation of the University Medical Center Groningen for collecting the data.

54 | Chapter 3

References 1.

McCracken LM, Turk DC. Behavioral and cognitive-behavioral treatment for chronic pain: Outcome, predictors of outcome, and treatment process. Spine (Phila Pa 1976). 2002;27(22):2564-2573.

Gatchel RJ, Okifuji A. Evidence-based scientiﬁc data documenting the treatment and cost-effectiveness of comprehensive pain programs for chronic nonmalignant pain. J Pain. 2006;7(11):779-793.

Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: A systematic review of interventions and outcomes. Rheumatology (Oxford). 2008;47(5):670-678.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. european guidelines for the management of chronic nonspeciﬁc low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

10.

Tait RC, Chibnall JT, Krause S. The pain disability index: Psychometric properties. Pain. 1990;40(2):171182.

11.

Gronblad M, Hupli M, Wennerstrand P, et al. Intercorrelation and test-retest reliability of the pain disability index (PDI) and the oswestry disability questionnaire (ODQ) and their correlation with pain intensity in low back pain patients. Clin J Pain. 1993;9(3):189-195.

12.

Soer R, Reneman MF, Vroomen PC, Stegeman P, Coppes MH. Responsiveness and minimal clinically important change of the pain disability index in patients with chronic back pain. Spine (Phila Pa 1976). 2012;37(8):711-715.

13.

Field A. Multilevel linear models. In: Discovering statistics using SPSS. 3rd ed. Los Angeles, London, New Delhi, Singapore, Washington DC: Sage; 2009:725.

14.

Twisk JWR. Applied longitudinal data analysis for epidemiology: A practical guide. Cambridge etc.: Cambridge University Press; 2003:301.

15.

Gross DP, Haws C, Niemelainen R. What is the rate of functional improvement during occupational rehabilitation in workers’ compensation claimants? J Occup Rehabil. 2012;22(3):292-300.

16.

Rainville J, Jouve CA, Hartigan C, Martinez E, Hipona M. Comparison of short- and long-term outcomes for aggressive spine rehabilitation delivered two versus three times per week. Spine J. 2002;2(6):402407.

Course of disability reduction during a PRP | 55

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Fran Fr anka an ka P.C C. Wa Wate ters te rsch rs choo ch oot, oo t,, Pie ete terr U. U Diji ks k tr traa Ph PhD, D D, Joha Jo hann ha n ess G.M. .M M. Bu Burg rg ger erho ho of Ph PhD, D Jan H.B D, B. Ge Geer e tzen er tzzen MD, PhD hD, Henr He n ic nr icaa R. R Sch chip ip pho horsst Pr Preu e pe eu perr MD MD,, Ph P D, Micchi hiel el F. Re R ne nema m n Ph ma PhD. D Subm Su bmit bm ittte ted d fo or pu publ blic bl icat ic atio at io on

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Abstract Objective: To analyze effects of multidisciplinary pain rehabilitation programs with different dosages; care as usual (CAU) versus care as usual short form (CAU-SF). Methods: A single blinded, 2 armed, randomized controlled trial, with non-inferiority design was performed. All patients with chronic musculoskeletal pain referred to an outpatient multidisciplinary pain rehabilitation program were eligible for this study. Only dosage in weeks differed between the 2 groups, content was similar. The pain disability index (PDI) was the primary outcome measure. Four points difference on PDI was applied as non-inferiority margin. Treatment effects within groups were expressed in standardized mean difference (SMD) and effect sizes (ES) were calculated between groups. Results: Both groups improved signiďŹ cantly on PDI (CAU: -10.8, CAU-SF: -8.3). Frequent extension of CAU-SF leads to limited contrast in dosage between groups. The 2.5 points difference on PDI falls within the non-inferiority margin but the conďŹ dence interval (CI) (-2.2 to 7.3) exceeded it. SMDs of CAU and CAU-SF were 0.8 and 0.7 respectively. ES was 0.2. Conclusions: Reduction of dosage of pain rehabilitation programs did not lead to non-inferior mean results. Difference in means showed non-inferiority but CI exceeded both the upper and lower border of the non-inferiority margin. The results of this trial are inconclusive. Keywords Dosage, Pain Rehabilitation Program, Chronic pain, Chronic musculoskeletal pain.

58 | Chapter 4

Introduction Multidisciplinary Pain Rehabilitation Programs (PRPs) are recommended for patients with chronic low back pain (CLBP) 1. They aim to reduce disability, distress and use of health care services by education of physical, psychological and practical techniques to improve function, work participation and health related quality of life 2. Multidisciplinary programs are effective for patients with chronic musculoskeletal pain (CMP) 1,3,4. Across pain rehabilitation centres nationally and internationally, there is much diversity in content and dosage of PRPs. Health care systems aim for the best treatments based on available evidence. However, there is paucity of evidence about the influence of dose on effects of PRP. In a recent systematic review 5, no randomized controlled trials (RCTs) were identified that were designed to analyse effects of differences in dose variables on outcome of PRPs. Additionally no studies were identified with the primary objective to analyze the association between dose and effect of PRP. If PRPs optimal dosage is known, this may benefit patients and could reduce direct or indirect costs. If similar effects are achieved with a shorter program, this could lead to earlier reduction of disability, regaining quality of life, and sooner participation in daily life, and minimize direct and indirect costs associated with PRP. Employers could also benefit from earlier return to work (RTW) for patients with work productivity loss, rehabilitation centres can reduce waiting lists and overall it can improve efficiency of care. Additionally shorter programs are also attractive for health care insurers and society as a result of reducing direct and indirect costs. The aim of this study was to analyze differences in effects of PRPs with different duration: care as usual (CAU) and PRP short form care as usual (CAU-SF) in a non-inferiority design. CAU-SF was hypothesized to be non-inferior to CAU in an outpatient pain rehabilitation setting.

Methods Design The study is a single blinded, 2 armed, RCT, with a non-inferiority design. The RCT took place at the University Medical Centre Groningen (UMCG), Department of Rehabilitation Medicine, Centre for Rehabilitation, location Beatrixoord Haren. The Medical Ethics Committee of the UMCG approved the study protocol [NL30094.042.11]. Trial registration: international clinical trials registry platform of WHO, TrialID= NTR3385.

Dosage of PRP: A RCT | 59

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Definitions Because dosage variables are relatively new to the scientific field of PRP, we have defined dosage related terms for this study. Duration means the total length of PRP and is expressed in weeks. Contact hours are the total amount of hours that a patient spends with his therapist during PRP. Participants All patients with CMP referred to an outpatient PRP at the Centre for Rehabilitation, UMCG, between September 2011 and October 2013 were potential participants for this study. This centre provides PRPs with duration of 8, 12, 16 or 20 weeks. Patients were eligible for the study when: 1. They were admitted for a 12, 16 or 20 week PRP; 2. they had CMP for more than three months without a specific pathological cause; 3. they experienced CMP induced disability; 4. social and psychological factors were complex and were assumed to be relevant in maintaining CMP induced disability 2; 5. they were willing to stop other treatments for CMP during PRP (except pain medication); 6. they were 18 years or older; 7. they were motivated to participate in PRP and 8. they were willing to participate in the study and signed informed consent. Patients were excluded when: 1. they were referred to the 8 weeks PRP; 2. they were unable to understand the Dutch language; 3. they had co morbidities such as heart failure, rheumatoid arthritis, or psychiatric disorders preventing a PRP. Interventions Common features of CAU and CAU-SF The objectives of CAU and CAU-SF were the same and the content was similar; both treatments were outpatient, multidisciplinary PRPs aimed to decrease CMP related disability, optimize participation and increase quality of life. PRP is intended to coach patients to self manage their pain and disability. The rehabilitation team consisted of rehabilitation physicians, occupational therapists (OTs), physiotherapists (PTs) and psychologists (PSYs). The PRP, based on cognitive behavioural principles, consisted of pain education, for example regarding differences between nociceptive and chronic pain. Patients were counselled to reflect on their pain management strategies (avoiding pain) and how these strategies could be changed into other management strategies (pain coping, alternating physical activity and rest and gradually increasing activities). Each patient set individual treatment goals. PTs applied exercise programs and sports activities, to improve patients’ confidence in movement and reduce pain-related fear, improve activity level, and physical functioning. OTs assessed current activities and patterns in daily living and educated patients how activities could be changed into, healthy activity levels and patterns. PSYs coached patients in understanding and dealing with the social and emotional impact of pain in daily life, pain beliefs, barriers for behaviour change and coached patients how to cope with pain. The rehabilitation physicians were responsible for the medical diagnosis and interventions (if any) and coordinated the overall treatment plan.

60 | Chapter 4

After the rehabilitation physician considered patients to be eligible for PRP, they were invited for an intake procedure, consisting of an interview by an OT, a PT and a PSY. These interviews were aimed at admitting a patient to PRP or not, and to formulate a treatment plan. If the patient was admitted to PRP, based on the assessments during intake, the team (rehabilitation physician, OT, PT, PSY) determined the required duration of 8, 12, 16 or 20 weeks based on assessment of the complexity of the physical, social and personal situation of the patient, motivation and ability to change behaviour. The team who determined treatment duration consisted of professionals with on average 9 years of experience in pain rehabilitation. PRP consisted of OT sessions, 30 minutes, 2 times a week, PT sessions, 30 to 60 minutes, 2 times a week and PSY sessions, 60 minutes, once a week. The session frequency of OT and PT reduced to once a week, the PSY sessions reduced to once every 2 weeks, based on the frequency protocols per duration. All sessions were delivered in a individual setting. For clinical reasons, the duration in weeks of PRP could be adapted (extended or shortened) in CAU and in CAU-SF, depending on the progression of the patient. Duration could be extended when additional coaching was needed to decrease in disability during the extra weeks of PRP. Duration was shortened when treatment goals were achieved earlier than expected, when the patient demonstrated continued lack of progress, or when the patient stopped the treatment, because it did not match patient expectations or reasons not related to the program (for instance holidays). Extending or shortening of the duration of PRP was based on agreement between the patient and the PRP team and decided by the rehabilitation physician. Reasons for extending or shortening PRP were registered. The difference between CAU and CAU-SF was the duration (in weeks) of PRP proposed to the patients after the randomization procedure. Differences in duration will also result in differences in contact hours. Patients allocated to the CAU group received 12, 16 or 20 weeks of PRP as proposed after the intake procedure. Patients allocated to the CAU-SF group received a 4 week shortened form of PRP than proposed (which will result in approximately 10 fewer contact hours on average). Thus patients in CAU-SF received 4 weeks less than proposed after intake. Patients allocated to 8 weeks of PRP in CAU were excluded because the treatment team thought it would not be feasible to offer PRP in 4 weeks with frequencies twice a week. Outcomes The primary outcome variable was self-reported disability assessed with the Pain Disability Index (PDI) 6. The questionnaire consists of seven items related to work, leisure and activities of daily life each item is scored on an 11-point scale (0 indicating no disability and 10 indicating maximum disability). The total scale ranges from 0 to 70. The high correlation between the PDI and Oswestry Disability Index (r=0.83) indicates good validity. Test-retest reliability is high (ICC=0.91) 7. Minimal Clinically Important Change (MCIC) ranges from 8.5 to 9.5 points 8. Dosage of PRP: A RCT | 61

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

The secondary outcome variable, quality of life, was measured with the Euroqol 5D-5L (EQ5D-5L; scale range -0.329 to 1.000) 9. The EQ-5D-5L assesses 5 dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/depression) of health on a 5 point scale: no problems, slight problems, moderate problems, severe problems, and extreme problems. The overall health is assessed on a 0-100 visual analogue scale of 20 cm (EQ-5D-VAS). A higher score indicates a better overall health state. The minimally clinical important difference (MCIC) was 10.5 points for the EQ-5D VAS and 0.03 points for the 5 dimensions of the EQ-5D-3L 10. The EQ-5D-5L version is based on the EQ-5D-3L version. The 5L version demonstrates improved discriminatory power 11, but the exact MCIC for this version has not been examined (yet). Sample size Sample size was calculated using the group sequential non-inferiority criterion. The standard deviation (SD) on PDI scores in CAU was 11. The non-inferiority margin was calculated as 40% of the mean change on PDI scores in CAU, which was 10 points, thus the non-inferiority margin was set at 4 points. For group sequential non-inferiority design an alpha of 0.045 and power of 0.8 was used. Using PASS software, these data gave an estimation of 124 patients needed per arm and 248 patients in total. Including 10% drop-outs, 276 patients should be included for the study. Randomization After determination of PRP duration by the team and after obtainment of informed consent, patients were randomized by an independent person. Patients were stratiďŹ ed on work status (paid work or unpaid/no work), resulting in 2 strata. For each stratum blocks of 6 were used. Sequentially numbered, sealed, opaque envelopes were used for each stratum. Blinding Patients were blinded for intervention. Before inclusion, they were informed about duration of CAU, which could be 8, 12, 16 or 20 weeks. Patients were explained that a duration was proposed by the team based on the assessment results. Furthermore, they were informed about the experiment of 4 weeks shortening duration. They were unaware of the treatment duration they supposed to receive in CAU, and, therefore, they were blinded of allocation to CAU or CAUSF. Patients were outcome assessor for outcome measurements; they were blinded for T0 scores. Rehabilitation physicians, OTs, PTs and PSYs could not be blinded for intervention for the 20 weeks duration because this part only exists in the CAU group. Eight week PRPs could be CAU-SF or patients not included in the study attending 8 weeks PRP.

62 | Chapter 4

Analyses An interim analysis was performed halfway the inclusion period. This enabled us to recalculate the a priori sample size based on trial data, as opposed to assumptions (a priori). Additionally, based on a-priori stopping rules, it was decided that the trial would be stopped if more than 25% of the PRPs in the CAU-SF were extended or if the difference in means on PDI between CAU and CAU-SF was more than 4 points. Differences in mean PDI scores were tested by one sided t-test. In consultation with a statistician, a priori it was decided to split the alpha of 0.05 in 0.005 and calculate a two-sided 99% conﬁdence interval (CI) for the interim analysis, and 0.045 with a twosided 91% CI for ﬁnal analyses. Because the study was stopped preliminarily, however, only one analysis was performed for which we used a two-sided 95% CI approach.

Means and SD’s of primary outcome are presented for CAU and CAU-SF group. All statistical analyses were performed according to an intention to treat analysis and a per protocol analysis. A CI approach was used to interpret non-inferiority (Figure 1). For PDI and EQ5D scores, 95% 2-sided CIs were used in the ﬁnal analyses. Non-inferiority was established when the difference in means was less than 4 (Figure 1). Treatment effects within groups were expressed in standardized mean difference (SMD): (meanchange/ SDchange). Effect sizes (ES) were calculated between groups (meanchange intervention group - meanchange control group)/ SDPooled

. Interpretation of SMD and ES: <0.10 no

effect; <0.30 little effect; <0.50 moderate effect; <0.80 large effect; >0.80 very large effect 13. Superior

Equivalent

Inferior

Non Inferior

A B1 B2 B3 C2 D

2.53

-2.2

7.3 C1

Figure 1 Illustration of the phenomenon of non-inferiority testing The dotted horizontal lines represent the theoretical width of the Conﬁdence Intervals (CI) of the differences in means between experimental group and control group. If the CI lies left of 0: (A) the experimental group is superior to the control group. If CI lies left of 4 (B1, B2, B3) the experimental group is non inferior. If CI includes 4 (C2) non inferiority is inconclusive. If CI lies is right to 4 the experimental group is inferior (D). The statistical analyses of the differences between groups resulted in line C1 with conﬁdence intervals for -2.2 to 7.3: which means non inferiority is inconclusive

Dosage of PRP: A RCT | 63

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Results Patient characteristics and trial flow Patients were recruited out of an estimated sample of 694 patients, admitted for PRP between September 2011 and October 2013. In total 257 (37%) were eligible after signing informed consent. Reasons for non-participation were for example: planning misunderstandings, trial participation was of low priority and thus forgotten to sign informed consent, unwilling to participate in a trial. Interim analyses (October 2013) revealed that in 36% of the patients in the CAU-SF PRP was extended. Based on the a priori stopping rules, the study was stopped directly after the interim analyses. At that point, n=201 (102 CAU, 99 CAU-SF) patients were randomized (Figure 2). In accordance with the rules of the METC, 30 patients, who were enrolled in the trial, were given the possibility to receive the admitted treatment duration and therefore these patients were excluded from analyses. Analyses were performed for 153 patients who had completed PRP: 81 in CAU and 72 in CAU-SF. The majority of patients experienced pain for more than 1 years. The most prevalent diagnoses were chronic back pain (42%), chronic neck pain (19%), widespread pain (9%) and fibromyalgia (7%). Patients who completed the study differed significantly on work status and welfare status, compared to those who did not complete the study (Table 1). Enrollment

Assessed for eligibility (n=257)

Excluded (n=56) • Not meeting inclusion criteria (n=43) • Excluded because stop RCT (n=13)

Randomized (n=201)

Allocation Allocated to CAU-SF (n=99) • Received allocated intervention (n=72) • Did not receive allocated intervention (n=12) • Excluded because stop RCT (n=14) • Declined to participate (n= 1)

Allocated to CAU (n=102) • Received allocated intervention (n= 81) • Did not receive allocated intervention (n= 5) • Excluded because stop RCT (n=16)

Analysis

Analysed (n= 81)

Figure 2 Flow diagram CAU: Care as usual, CAU-SF: Care as usual short form

64 | Chapter 4

Analysed (n=72)

Table 1 General characteristics of participants in the CAU and the CAU SF and participants who completed the study and those who did not. CAU n=81 Age (years) (n=153) Average pain last week (n=153) Worst pain last week* (n=153) Gender (% female) (n=153) Duration of pain (n=131) 3 months to 1 year > 1 year Assigned treatment duration (weeks) (n=153) 8 12 16 20 Marital status (n= 132) Single Married/ living together/LAT Otherwise Education (n=136) No or low education Middle education High education Work status** (n=135) Employed, fulltime Employed, parttime houseman/-wife Unemployed Otherwise Welfare status** (n=135) Not working and no welfare Partial sick leave/disability pension Full sick leave/disability pension unemployed Otherwise Pain location (n=333) Head or face Upper extremities Neck or shoulders Low back Lower extremities Otherwise Number of sites (median, IQR)

mean 44 6 8 n 45

CAU-SF n=72

Completers n=120

Drop-outs n=33

SD mean SD median IQR median IQR 12.4 44 11.9 44.6 33.6 to 52.3 47.4 40.4 to 54.9 1.6 7 1.6 7 6 to 8 7 5 to 8 1.6 8 1.3 8 7 to 9 8 7 to 9 % n % n % n % 56 49 68 79 66 15 45

15 23 50 77

20 30 46 70

27 26 78 74

8 31 18 69

0 37 41 3

26 45 1 0

23 66 30 1

3 16 12 2

0 46 51 4

36 63 1 0

19 55 25 1

9 49 36 6

9 14 53 80 4 6

11 17 47 71 8 12

16 15 77 73 12 12

4 15 23 85 0 0

20 30 30 44 18 17

13 19 34 50 21 31

25 23 51 47 33 30

8 30 13 48 6 22

19 22 9 8 11

27 32 13 12 16

20 18 8 8 12

31 27 12 12 18

34 36 9 12 16

32 34 8 11 15

5 4 8 4 7

18 14 29 14 25

25 30 6 4 3

37 44 9 6 4

23 30 9 0 5

34 45 13 0 7

43 40 14 4 6

40 37 13 4 6

5 20 1 0 2

18 71 4 0 7

16 21 38 47 32 15

9 12 22 28 19 9

21 16 39 45 33 10

13 10 24 27 20 6

31 27 59 75 50 20

12 10 23 29 19 8

6 10 18 17 15 5

8 14 25 24 21 7

2 [1,4]

3 [1,3]

CAU: Care as usual, CAU-SF: Care as usual short form * SigniďŹ cant difference between CAU and CAU-SF (<0.05) **SigniďŹ cant difference between completers and drop-outs (<0.05)

Dosage of PRP: A RCT | 65

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Outcomes and estimation The reduction of PDI scores was 2.5 larger in the CAU group compared to the CAU-SF group (Figure 1). The point estimate of the difference was 2.5 and lies within the non-inferiority margin from 0.0 to 4.0. Because the CI of this difference exceeded both the upper and the lower border of the non-inferiority margin, the results were inconclusive. In 74% of the cases, PRPs were delivered per protocol or shorter, and in 26% of the cases PRPs were extended (Table 2: CAU-SF 36%; CAU 17%). Differences in dosage between shortened, dosage as planned and extended PRPs were non-significant (p=0.066). Reasons for extending were: changes in the behaviour or situation of the patient, RTW assistance needed more support and/or more time than anticipated, patient case was more complex than estimated a priory, and treatment logistics. Both groups improved significantly over time. SMD and ES for PDI and EQ5D scores are presented in Table 3. Differences in results were not significant (p>0.05) between groups for both outcome measures. Table 2 Treatment duration CAU n=81

CAU-SF n=72

P value

95% CI

Received treatment duration (weeks) (mean sd)

11.7

4.5

10.8

3.9

0.323

-0.525 to 2.219

Received contact hours (mean, sd) PRP Dropouts (n,%) Weeks (mean, sd) PRP Shortened (n,%) Weeks (mean, sd) PRP As planned (n,%) PRP Extended Weeks (mean, sd)

30.7 22 6.0 23 -5.7 22 14 2.9

11.3 27% 3.7 29% 4.0 27% 17% 2.0

29.8 11 5.1 15 -3.4 20 26 4.0

10.4 15% 2.2 21% 2.6 28% 36% 2.0

0.622

-2.651 to 4.389

CAU: Care as usual, CAU-SF: Care as usual short form, PRP: Pain Rehabilitation Program.

66 | Chapter 4

Table 3 Trial results (pain disability and quality of life) between and within treatment arms

PDI total CAU CAU-SF ES for between group difference EQ5D index total CAU CAU-SF ES for between group difference

T0 Mean (sd)

T1 Mean (sd)

Difference in 95% CI means (sd)

SMD

144 74 70

37.0 (13.3) 36.1 (12.5) 37.9 (14.2)

119 63 56

25.8 (16.2) 25.1 (15.0) 26.6 (17.7)

112 58 54

9.6 (12.8) 10.8 (13.2) 8.3 (12.2) 0.20

7.2 to 12.0 7.3 to 14.3 5.0 to 11.6

0.8 0.8 0.7

125 66 59

0.55 (0.20) 0.56 (0.19) 0.54 (0.21)

101 54 47

0.70 (0.18) 0.70 (0.17) 0.70 (0.20)

93 51 42

0.13 (0.17) 0.11 (0.19) 0.15 (0.15) -0.19

0.09 to 0.17 0.06 to 0.17 0.10 to 0.19

0.8 0.6 1.0

T0: before pain rehabilitation program (PRP); T1: direct after PRP. PDI: Pain Disability Index (scale 0-70); EQ5D: Euroqol 5D index (scale 0.0-1.0),CAU: Care as usual, CAU-SF: Care as usual short form, SMD: Standardized mean difference; ES: Effect Size

Ancillary analyses Further analyses were performed on differences in trial results between subgroups, patients for whom the PRP was delivered per protocol or shortened, and patients for whom the PRP was extended. The mean duration of the per protocol subgroup was 10.0 (SD 3.8) weeks. The mean number of contact hours was 27.1 (SD 9.7). For the extended subgroup mean duration was 15.0 (SD 3.3) weeks, and 39.2 (SD 8.9) contact hours (Table 4). SMD of PDI of the per protocol subgroup was 0.9; ES of the extended subgroup was 0.6. Because the duration of the PRPs was similar between groups, a regression analysis was performed to analyze predictors of PRP results. Dependent variable was change in disability (PDI T1-T0). Independent variables were: PDI baseline, PRP duration (weeks) and contact hours, gender, average pain at baseline and interaction term of gender and duration (weeks). Dosage of PRP or any of the other variables did not signiďŹ cantly contribute to the regression model (results not presented).

Table 4 Differences between subgroups of patients, who completed the trial, according to duration of PRP, per protocol or shorter and extended

PDI T0 PDI T1 PDI T0-T1

PRP per protocol or shorter mean sd n 37.7 12.3 77 26.7 16.6 71 10.4 10.4 68

PRP extended Mean Difference 95% CI P value mean sd n 33.0 14.0 37 4.7 0.42 to 9.79 0.07 22.8 16.2 38 3.9 -2.67 to 10.49 0.24 9.4 16.0 35 1.1 -4.12 to 6.26 0.68

PDI: Pain Disability Index (scale 0-70), PRP: Pain Rehabilitation Program, T0: before treatment, T1: after treatment.

Dosage of PRP: A RCT | 67

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Adverse events No trial related adverse events were reported.

Discussion The statistical analyses showed a point estimate difference between groups of 2.5 with CI from -2.2 to 7.3, which means non-inferiority was inconclusive. Within both groups, short term ES were moderate to large for outcome measures disability and quality of life. In a quarter of all cases the duration of PRP could be reduced. Because of extension of PRP in patients randomized to CAU-SF, this study was not able to detect differences in dosage or effect. The results could imply that shortening of PRP duration may be considered in some cases, without loss of short term results. A lower dosage may benefit patients and other stakeholders. However, 26% of all PRPs (36% within CAU-SF and 17% within CAU) were extended to achieve a desired outcome. Extending (and shortening) of PRP was always agreed upon by the patients and the PRP team, and was a result of a difference between baseline needs and progress during PRP. At this point in time it is not possible to identify characteristics of patients whose PRPs were extended or shortened, and thus to plan a correct dosage at the start. This topic should be subject of further investigation. On the other hand, the results of the extended subgroup were similar to the per protocol group, indicating that extension of treatment might not be a solution to improve treatment outcome. Significant differences between completers and drop-outs may be hypothesized by less motivation of patients with disability compensation or housewives to complete treatment. However, characteristics of dropouts have not been subject of study. It is unknown whether this has systematically influenced the results of this study. Because this trial is the first of its kind, the results of this trial cannot be compared to other pain rehabilitation trials with similar designs. Patient characteristics such as pain disability, pain intensity, pain duration, gender distribution, appear similar to other PRP trials, both in the Netherlands and internationally 5, and regular clinical secondary and tertiary rehabilitation programs in the Netherlands. Dosage of PRP in this study was ‘midrange’ when compared to dosage reported in other studies

. The results of this study stress the relevance of further dosage studies in

different settings 5. One of the main challenges expressed in pain rehabilitation is the issue of ‘what works for whom’. We suggest adding another challenge: ‘how much works for whom?’ Within pain rehabilitation, this may apply to the multidisciplinary program as a whole, but also to its components. This question may open a new line of research that may lead to major new insights. Because dosage has been a methodological blind spot in PRP research, and dosage variables were deﬁned and interpreted differently across studies, results of previous trials may have been biased by lack of control for dosage. If A is compared to B and A leads to superior

68 | Chapter 4

results, it is expected that A is the treatment of choice. However, theoretically, if the dose of A was higher than B, the difference also might be explained by under dosage of B 15. Differences in dosage could also explain differences between studies and could be reasons why some studies were not able to detect signiﬁcant differences between interventions 14. Weaknesses and strengths Weakness of the study was that it had to be discontinued prematurely because too many PRPs in de CAU-SF group were extended. This caused lack of contrast regarding dosage between the CAU and CAU-SF groups, which may explain partly the inconclusiveness of the results. In addition the small sample size caused wide CI, which also led to inconclusiveness. Different scenario’s were considered and discussed with the team to shorten the program. CAU was offered in programs of 8,12,16 and 20 weeks. In absence of evidence, it was decided to test shortening of PRP with 1 step (4 weeks). This could be considered a methodological weakness of the study, because the relative impact differed between programs (from 12 to 8 weeks: 33% reduction, 12 to 16 week: 25% reduction, 20 to 16 weeks: 20% reduction). On the other hand, adherence to CAU strengthens this study because it reduced the risk of allocation bias in patients and clinicians. Extension of treatment duration may be regarded as a limitation or even as adverse event or program violation. However, during this study, we were obliged to adhere to the Dutch health care regulations and the Medical Ethics Committee. As a consequence, we were not allowed to deny health care that was deemed necessary for good patient care. Within the area of pain rehabilitation there are no guidelines regarding dosage of PRPs. As described in the introduction, PRPs aim to reduce disability and use of health care services. During PRP patients are coached to self manage pain and disability. As far as we know there are no published validated measures to assess these self management skills. Consequently, it is unclear at what point a patient will be ready to self manage his pain and disability, and when PRP is no longer of added value. Partly because of lack of evidence regarding dosage of PRP, choices of dosage of PRP are unclear and arbitrary. This may result in differences in dosage between and within PRP teams. This is a weakness of this study because, for some cases this resulted in shortening or extending of PRP in both the CAU and CAU-SF groups. On the other hand, this is an important reason for future research to focus on the rationale underlying dosage of PRP, including transparency of choices in dosage of treatment. This should eventually lead to more rational dosage of PRP, including explicit arguments on which dosage for individual patients is based. This could result in development of validated measurements to determine dosage of PRP. A strength of this study is the pragmatic, clinical design; including the struggle of clinical practice regarding dosage of PRPs. This study shows results of shortening and extending PRPs. Based on the regression analyses, which showed that dosage did not signiﬁcantly contribute to the model, reducing PRP is possible and extending does not automatically lead to better results. As this

Dosage of PRP: A RCT | 69

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

was the ďŹ rst trial within pain rehabilitation to study dosage, we suggest replication in different settings, and that all future trials in this ďŹ eld clearly describe dosage issues to enable future (meta-)analyses of trial results. In conclusion, reduction of dosage of PRP did not lead to non-inferior mean results. Because the CI of the mean difference exceeded both the upper and the lower border of the non-inferiority margin, the results of this trial are inconclusive.

Acknowledgements The authors acknowledge the participation of the patients in this study, and the clinical and support staff of the Pain Rehabilitation Department of the Centre for Rehabilitation of the UMCG.

70 | Chapter 4

References 1.

Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. european guidelines for the management of chronic nonspeciﬁc low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

Köke A, Brouwers M, Heuts P, et al. Consensus rapport pijnrevalidatie Nederland. 2005.

Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: A systematic review of interventions and outcomes. Rheumatology (Oxford). 2008;47(5):670-678.

Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain. Cochrane Database Syst Rev. 2014;9:CD000963.

Waterschoot FP, Dijkstra PU, Hollak N, de Vries HJ, Geertzen JH, Reneman MF. Dose or content? effectiveness of pain rehabilitation programs for patients with chronic low back pain: A systematic review. Pain. 2014;155(1):179-189.

Tait RC, Chibnall JT, Krause S. The pain disability index: Psychometric properties. Pain. 1990;40(2):171182.

Herdman M, Gudex C, Lloyd A, et al. Development and preliminary testing of the new ﬁve-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20(10):1727-1736.

10.

Soer R, Reneman MF, Speijer BL, Coppes MH, Vroomen PC. Clinimetric properties of the EuroQol-5D in patients with chronic low back pain. Spine J. 2012;12(11):1035-1039.

11.

Janssen MF, Pickard AS, Golicki D, et al. Measurement properties of the EQ-5D-5L compared to the EQ5D-3L across eight patient groups: A multi-country study. Qual Life Res. 2013;22(7):1717-1727.

12.

Borenstein M. Introduction to meta-analysis. Chichester, UK. Wiley; 2009:421.

13.

Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988:567.

14.

15.

Dosage of PRP: A RCT | 71

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Frran Fran anka ka P.C C. Wa W te ters rsch rs choo ch o t, Elsel oo lsse in ne Be Benn nnen en Mscc, He Henr nric nr icaa R. R Sch chip ip pho hors rsst Pr Preu eu upe perr MD MD,, Ph PhD, D D, Piiet Piet eter er U. Di Dijk jk kst s ra a PhD hD, Ja Jan n H. H.B. B. Gee eert rtze z n MD ze MD, Ph PhD, D, Micchi hiel e F. Re el Rene ne emaan Ph PhD D

Su ubm bmit itte it ted d fo forr pu publ blic icat atio at ion io n

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Abstract There is a large variety in prescribed dosages of pain rehabilitation programs (PRP) but evidence regarding the optimum dosage is unknown. The aim of this study was to explore perspectives of patients and rehabilitation professionals regarding dosages (e.g. total duration, number of contact hours) of PRP in three different rehabilitation centers in the Netherlands. A study using an explorative qualitative research design was performed with thematic analysis. Individual semistructured interviews were conducted with patients undergoing PRP, and focus groups were formed with rehabilitation professionals involved in PRP. In total, 12 patient interviews and three focus groups with 17 rehabilitation professionals were analyzed. All patients were satisďŹ ed with offered dosage. Several factors were considered important in relation to the choices in dosage of PRP. All factors were categorized as patient-related factors, treatment-related factors and external factors. Although dosage of PRP differed among the participating rehabilitation centers, patients and rehabilitation professionals mentioned similar factors. In absence of evidence, differences in choices of PRP dosage appear mainly based on historical grounds and clinical expertise. Keywords Pain rehabilitation programs, dosage, qualitative analysis, chronic pain.

74 | Chapter 5

Introduction Multidisciplinary Pain Rehabilitation Programs (PRPs) are recommended to treat patients with chronic musculoskeletal pain (CMP)

1,2

. These PRPs, based on the biopsychosocial model, aim to

decrease disability and optimize participation of patients with CMP. A systematic review showed that PRPs have a moderate, but consistent, positive effect on disability and pain, compared to usual care or physical treatment programs for patients with chronic low back pain 3. PRPs result in better self-management of pain and disability, and a reduction in healthcare utilization in treated patients was found, which may contribute to a decrease of direct and indirect costs over the long term 4. Although effective, PRPs are relatively expensive. The multidisciplinary characteristics, as well as the high number of contact hours and total duration of PRP, provide relatively high direct costs and travel expenses for the patients. An aspect of PRP is dosage, which includes the total duration, the total number of contact hours, and intensity of treatment (number of contact hours per week). Differences in the dosage of PRP may lead to differences in direct and indirect costs. Choices for dosage and how dosage is established for an individual patient is unknown. A recent systematic review showed that dosage of PRP has never been studied as a primary aim and the optimum dosage of PRPs is currently unknown 5. The studies included in that review differed in terms of dosage (total duration, from weeks to months, and regarding contact hours, from fewer than 10 to over 100 hours) and effect. Better understanding of dose variables could lead to better and more efďŹ cient patient care, which will beneďŹ t patients, rehabilitation facilities, insurers and employers. To acquire insight into factors related to dosage of multidisciplinary PRP, this study aimed at exploring perspectives of patients who underwent PRP, as well as of the rehabilitation professionals working in PRPs. In addition, this study aimed at examining the argumentations/reasons and underlying choices in dosage of PRPs in three rehabilitation centers in the Netherlands.

Methods A study using an exploratory qualitative design with thematic analyses was used to gain in-depth information about the perspectives of patients and rehabilitation professionals involved in PRPs regarding the doses of these programs. The study was performed in three rehabilitation centers in the Netherlands. These centers were selected because of their differences in dosage, while patient characteristics were similar 6. Individual semi-structured interviews with patients who underwent multidisciplinary PRP, as well as focus groups with professionals working in multidisciplinary PRP,

Patient and professionalsâ&#x20AC;&#x2122; perspectives on dosage of PRP | 75

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

were used. The medical ethics committee of the University Medical Center of Groningen (UMCG) granted a waiver for this study. All participants signed informed consent for participation, digital recording and the use of data. Participants A purposeful sample of patients was recruited from three rehabilitation centers in the Netherlands: Adelante in Hoensbroek (RC1), located in the south; Roessingh in Enschede (RC2), located in the east and the Center of Rehabilitation of the UMCG (RC3), located in the north. We aimed at including patients with varying characteristics, who completed a multidisciplinary PRP: working/ not working, differences in dosage of PRPs, males/females, and a variety of ages. PRP should have been completed for at least two weeks. The first author contacted the recruited patients via telephone to check their willingness to participate in the study and to plan the interview. They received written information about the purpose of the interview. Rehabilitation professionals from the three rehabilitation centers were included for participation in the focus groups, provided that they were working within the area of multidisciplinary PRP for a minimum of two years, for at least 0.5 fte. We aimed to include rehabilitation professionals from different disciplines (i.e. occupational therapist/ physiotherapist/ psychologist/ physiatrist). These rehabilitation professionals should be involved in determining the individual dosages of the program for the patient. Data collection Data were collected between May and September 2014. A total sample of 15 patients was planned for the interviews, five patients from each rehabilitation center, with the expectation of reaching data saturation. A semi-structured interview scheme was constructed [Appendix] for patients using open questions enabling reflection about experiences and perceptions on dosage of treatment received. The interview started with questions about the total duration and intensity of PRP, followed by questions regarding personal experiences and perceptions of treatment dosage, such as: “what did you like and dislike about the duration and intensity of the program” and “which factors, in your opinion, influence dosage of PRP?”. Also patients’ opinion regarding the relationship of dosage to return to work and personal and general costs were asked. Face-to-face interviews were performed at the rehabilitation center where the patient was treated. The interviews were scheduled for 60 minutes and were audio recorded. Prior to this study, an interview was pilot tested. Patient interviews were planned prior to the focus group at each center. For the focus groups meetings, a minimum of five rehabilitation professionals of different disciplines per center were invited in order to achieve variation in perspectives on PRP dosage per

76 | Chapter 5

rehabilitation center. Meetings were planned for 90 to 120 minutes and were audio recorded. An interview scheme was constructed [Appendix], aiming to gain insight into the line of reasoning underlying dosage choices, as well as rehabilitation professionals’ perspectives on optimum dosage of PRPs. Measures of validity The ﬁrst author (FW), who has over 10 years of experience as an occupational therapist in pain rehabilitation, conducted all patient interviews and also led the focus group meetings in the Adelante and Roessingh rehabilitation centers. Because she was closely acquainted with the rehabilitation professionals at UMCG, a psychologist with experience in chronic pain treatment but who was not involved in PRP of the UMCG led this focus group. The second author (EB) assisted in all focus groups.

Data analysis Interviews and focus groups were audio recorded and transcribed verbatim by the ﬁrst and second authors (FW and EB). Transcripts were imported into ATLAS.ti, version 5.2, to analyze data. Thematic analysis was considered as the appropriate approach for identifying, analyzing and reporting factors related to dosage of PRP 7. Because of the lack of evidence regarding dosage of rehabilitation programs, especially PRP aimed at behavior change, it was not possible to provide a detailed theoretical framework. Initially, the ﬁrst and second authors (FW and EB) conducted the inductive coding of the interviews and another independent researcher validated the codes. All codes and quotations were re-analyzed, merged, and renamed.

Results Participants In total, 13 patients were interviewed and 12 interviews were analyzed. Initially, ﬁve patients per center were planned, two dropped out because of logistics problems and one was excluded because the audio was impossible to transcribe. Seventeen rehabilitation professionals participated in the focus groups (RC1: 5; RC2: 5; RC3: 7). Different disciplines participated: occupational therapists (4), physiotherapists (3), movement teacher (1), psychologists (4), social workers (2) and physiatrists (3). The characteristics of participants are described in Table 1. Data saturation was reached for patient interviews and focus groups for rehabilitation professionals.

Patient and professionals’ perspectives on dosage of PRP | 77

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Table 1 Characteristics of participants and PRP dosage

Patients Age mean (SD) Female % Duration of pain n 1 to 2 years 2 to 5 years > 5 years Initial PDI score mean (SD) Missing n

RC1 (n=4)

RC2 (n=4)

RC3 (n=4)

42 (16)

37 (11)

53 (7)

100

0 2 2 27 3

1 0 3 43.3 (14.2) 0

2 0 2 31.5 (16.4) 0

(n=5) 41 (6) 60 7.6 (4.7)

(n=5) 43 (9) 80 9.0 (5.6)

(n=7) 39 (6) 71 7.4 (5.5)

3 12

1 to 2 3 to 36

1 to 2 8 to 20

30 120

2 10 to 150

3 30 to 70

Rehabilitation professionals Age mean (SD) Female % Years of experience in PRP mean (SD) PRP Range in total duration (wks) Assessment Treatment Range in contact hours Assessment Treatment

PDI: Pain Disability index, PRP: Pain Rehabilitation Program; RC1: Adelante in Hoensbroek; RC2: Roessingh in Enschede, RC3: Center for Rehabilitation of the UMCG

Factors related to dosage Different codes related to dosage of PRP were derived from the interviews and focus groups. Codes were categorized as patient-related factors, treatment-related factors or external factors. In addition, the codes were arranged as “shared” (indicating that they were mentioned by patients and rehabilitation professionals), “patients” (those that were mentioned by patients only), and “rehabilitation professionals” (those mentioned by rehabilitation professionals only) (table 2).

78 | Chapter 5

Patient and professionalsâ&#x20AC;&#x2122; perspectives on dosage of PRP | 79

External factors

Team & treatment related factors

Patient related factors

Table 2 Overview of codes per category

Case complexity General status Insecurity Self-knowledge Motivation and ability to change behavior Expectancies Focus on yourself Acceptance of pain Self management Apply lessons learned into practice Waiting time Clarity about dosage and time plan Individually tailoring Shared decision making Contact with rehabilitation professionals Interdisciplinary functioning Content of treatment Treatment goals Format of treatment Direct costs Indirect costs Investment for the future Personal and work factors

Shared factors -

Factors from focus groups with rehabilitation professionals

Support from environment

Travelling time Format of treatment Injury compensation Test results or other treatments

Having time and opportunity to explore Ehealth Saturation Lack of evidence Expertise of the team Prediction of dosage of PRP Timing to ďŹ nish PRP

Assertiveness Relapse

Factors from interviews with patients

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Patient-related factors: Shared factors: Both patients and rehabilitation professionals mentioned case complexity of a patient as an important factor related to dosage of PRP as well as an important issue to take into account for indication of dosage of PRP. Both patients and rehabilitation professionals described characteristics that affect dosage of PRP, such as the general status (the level of physical condition or the degree of experienced disability), as well as the degree of insecurity, self-understanding, motivation and ability to change behavior, in addition to the expectancies patients have regarding treatment and treatment results. (P3): “It depends on what is going on with someone. Is he traumatized? Does he want to talk about it? Is he able to manage his problems? Does he have some self-understanding; does he know what his problem is?” Both patients and rehabilitation professionals mentioned that PRP requires focusing on yourself. Patients and rehabilitation professionals involved with both high and low PRP dosages mentioned this focus as being related to dosage. A patient (P7) who underwent semi-inpatient PRP described: “The inpatient format of treatment was a good experience. It was quite intense, but actually very good because you really have to focus on yourself. If I would have been at home, I would have been distracted by all the things you have to do at home”. Focusing on oneself and on the treatment was also mentioned by a patient (P9) as motivation for preferring twice a week PRP instead of once a week: “Yes, two times a week, than you are really engaged with the treatment. Treatment only once a week…. well, then a whole week is in between…”. Rehabilitation professionals mentioned that being able or motivated to focus on oneself results in being motivated to follow the prescribed dosage of PRP. In some cases, for example (T13) “if a patient is overloaded”, being able to focus on oneself is a motivation for the choice of dosage and the format of PRP, such as inpatient PRP. Acceptance of pain is related to the content of PRP because PRP is aimed at improving function, not reducing pain. Participants mentioned that PRP contributed to the acceptance of pain and that patients need time to go through this process of acceptance (P4): “I think the minimum duration of PRP should be 12 weeks. Especially because you experience a process of change in which you need time… time to make a start, because after those 12 weeks you still need to go on for yourself”. If a patient is able to self-manage his pain and disabilities, participants suggest needing a lower dosage of PRP compared to a patient that is not able to. A rehabilitation professional in a focus group (T16) described: “Sometimes the minimum dosage of PRP is just one hour of intake. There are patients who just need some explanation to self manage their lives again. Being able to self manage pain is related to patient characteristics and behavior, the way people are, and whether

80 | Chapter 5

they are able to cope with their pain and disabilities themselves”. Patients also mentioned that if they are able to learn the principles of self-management during the process of PRP, they will ultimately need a lower dosage. To apply lessons learned into practice, patients need time. This is a motivation of patients and rehabilitation professionals for dose variables such as total duration of PRP. (P9) “It is a process which requires time, 6 or 8 weeks… well, that’s quite short…. You need to be aware of things, you have to practice, adjust and adopt”. Needing time to apply lessons learned was additionally mentioned with regard to intensity (contact hours per week) and the reduction of intensity of contact hours during PRP. (T13): “At the end of the program, reducing the intensity of the program (fewer days or fewer contact hours per week), could support the process of putting the lessons learned into practice”.

Factors from interviews with patients: Assertiveness was mentioned by patients with regard to the ability to stand up for themselves. This code is part of the goals and content of PRP and therefore patients mentioned this in relation to needing a lower dose when patients are already able to stand up for themselves. However, the code has also been reported in terms of “shared decision making” regarding dosage of PRP. (P8) “I had to learn to set boundaries. My day was fully planned and I used to do whatever I was told, which resulted in being exhausted at the end of the day”. Some patients reported that they needed support during times of relapse. (P2): “Everything was going well; I also felt it that way. However, I did not have the feeling that I was ready. I was afraid that I would relapse and that was why I needed some sessions”. Factors from focus groups with rehabilitation professionals: No other patient-related factors were reported by rehabilitation professionals only. Treatment-related factors: Shared factors: Waiting time before and during treatment can influence, in a negative or positive way, the general status of a patient at the start of treatment, therefore having an influence on the dose of PRP. One patient (P1) described how she was able to turn the negative result of the waiting time into a positive action: ”I had to wait for a long time before start of PRP. For me personally, it has led to an increase of pain and I became depressed. Consequently, I realized I had to do something and I started doing things on my own”. Most patients knew the total duration of PRP beforehand. Patients were satisfied with this clarity about dosage and time plan for PRP: (P4) “I compare it with knowing when to go on holidays: the last week at work you are really into holidays. With therapy, it is the same – you have a set goal to work on in time”. Nevertheless,

Patient and professionals’ perspectives on dosage of PRP | 81

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

besides structured dosages providing this clarity, participants emphasized that the possibility of individually tailoring dosage to their needs is important. The individually tailored dosage is also related to shared decision-making. Patients and rehabilitation professionals attached value to shared decision-making in terms of PRP dosage. Good contact between patient and rehabilitation professionals supported therapy and therefore could reduce dosage: (P11): “You need to have a connection with your rehabilitation professionals; otherwise you will not make any progress”. However, too many changes in rehabilitation professionals during therapy can impede the connection, and therefore the process, because: “you have to tell your story again and again”. Related to this code, patients and rehabilitation professionals mentioned that good interdisciplinary functioning of the team supported effectiveness and therefore had a positive influence on PRP dosage. (P9): “I experienced that everyone knew who I was and what I needed. Everyone had read all reports and you did not need to tell the same stories”. Both patients and rehabilitation professionals made it clear that dosage of PRP was strongly related to the content of treatment, treatment goals and the format of treatment (e.g. inpatient versus outpatient). Factors from interviews with patients: Most patients had the experience of having time and opportunity during PRP to explore themselves and reach their goals. However, some patients also experienced this as a problem when the frequency was reduced from five days to two days per week. “Content became more intensive, everything was put together in two days and everything had to go faster…I had to hurry….”. Some patients denoted that towards the end of treatment, saturation of treatment content occurred. Therefore, they suggested that retrospectively, the dosage could have been reduced. (P2): “At a certain moment, you are done with it”. Patients stated that they trust the expertise of the team, both with regard to the content of treatment as a prediction of dosage and when decisions are made regarding the need to prolong or reduce PRP dosage. (P1): “I thought, well if that is the success of the therapy, then I am going to do it that way”. Rehabilitation professionals felt that they had to be experts in the field and should offer a high PRP quality. They related this quality and expertise to content and dosage of PRP. Factors from focus groups with rehabilitation professionals: Rehabilitation professionals suggested that different forms of ehealth could influence the dosage of PRP in the near future. “Patients can receive home assignments by mail after intake and before starting. Also aftercare can be supported by ehealth”. The focus groups revealed that there was a lack of evidence for the delivered doses of PRP. (T15) “These programs have been developed and adapted over the years based on skills acquired by the team, and developments such as group based treatments. This provides certain components in the program. This is what we have; we did not question ourselves if it is necessary for patients”. Additionally, with that lack of evidence, it is hard to predict dosage of PRP and to know when the timing is right to finish

82 | Chapter 5

PRP: “It is a difficult decision, at what point a patient is ready, and to reach consensus about this within the team and also with the patient…. “ External factors: Shared factors: Opinions regarding financial factors were of interest to the interviewer. The majority of participants stated that financial factors should not influence dosage of PRP. Patients could imagine that direct costs that patients have to pay themselves, like travel costs, could be of concern for some patients, although they did not experience that for themselves. (P11): It is regrettable if your own development depends on a few Euros extra per month”. Direct costs of therapy and indirect costs, for example those related to returning to work, should not influence treatment because patients saw the costs of therapy as an investment in the future: (P9): Well, I think, on the longer term, if therapy works, you save money. It also applies for work: if my boss did not allow me time for this therapy, I might not yet be at work for 100%”. Personal and work factors could have a positive or negative influence on the opportunity to receive therapy, and therefore could influence the satisfaction or choice for dosage of PRP. Some patients had to adjust their activities at home to their therapy: (P11): “On therapy days, I did not have the energy to do things at home. I was happy I made the choice to delegate the household tasks. The days at home I had time to do things with my children and husband”. A rehabilitation professional mentioned that sometimes, dosage of therapy can be adjusted to work or personal situations: “If patients are not able to organize the required dosage of two times a week and they have strong reasons, we sometimes decide to reduce frequency of treatment to once a week”. Factors from interviews with patients: Patients perceived support from environment (e.g. partner, colleagues, employer) as a positive factor because it helped with completing the program well and spending the time needed for the program. Lack of support could have a negative influence on the effects of the program because it could interfere with the process of therapy and lead to extension of PRP. Factors from focus groups with rehabilitation professionals: In the focus groups, rehabilitation professionals mentioned that travel time could influence the choice of the offered format of treatment, therefore influencing the dosage of PRP. “This center is a center of expertise regarding pain rehabilitation, so patients come from across the province, which makes outpatient rehabilitation not appropriate”. If the patient received injury compensation, it sometimes can impede progression in treatment. Therefore, rehabilitation professionals stated that in some cases, this could be a reason for not starting PRP or for stopping it prematurely. Test results or other treatments can also influence the dosage of PRP: “More often you see a sort of intermediate treatment if patients need other treatments, like EMDR

Patient and professionals’ perspectives on dosage of PRP | 83

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

….. Or if test results are not available; waiting for these results causes delay” (EMDR is the abbreviation for Eye Movement Desensitization and Reprocessing; trauma treatment). Differences and similarities Interviews and focus groups were performed in three rehabilitation centers across the Netherlands. All centers offered multidisciplinary PRP for patients with CMP. Although dosages differed per center, factors were similar for all centers. All patients, in general, mentioned being satisﬁed with the received dosage, although some had suggestions for improvements related to dosage of PRP. Rehabilitation professionals reported that they supported the choices for current dosages of PRP; however they also had suggestions for improvements and implications for further research in order to eliminate the lack of evidence for an optimal dosage of PRP.

Discussion As a contribution to gather knowledge about factors related to dosage of PRP, this study provides insight into perspectives of patients and rehabilitation professionals regarding dosage of PRP. Several general factors were considered relevant. Overall, all codes could be summarized into patient, treatment and external factors. Interestingly, despite differences in offered and received dosages of PRP, these factors were similar across centers with different PRP dosages. For example, in general, patients need clarity about dosages at the start of PRP. However, during PRP, they would like to have the opportunity to adapt the dosage to their individual needs, in consultation with their rehabilitation professionals. Patients and rehabilitation professionals mentioned that different personal characteristics, such as motivation and ability to change, general status of a patient, acceptance, and self-understanding, are related to dosage of PRP. However, similar “ability to change” can lead to 12 weeks, divided into treatment of six weeks, five times a week, eight hours a day, and six weeks, three times a week, eight hours a day of PRP in RC1, and 12 weeks, twice a week, 2,5 hours a day of PRP in RC3. Remarkably, no explicit reasons for these differences in choices of dosage were expressed during the focus groups. Treatment outcome is caused by specific and aspecific factors of treatment. The content of PRP is a specific factor. Outcome, defined as “being satisfied” with treatment, is an example of an aspecific factor influencing the effect of treatment. The factors in this study regarding dosage of treatment seemed to be linked both to specific and aspecific factors. Ehealth and evidence regarding optimum dosage are specific factors which can contribute to better prediction of dosage and more efficient PRP. Content and format of treatment, and external factors (e.g. direct and indirect costs, investment in future, personal and work factors, support, test results) are specific factors influencing choices of, and satisfaction with dosage. These specific factors could

84 | Chapter 5

explain part of the differences in choices of the offered PRP dosage. However, the results of this study showed that also other, more aspeciﬁc factors seemed to be relevant although not speciﬁc related to dosage. Some of these latter factors are similar to the common factors analyzed for explaining equivalence 8,9

. In a systematic review ﬁve categories of common factors

of various treatment approaches

were analyzed in psychotherapy research: patient characteristics, therapist qualities, change processes, treatment structure and therapeutic relationships 8. Patient characteristics, such as expectations of patients with chronic pain, are related to outcomes of pain treatment

10-12

Rosenzweig 9 already in 1934 described the effect of therapy characteristics with the introduction of the concept of common factors. In physical therapy care, a systematic review analyzing patient satisfaction concluded that characteristics of the therapist and the process of care were determinants of satisfaction of musculoskeletal physical therapy care 13. However, the association between satisfaction and treatment outcome was inconsistent. The category of change processes consists of, for example, the opportunity for catharsis, practice of new behaviors in a safe environment, provision of rationale and foster insight 8. These common factors are similar to the factors identified in this study: having time and opportunity to explore, applying lessons learned into practice, acceptance of pain, and self-management. Hush et al. 13 also described the relationship between patient satisfaction and the process and organization of care. Timely and efficient treatment, adequate treatment frequency and follow up, patient involvement in decision making and individualized care were variables of the process of care, contributing to patient satisfaction. Variables contributing to patient satisfaction regarding the quality and efficiency of the organization of care were: good access to services, location, parking and approachable support staff

. Communication, as part of the treatment structure 8, is concluded to be an

important factor in healthcare

. This is in line with the factors “clarity”, “shared decision-

making” and “contact with rehabilitation professionals” mentioned by patients and professionals in this study. Communication based on patient-centered care strategies, is also associated with a stronger therapeutic relationship 15. In addition, studies have analyzed therapeutic relationships as important common factors that could be related to the effects of different healthcare treatments 16-18

. All common factors were analyzed in different contexts of treatments and in the associations

between different variables of treatment. Remarkably, however, in this study with similar contexts and aims of PRP, similar common factors led to different dosages of PRP. Therefore, it is a challenge to analyze the effect of the different components of PRP on outcome. As a result of the lack of evidence regarding dosage of PRP, as mentioned in the focus groups, all rehabilitation centers made different choices regarding dosage. These choices were based on different motives related to content, dosage, and composition at the time of developing the multidisciplinary PRP. Over time, dosage was adapted because of changes in society, costs,

Patient and professionals’ perspectives on dosage of PRP | 85

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

experiences, and new insights related to content or format of therapy. At present, no evidence can support the choices of dosage of PRP and therefore, rehabilitation professionals are not able to rationally choose optimum dosage of PRP. Strengths and limitations Analyzing factors associated with dosage using a qualitative method from the perspective of patients and rehabilitation professionals is considered a strength of this study because it is the ďŹ rst study providing an overview of factors related to dosage from patient and professionalsâ&#x20AC;&#x2122; perspectives. Taking into account three different rehabilitation centers with different dosages of PRP and similar content across the Netherlands is a strength because this provides insight into the differences and similarities regarding perspectives of dosage, resulting in similar factors regardless of variety in the offered dosage. Although this study was performed in three different centers, it focused solely on centers in the Netherlands, limiting generalization to PRPs performed in centers in other countries. In addition, the qualitative nature of this study resulted in perspectives of patients and rehabilitation professionals and cannot simply be generalized to all other groups of patients and rehabilitation professionals involved in PRPs. For example, we were not able to include more male patient participants. Although, the majority of patients in PRP are female, more male patient participants would have improved comparability to clinical practice. Recommendations This study is an initial exploration of the perspectives of patients and rehabilitation professionals regarding PRP dosage. It resulted in an overview of factors that can be used for future research and clinical practice regarding PRP dosage. The overview can contribute to further analysis of these factors and their relationship to dosage. Also, it can improve the awareness of these factors related to dosage of PRP in clinical practice. The results indicate that factors other than content and dosage of PRP were perceived as being important factors by patients and rehabilitation professionals. Clinical practice can be improved by taking these other factors into account as contributors to the overall success of PRP. Conclusion This study shows that, although dosage of PRP differed, patients and rehabilitation professionals generally mentioned the same factors related to characteristics of patients and treatment, as well as similar external factors. In absence of evidence, differences in choices of PRP dosage appear mainly based on historical grounds and clinical expertise. Therefore, research is needed to guide choices of optimum PRP dosage.

86 | Chapter 5

Acknowledgements The authors would like to thank all patients and rehabilitation professionals who participated in the interviews and focus groups at all three rehabilitation centers. They also extend their appreciation to Ernst Schrier for leading the focus group and Annemieke de Jong and Suzanne Broekema for their support with data analysis.

ConďŹ&#x201A;ict of interest We certify that no party having a direct interest in the results of the research supporting this article has or will confer a beneďŹ t on us or on any organization with which we are associated.

Patient and professionalsâ&#x20AC;&#x2122; perspectives on dosage of PRP | 87

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1.

Turk DC. Clinical effectiveness and cost-effectiveness of treatments for patients with chronic pain. Clin J Pain. 2002;18(6):355-365.

Airaksinen O, Brox JI, Cedraschi C, et al. Chapter 4. European guidelines for the management of chronic nonspeciﬁc low back pain. Eur Spine J. 2006;15 Suppl 2:S192-300.

Kamper SJ, Apeldoorn AT, Chiarotto A, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain: Cochrane systematic review and meta-analysis. BMJ. 2015;350:h444.

Lambeek LC, van Tulder MW, Swinkels IC, Koppes LL, Anema JR, van Mechelen W. The trend in total cost of back pain in the Netherlands in the period 2002 to 2007. Spine (Phila Pa 1976). 2011;36(13):10501058.

Preuper HR, Boonstra AM, Wever D, et al. Differences in the relationship between psychosocial distress and self-reported disability in patients with chronic low back pain in six pain rehabilitation centers in the netherlands. Spine (Phila Pa 1976). 2011;36(12):969-976.

Braun V, Clarke V. Using thematic analysis in psychology. Qualitative Research in Psychology. 2006;3(2):77-101.

Grencavage LM, Norcross JC. Where are the commonalities among the therapeutic common factors? Professional Psychology: Research and Practice. 1990;21(5):372-378.

Rosenzweig S. Some implicit common factors in diverse methods of psychotherapy. Journal of Psychotherapy Integration. 2002;12(1):5-9.

10.

Kalauokalani D, Cherkin DC, Sherman KJ, Koepsell TD, Deyo RA. Lessons from a trial of acupuncture and massage for low back pain: Patient expectations and treatment effects. Spine (Phila Pa 1976). 2001;26(13):1418-1424.

11.

Bhana N, Thompson L, Alchin J, Thompson B. Patient expectations for chronic pain management. J Prim Health Care. 2015;7(2):130-136.

12.

Burns JW, Nielson WR, Jensen MP, Heapy A, Czlapinski R, Kerns RD. Speciﬁc and general therapeutic mechanisms in cognitive behavioral treatment of chronic pain. J Consult Clin Psychol. 2015;83(1):1-11.

13.

Hush JM, Cameron K, Mackey M. Patient satisfaction with musculoskeletal physical therapy care: A systematic review. Phys Ther. 2011;91(1):25-36.

14.

Roberts L, Bucksey SJ. Communicating with patients: What happens in practice? Phys Ther. 2007;87(5):586-594.

15.

Pinto RZ, Ferreira ML, Oliveira VC, et al. Patient-centred communication is associated with positive therapeutic alliance: A systematic review. J Physiother. 2012;58(2):77-87.

16.

Burns JW, Glenn B, Loﬂand K, Bruehl S, Harden RN. Stages of change in readiness to adopt a selfmanagement approach to chronic pain: The moderating role of early-treatment stage progression in predicting outcome. Pain. 2005;115(3):322-331.

17.

Ferreira PH, Ferreira ML, Maher CG, Refshauge KM, Latimer J, Adams RD. The therapeutic alliance between clinicians and patients predicts outcome in chronic low back pain. Phys Ther. 2013;93(4):470478.

88 | Chapter 5

18.

Degnan A, Seymour-Hyde A, Harris A, Berry K. The role of therapist attachment in alliance and outcome: A systematic literature review. Clin Psychol Psychother. 2014.

Patient and professionalsâ&#x20AC;&#x2122; perspectives on dosage of PRP | 89

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

F an Fr anka k P.C ka C. Wa Wate ters te rsch rs choo ch oot, oo t, Elssel e in ne Be Benn nnen nn en Msc, scc, Lu Luca cass H. ca H V. V van der er Wou o de PhD hD, Henr He n icca R. Sch chip ipho ip hors ho rstt Prreu rs eupe p r MD pe MD,, Ph PhD, D, Micchi hiel e F. Rene el Re enema man n Ph PhD. D Accce cept pted pt ed Int J Reh ehab abilil Res ab

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Abstract Assessment of case complexity of patients with chronic non-specific musculoskeletal pain (CMP) is currently clinician based, not transparent and with low reliability. The objective of this study was to explore case complexity and to initiate development of a Case Complexity Index (CCI). A three-round Delphi study among clinicians involved in multidisciplinary Pain Rehabilitation Programs was performed to identify important factors which are assumed to influence functioning in patients with CMP. The 10 most important factors were used to initiate development of a CCI with the mean ratings of importance per factor as weights. The feasibility of the CCI was tested in a pilot study on 16 patients with CMP. In the first round, 166 factors were identified; in the second round, the 10 most important factors were selected; in the third round, relative weights of each factor were calculated, ranging from 1.75 (features of complaints) to 3.56 (psychiatric disorders) on a scale of 0 (no weight) to 4 (very heavy weight). The assessments for the factors were mainly based on clinical examination and reasoning. Clinicians could rate all patients using the CCI, which establish feasibility of the CCI. Ten, mainly psychosocial factors were identified, which were assumed to be most important for assessment of case complexity of a patient with CMP. With these factors a CCI was created for which feasibility was established. This CCI is transparent, easy to use and might provide a basis for further developments of a structured assessment of case complexity, which may have scientific and clinical relevance. Keywords Chronic pain, chronic non-specific musculoskeletal pain, multidisciplinary pain rehabilitation program, case complexity, complexity index, Delphi technique

92 | Chapter 6

Introduction Development of chronic non-speciﬁc musculoskeletal pain (CMP) is often explained with the biopsychosocial model, because interacting factors may be involved inﬂuencing CMP and its consequences

1,2

. The number of factors and their relative involvement creates a highly variable

case complexity of patients with CMP. Due to the multifactorial nature of the problem, a multidisciplinary approach is needed to treat patients with CMP. Multidisciplinary Pain Rehabilitation Programs (PRP) are effective in improving pain and daily functioning in patients with CMP 3, but improvement is modest. PRPs are highly variable and optimal content and dosage are unknown 4. To analyze the clinical situation of patients eligible for PRP, to set goals and to estimate dosage of PRP, clinicians assess the case complexity of patients. This study defined case complexity as the extent to which combinations of factors influence functioning of patients with CMP. It comprises both a number of factors and their relative weight. Theoretically, a patient presenting a large number of factors that each weigh heavily, is considered a more complex case than a patient with fewer factors that weigh lighter. Complex cases might need different treatment and different dosage of PRP than low complex cases. A patient presenting only one heavy weight factor could need high dosage of mono-disciplinary therapy, but will not be considered as complex case for this study, because there is only one factor influencing functioning. This patient is not eligible for PRP. Although never investigated, for the evaluation of PRPs, case complexity could be considered as potential moderator, because besides influencing content of treatment it might also influence dosage. However, we are presently unable to reliably and validly assess case complexity, and therefore unable to test relationships between case complexity and dosage of PRP. Case complexity of patients admitted to PRPs, is currently assessed ‘intuitively’ by clinicians, partly based on validated questionnaires. However, psychometric properties of the final complexity assessment are unknown. A standardized, transparent and psychometrically sound method assessing case complexity specifically for patients with CMP is unavailable. In the Netherlands, the Workgroup Pain Rehabilitation Netherlands (WPN) developed a four level classification system for case complexity of patients with CMP 5 (textbox 1). Patients with WPN level 3 or 4 are eligible for PRP. While this classification is considered as gold standard for the complexity assessment by rehabilitation specialists in the Netherlands, its criteria are not described, lack transparency and objectivity, and has moderate interrater reliability 6. Moreover, differentiation is low because levels are very wide, and PRP content and dosage cannot be derived from this classification. Thus, a standardized, transparent and psychometrically sound method to assess case complexity will help improve PRP scientifically and clinically.

Case complexity of patients with CMP | 93

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Textbox 1 Deﬁnitions of WPN levels (author’s translation) WPN level Criteria 1 Patients with pain and (imminent) limitations in activities without inﬂuence on (work) participation, no or minimal psychosocial problems. 2

Patients with pain and limitations in activities and (work) participation, no or minimal psychosocial problems.

Patients with pain and limitations in activities and (work) participation, psychosocial problems that have moderate to severe inﬂuence on the (self-reported) level of functioning.

Patients with pain and limitations in activities and (work) participation, psychosocial problems that have very severe inﬂuence on the (self-reported) level of functioning.

The objectives of this study were to: 1.

Explore the most important factors inﬂuencing case complexity of patients with CMP eligible for PRP.

Initiate development of a Case Complexity Index (CCI) based on factors, deemed most relevant by clinicians.

Explore means of assessment used to determine the relevance of CCI individual factors.

Test the CCI feasibility.

Materials and Methods A 3-round online Delphi study was conducted among clinicians working in PRPs in the Netherlands. Participants Rehabilitation specialists were selected through purposive sampling and invited by email. The rehabilitation specialists were asked to invite physiotherapists, psychologists and occupational therapists from their rehabilitation team to participate (snowball sampling). Rehabilitation specialists, physiotherapists, psychologists and occupational therapists who were for more than two years substantially (0.5 FTE or more) involved in a PRP for patients with CMP were considered experts for this Delphi study. The experts worked in rehabilitation centres spatially dispersed across the country. For the sample size in a Delphi study there are no strict rules, but a minimum sample size of 10-15 for a homogenous group of experts has been suggested 7. Because the sample consisted of four different disciplines, a minimal sample size of 10 in each discipline was chosen. First round Objective: to determine all factors that are assumed to inﬂuence the case complexity of patients eligible for multidisciplinary PRP. The experts were sent a link to an online survey. This survey consisted of one open question in which the experts were asked to list all factors that might inﬂuence case complexity.

94 | Chapter 6

Analysis Duplicates of answers were removed and the remaining answers were clustered into comprehensive factors as described in the International Classiﬁcation of Functioning, Disability and Health (ICF) 8

. The ﬁnal clustering and disposition of the factors were consented on by three authors (FW, EB

and MR). Second round Objective: to identify the most important factors for determination of the case complexity. The experts were sent a link to a second online survey, and they received the overview of the factors arranged in the domains of the ICF model. The experts were asked to select the 20 most important factors for the determination of the case complexity of patients eligible for a multidisciplinary PRP. They were asked to select the factors in such a way that the factors comprised the complexity of 80% of the patients. Analysis Per factor the number of selections was counted. If possible, factors were clustered under higher level factors in order to develop a manageable number of (comprehensive) factors for the third round. The number of selections of all factors was calculated and the top 10 factors with the highest sums were selected for the third round. Percentages of the total amount of selections per factor were calculated. The factors were all linked to ICF categories using the ICF linking rules 9. Third round Objectives: to weigh the top 10 most selected factors and investigate how these factors are assessed for a PRP. The experts were sent a link to a third online survey, which included the 10 factors with the highest sums of selections. This survey consisted of three questions: Question 1: Please rate to what extent the factor accounts for the case complexity on a 5-point Likert scale (0= no weight, 1= light, 2= average, 3= heavy, 4= very heavy). Question 2: Please indicate how you determine the relevance of each individual factor (1= based on measurement tool only (100%/0%); 2=more based on measurement tool than clinical reasoning and examination (75%/25%); 3=based on measurement tool as much as on clinical reasoning and examination (50%/50%); 4=more based on clinical reasoning and examination than measurement tool (25%/75%); 5=only based on clinical reasoning and examination (0%/100%)). An expert could also select the option ‘I do not determine this factor’. Question 3a: If you based your assessment (partly) on measurement tools, please list the measurement tools and the cut-off points you use. Question 3b: If you based your assessment (partly) on clinical reasoning and examination, please explain how you determine the relevance of the factor and which ‘cut-off points’ you use.

Case complexity of patients with CMP | 95

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Analysis Question 1: For each factor, the mean, standard deviation, median and interquartile range (IQR) of the ratings were calculated. The differences between the four different disciplines were analyzed by means of the Kruskal-Wallis test. The differences between rehabilitation specialists and the other disciplines as one group were analyzed by means of the Mann-Whitney U test. Question 2: medians and IQRs were calculated for each factor. Question 3: measurement tools and cut-off points were listed. Case Complexity Index The mean ratings of relevance (Q1, 3rd round), were used to calculate the weights. The weights were linearly transformed to obtain a CCI on a 0-100 scale. The formula for the complexity index was formed by calculating the sum of the products of the factors and their mean weights. The options for the scoring of every individual factor were 0 (not present), 0.5 (partially present) or 1 (fully present), or ‘cannot determine’. The minimum score of the presence of the factors was 0, calculated by scoring every factor 0. The maximum score was 10, calculated by scoring every factor 1. The minimum total CCI score is 0 (no presence of any factor) and the maximum score is 100 (all factors are fully present). Feasibility A feasibility study of the CCI was performed at the University Medical Centre Groningen, Centre for Rehabilitation. A convenience sample of 16 consecutive patients with CMP who underwent a clinical assessment for a PRP, was included. The rehabilitation specialists, occupational therapists, physiotherapists and psychologists acted as assessors by filling in a form following the clinical assessments. The rehabilitation specialists filled out the form independently from the other member of the team. The other members of the team filled out the form together during team meetings after completion of the individual assessments. The form was designed to identify the WPN level and the presence of the 10 factors of the CCI. The definitions of the WPN levels were provided for this pilot study, each WPN level was divided into three categories: ‘-‘ (light), ‘normal’ (typical) or ‘+’ (severe). For the CCI, the assessors scored every factor to what extent the factor was present, based the clinical judgement. The options were: ‘not present’ (score 0), ‘partially present’ (0.5), ‘fully present’ (1.0) and ‘cannot determine’. The option ‘cannot determine’ was considered as a missing value. Analysis CCI was calculated for every patient. To deal with the missing values, two methods were used: (1) Fully Conditional Specification Method in SPSS to replace the missing values (imputation), and (2) dividing the complexity index by the number of filled in factors (mean complexity index). The medians, means and standard deviations of the values of the factors, complexity index and WPN

96 | Chapter 6

level were calculated. Spearman’s rho was calculated between CCI and WPN level. To investigate the relevance of the factors without the weights, Spearman’s rho was calculated between the unweighted CCI (sum of the values of the factors for every patient) and the WPN level. The analyses were conducted separately for the rehabilitation specialists and treatment team. The interrater agreement between the rehabilitation specialists and treatment team on the WPN level and the assessment of the factors was assessed by calculating Kappa with linear weighting.

Results Participants A total of 52 experts participated in the Delphi study: 21 rehabilitation specialists, 12 physical therapists, 11 occupational therapists and 8 psychologists. Round 1: 44 experts; 17 rehabilitation specialists, 11 occupational therapists, 10 physical therapists and 6 psychologists. Round 2: 41 experts (response rate= 93.2%); 16 rehabilitation specialists, 9 occupational therapists, 9 physical therapists and 7 psychologists. Round 3: 36 experts (response rate= 81.8%); 13 rehabilitation specialists, 9 occupational therapists, 8 physical therapists and 6 psychologists. The experts worked in 15 rehabilitation centres in 11 out of 12 provinces in the Netherlands. Three nonresponding experts declared that lack of time was their reason for not responding. First round After removal of duplicates, the experts listed a total of 166 factors that were deemed important for the determination of the case complexity (Appendix 1). Second round All experts selected 20 factors which they deemed most relevant for determining the case complexity. Three experts selected less than 20 factors, which resulted in a total of 802 selections. The majority of unique selections were clustered into 25 (comprehensive) factors in the second round. The top 10 most selected factors are presented in table 1. These 10 factors represent 59.1% of the total amount of selections. Appendix 2 shows the deﬁnitions of the 10 factors.

Case complexity of patients with CMP | 97

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Table 1 Top 10 most important factors selected in the second round Factor

ICF domain

ICF Category

Psychiatric disorders Motivation Family system problems Mentalization capacity Somatisation Financial and work-related problems Features of complaints Treatment-interfering personality traits Life events Personal injury proceeding Total selections / %

ICD Body Functions External Factors Body Functions Body Functions External Factors Body Functions Body Functions Personal Factors External Factors

b130 e310 b164 b1602 e590 b280 b126 nc e550

N selections

% of total amount of selections

123 70 47 43 42 34 34 30 27 24 471

15.3 8.7 5.9 5.4 5.2 4.2 4.2 3.7 3.4 3.0 59.1

ICF: International Classification of Functioning, Disability and Health; ICD: International Classification of Diseases and Related Health Problems, Nc: not classified

Third round The results of the ratings of relevance of the individual factors are presented in table 2. The mean ratings ranged from 1.8 for ‘features of complaints’ to 3.6 for ’psychiatric disorders’. The mean ratings per discipline are not presented, because differences in medians between the four disciplines were non-signiﬁcant. The rating of relevance per factor is shown in ﬁgure 1. Table 2 Ratings of relevance of factors, differences between disciplines and between rehabilitation specialists and treatment team. Factor Psychiatric disorders Motivation Treatment-interfering personality traits Somatisation Family system problems Life events Mentalization capacity Personal injury proceeding Financial and work-related problems Features of complaints a

Meanb 3.6 3.2 3.1 2.8 2.6 2.4 2.4 2.4 2.3 1.8

SD Medianb IQR 0.65 4 1 1.03 3.5 1 0.67 3 0 0.75 3 1 0.87 3 1 0.84 2 1 0.87 2 1 0.72 2 1 0.69 2 1 0.84 2 1

All disciplines h (df=3)c p 5.14 .162 7.70 .053 6.70 .082 3.14 .371 2.34 .505 4.12 .249 4.15 .245 0.90 .826 0.16 .984 3.17 .367

RSa vs team Ud p 144.0 .831 74.5 .007 145.5 .877 103.5 .099 119.5 .288 148.5 .972 104.0 .107 129.0 .461 147.5 .941 101.5 .090

RS = Rehabilitation Specialists Options of 5-point Likert scale for importance: 0= no weight, 1= light, 2= average, 3= heavy, 4= very heavy. c h= test-statistic Kruskal Wallis test for differences between all disciplines. d U=test-statistic Mann-Whitney U test for differences between rehabilitation specialists and treatment team. b

98 | Chapter 6

Psychiatric disorders

Motivation

Treatment-interf ering personality traits

Somatization Rehabilitation Specialists Family system problems

Occupation therapists Physical therapists

Lif e events

Psychologists Total

Mentalization capacity

Personal injury proceeding

Financial and work-related problems

Features of complaints

Figure 1 Mean rating of importance of factors per discipline (0=no weight, 1=light, 2=average, 3=heavy, 4=very heavy) Factors are ordered by decrease of mean rating. Disciplines are divided and ordered by rehabilitation specialist and the rest of the team.

The results of the methods of assessment of the factors are shown in table 3. For all factors, the medians were 3 or higher, indicating a tendency towards a dominance of clinical examination and reasoning. In total 40 different instruments were named. The number of instruments that are used to measure a factor ranged from 0 (life events) to 17 (treatment interfering personality traits). In 49 cases cut-off points were not mentioned for an instrument used for a speciﬁc factor. The most mentioned cut-off point for clinical examination and reasoning is ‘negative inﬂuence of factor on functioning’.

Case complexity of patients with CMP | 99

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Table 3 The determination of the factors (n=36). Factor Psychiatric disorders Motivation Treatment-interfering personality traits Somatisation Family system problems Life events Mentalization capacity Personal injury proceeding Financial and work-related problems Features of complaints

Mediana 3 5 3 4 5 5 5 5 5 4

IQR 1 1 1 1 1 1 0 0 1 2

Missing (%)b 36.1 0.0 27.8 5.6 22.2 22.2 25.0 33.3 30.6 8.3

Options of 5-point Likert scale for determination: 1=only based on measurement tool, 2=more based on measurement tool than clinical reasoning, 3=based on measurement tool as much as on clinical reasoning, 4=more based on clinical reasoning than measurement tool and 5=only based on clinical reasoning. b The missing values represent the option ‘I do not determine this factor’

Case Complexity Index The CCI is based on the 10 factors (round 2) and their weights (0-4; 3rd round). The weights were linearly transformed to a 0-100 scale by multiplying the mean ratings by 3.79 (= 100/ sum weights). The final formula for the CCI = 13.5*psychiatric disorders + 12*motivation + 11.6* treatment-interfering personality traits + 10.6*somatisation + 10*family system problems + 9.2*life events + 9.1*mentalization capacity + 8.9*personal injury proceeding + 8.5*financial and work-related problems + 6.6*features of complaints. The options for the scoring of every factor are 0 (not present), 0.5 (partially present) or 1 (fully present), or ‘cannot determine’ (missing value). The minimum score is 0 (no presence of any factor) and the maximum score is 100 (all factors are fully present). A higher CCI indicates a more complex case. The factors can be scored on a CCI form (table 4). For example, a fictive patient scored 1 on motivation, mentalization capacity and financial and work-related problems, and scored 0.5 on family system problems, somatisation, treatment-interfering personality traits and life events. The scores were multiplied with their corresponding weights, summed and CCI of this patient is 50.3.

100 | Chapter 6

Table 4 Example of the CCI for a ﬁctive patient.

Psychiatric disorders Motivation Treatment-interfering personality traits Somatisation Family system problems Life events Mentalization capacity Personal injury proceeding Financial and work-related problems Features of complaints CCI

Not present 6 6 6

Optionsa Partially Fully Cannot Score Weight present present determine 0 13.5 6 1 12.0 6 0.5 11.6 6 0.5 10.6 6 0.5 10.0 6 0.5 9.2 6 1 9.1 0 8.9 6 1 8.5 0 6.6

Score* weight 0.0 12.0 5.8 5.3 5.0 4.6 9.1 0.0 8.5 0.0 50.3

a Not present=0, Partially present=0.5, Fully present=1, Cannot determine=missing value. CCI=Case Complexity Index

Feasibility In the feasibility study 16 patients were assessed. Fourteen patients were assessed by the treatment team and, 14 patients were assessed by the rehabilitation specialists. Twelve patients were assessed by both the rehabilitation specialists and treatment team. The patients had a mean age of 41.2 years (SD=15.1), 7 were female, and 8 patients were diagnosed with lumbago, 6 with wide spread pain and 2 with chronic pain syndrome. The mean Pain Disability Index (PDI) 10

score of the patients was 39.5 (SD=11.5, range: 16 to 56). The medians, means and standard

deviations of the values of the factors, CCI and WPN level are shown in table 5. The correlations between CCI and WPN were strong and significant for the treatment team, lower and nonsignificant for the rehabilitation specialists. Table 6 shows both methods used for missing values: unweighted CCI, (mean CCI was used for missing values), and CCI (missing values were imputed). The agreement of the assessment of the WPN level between the rehabilitation specialists and the treatment team was non-significant (k=0.17, 95% CI 0.000 to 0.402). The agreement of the assessment of the CCI between the rehabilitation specialists and the treatment team was k=0.48 (95% CI 0.355 to 0.607), which shows a moderate agreement

. All assessors were able to

comprehend the CCI form and its concepts with brief written instructions only. The time needed to ﬁll out the CCI was less than 1 minute per patient.

Case complexity of patients with CMP | 101

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Table 5 Descriptives of the scores on the factors of the CCI and WPN level in the feasibility study Rehabilitation specialists (n=14a) Psychiatric disorders Motivation Treatment-interfering personality traits Somatisation Family system problems Life events Mentalization capacity Personal injury proceeding Financial and work-related problems Features of complaints CCI (missing values imputed) WPN level

Median 0.25 0.00 0.50 0.50 0.50 0.50 0.00 0.00 0.50 1.00

Mean 0.25 0.21 0.50 0.36 0.46 0.54 0.18 0.04 0.54 0.86 37.32

Treatment team (n=14a) SD 0.26 0.38 0.28 0.36 0.24 0.37 0.32 0.13 0.43 0.23 16.33

Median 0.00 0.00 0.50 0.50 0.50 0.50 0.00 0.00 0.50 1.00

3/3+

Mean 0.31 0.25 0.61 0.43 0.46 0.65 0.29 0.04 0.50 0.71 41.82

SD 0.43 0.38 0.35 0.43 0.37 0.38 0.38 0.13 0.44 0.38 19.76

CCI=Case Complexity Index Table 6 Correlations between complexity and WPN level using different methods CCI.

Unweighted CCI CCI (missing values imputed)

WPN level Rehabilitation specialists r p 0.45 0.146 0.43 0.134

Treatment team r p 0.73 0.003* 0.72 0.003*

WPN level: r= spearman’s rho,CCI=Case Complexity Index, * p<0.05

Discussion This study explored the most important factors to assess case complexity of patients with CMP eligible for PRP, according to the experts in the Dutch chronic pain rehabilitation field. These factors were used to initiate the development of a CCI. The CCI consists of 10 factors: financial and workrelated problems, features of complaints, personal injury proceeding, life events, mentalization capacity, motivation, personal injury proceeding, psychiatric disorders, somatisation and family system problems. Of these 10 factors, according to the experts, psychiatric disorders have the most influence on the case complexity and features of complaints have the least influence. The four professions agreed on the ratings of influence of the factors. The assessment of the factors is mainly based on clinical examination and reasoning. Feasibility of the CCI was established. The factors in our study are mainly psychosocial factors. Indicating that the experts rate the biological factors to be of lower relevance in the assessment of the case complexity of patients eligible for PRP. This is remarkable considering PRPs are based on the biopsychosocial model. The

102 | Chapter 6

results of this study showed that there are no unambiguous measurement tools for the factors, which is in line with a systematic review that there is a large variability in measurement tools for predictors of outcome of multidisciplinary PRPs for patients with chronic low back pain 12. These predictors overlapped partially with the factors in the current study. The CCI is based on a combination of 10 factors, which is a number chosen by the authors to construct a manageable tool. These factors account for 59.1% of the total number of selections. Therefore, it does not capture all factors that determine case complexity, but it does contain the most important factors as viewed by the experts. Also, the case complexity does not take the interaction between the factors into account. The goal of this Delphi study was to gain insight into the assessment of the case complexity based on opinions of experts and take ﬁrst steps to create a new concept for the assessment of case complexity based on this insight. Therefore, the results of this Delphi study do not represent “the truth”. It presents an expert view about factors of case complexity, and a state of the art on how these factors are assessed in rehabilitation practice. The experts were all from the Netherlands, which may limit generalizations to pain rehabilitation settings in other countries.

The CCI might be a useful tool for the assessment of the case complexity of patients with CMP. It is transparent, easy to score, takes very little time to fill out and provides a numeric value to the case complexity. The complexity index can be used by the rehabilitation specialists and other members of the treatment team. For the assessment of the treatment team, the results of the pilot study show that the CCI is associated with the WPN level. The CCI is currently still a rather subjective measure, just as the gold standard WPN method, because the decision on the presence of each factor is based on clinical reasoning. However the criteria of the CCI are more specific, which can provide a basis for structured assessment of case complexity and stimulation to validate the decision on assessing each factor. Other advantages of the CCI compared to the WPN level is the better differentiation between patients, because of its numeric value (0-100) compared to (broad) categories and the possibility to convert the factors into treatment goals. On the limitations side, contra-indications for the treatment were not considered. Factors may be too dominantly present that it cannot be treated with the expertise of a typical PRP team. For example, a patient with a dominantly present personality disorder that interferes with PRP may not benefit from a PRP, and may be better referred to other specialists. Also one factor can be heavily present, needed much therapy, but need other therapy than multidisciplinary PRP. The present CCI does not account for this. Some other attempts to capture complexity have been described the Rehabilitation Complexity Scale 13, the Vector Model of Complexity 14, COMPRI 15 and INTERMED 16. However none of these methods are designed to assess case complexity specifically for patients with CMP. As stated 17,

Case complexity of patients with CMP | 103

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

there is need for more research to assess case complexity for more specialist rehabilitation programs for individual patients within the rehabilitation process. The CCI is a more specific measure, only addressing relevant factors for the clinical assessment of patients with CMP who are eligible for a multidisciplinary PRP and includes factors derived from experts in the field. Therefore, the factors of the CCI and its score seem to be more appropriate for the use within this specific setting of multidisciplinary PRP. Our study showed that the assessment is mainly based on clinical examination and reasoning. Also, the used cut-off points for the clinical examination and assessment are highly variable for each factor. Because the clinical examination and reasoning will be influenced by the assessors’ experiences and beliefs and various implicit cut-off points are used, it is challenging to reach consensus about the case complexity of an individual patient. In this way, the clinical examination and reasoning of the assessor will likely influence the determined case complexity of a patient. This can still be the problem with the CCI, which showed a fair interrater agreement between the rehabilitation specialists and treatment team. However, these results are based on a small sample size and many different assessors were involved. The definitions of the factors were also not given to the assessors, which may have lead to variability in the interpretation of the factors. A few assessors were also involved in the Delphi study and therefore had more knowledge of the factors. These exploratory analyses were performed within the context of feasibility. Clearly, formal studies are needed to study psychometric properties of the CCI. The first steps of the development of the CCI have been reported in this paper. However, we realize that the present CCI is not an end-product. The contribution of each factor to the case complexity should be tested in a sensitivity analysis. In this analysis other factors should also be considered. To improve the CCI, it is recommended to objectify the assessment of the individual factors. For some factors validated instruments are available and these could be incorporated in the CCI. It might be necessary to have an extra option between ‘partially present’ and ‘fully present’ for the scoring of the presence of the factors (e.g. similar to the ICF qualifiers), as there is a large gap between those options. Because the present CCI represents 59.1% of the case complexity, the addition of an extra open factor may be considered to facilitate for the 40.9% that are unaccounted for. Contra-indications for the treatment should be explored and added to the CCI. The validity, reliability and the practicality of the CCI need to be tested in several settings to ensure the generalizability within chronic pain rehabilitation. Future research is also needed to explore the usability of the total CCI score compared to the scores and weights per factor and analyse relationships of case complexity with dosage of PRPs.

104 | Chapter 6

Acknowledgements The authors acknowledge the participation of all experts in the Delphi study and the participation of the rehabilitation physicians and the other team members in the feasibility study.

Case complexity of patients with CMP | 105

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1.

Gatchel RJ, Peng YB, Peters ML, Fuchs PN, Turk DC. The biopsychosocial approach to chronic pain: Scientiﬁc advances and future directions. Psychol Bull. 2007;133(4):581-624.

Dansie EJ, Turk DC. Assessment of patients with chronic pain. Br J Anaesth. 2013;111(1):19-25.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

Köke A, Brouwers M, Heuts P, et al. Consensus rapport pijnrevalidatie Nederland. 2005.

Boonstra AM, Schiphorst Preuper HR, Brouwer MAH, Weerdt de M, Knol D, Köke A. Classiﬁcatie van patiënten met chronische pijnklachten: Onderzoek naar de betrouwbaarheid van de WPN niveaus. Ned T Pijnbestrijding. 2011;30:5-10.

Delbecq AL, Van de Ven AH, Gustafson DH. Group techniques for program planning: A guide to nominal group and delphi processes. Glenview, Ill: Scott, Foresman; 1975:174.

World Health Organization. ICF: International classiﬁcation of functioning, disability and health. Geneva: World Health Organization; 2001:299.

Cieza A, Geyh S, Chatterji S, Kostanjsek N, Ustun B, Stucki G. ICF linking rules: An update based on lessons learned. J Rehabil Med. 2005;37(4):212-218.

10.

Tait RC, Chibnall JT, Krause S. The pain disability index: Psychometric properties. Pain. 1990;40(2):171182.

11.

Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159-174.

12.

van der Hulst M, Vollenbroek-Hutten MM, Ijzerman MJ. A systematic review of sociodemographic, physical, and psychological predictors of multidisciplinary rehabilitation-or, back school treatment outcome in patients with chronic low back pain. Spine (Phila Pa 1976). 2005;30(7):813-825.

13.

Turner-Stokes L, Williams H, Siegert RJ. The rehabilitation complexity scale version 2: A clinimetric evaluation in patients with severe complex neurodisability. J Neurol Neurosurg Psychiatry. 2010;81(2):146-153.

14.

Safford MM, Allison JJ, Kiefe CI. Patient complexity: More than comorbidity. the vector model of complexity. J Gen Intern Med. 2007;22 Suppl 3:382-390.

15.

Huyse FJ, de Jonge P, Slaets JP, et al. COMPRI--an instrument to detect patients with complex care needs: Results from a European study. Psychosomatics. 2001;42(3):222-228.

16.

de Jonge P, Huyse FJ, Slaets JP, Sollner W, Stiefel FC. Operationalization of biopsychosocial case complexity in general health care: The INTERMED project. Aust N Z J Psychiatry. 2005;39(9):795-799.

17.

Wade D. Complexity, case-mix and rehabilitation: The importance of a holistic model of illness. Clin Rehabil. 2011;25(5):387-395.

106 | Chapter 6

Case complexity of patients with CMP | 107

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 110 | Chapter 7

The main aim of this thesis was to acquire insight into aspects of dosage for multidisciplinary Pain Rehabilitation Programs (PRPs) for patients with chronic pain. Five studies were performed. In this concluding chapter, the main ﬁndings and factors related to PRP dosage will be discussed, followed by clinical, scientiﬁc and societal implications, strengths and limitations of this thesis and recommendations for future research.

Main Findings A systematic review and meta-analysis (Chapter 2) revealed that PRP dosage had never been studied as a primary objective within pain rehabilitation research. Analyzing studies that investigate the effects of PRPs and comparing the reported dose variables resulted in an overview of PRPs with a wide variety of dosages having different outcome measures and effects on disability, quality of life (QoL) and work participation. Evaluation moment, number of disciplines, type of intervention, duration of intervention, percentage of women, and age influenced the effect of PRP. However, because dose variables and content variables were strongly interrelated, a distinction between these variables regarding their effects was impossible. Chapter 3 describes a clinical study in which the course of disability reduction during PRP was analyzed. At a group level, the course of disability reduction occurred following a quadratic time model. This course was influenced by initial Pain Disability Index (PDI) scores, treatment week, average pain, and the interaction between initial PDI scores and treatment week, resulting in, for example, a steeper slope (faster disability reduction) for patients with higher initial PDI scores and average pain. Chapter 4 describes a RCT in which the effects of PRPs with similar content but different dosages were compared. Frequent extension of the Care as Usual – Short Form (CAU-SF) program led to exceeding predefined stopping rules, and the study had to be stopped prematurely. Because of lack of power and extension of CAU-SF, differences in dosages between the control and experimental group were too small to detect significance of non-inferiority. Hence, no robust conclusions could be drawn. Ancillary regression analysis showed that dose variables did not contribute to the prediction model, which could be interpreted as a possibility of shortening PRP without loss of effect. Chapter 5 describes a study in which patient and therapists’ perceptions were explored regarding dosages of PRP in three rehabilitation centers in the Netherlands with different dosage of PRP. Factors that determined the dosage were related to patient, treatment and external factors. The results showed that although PRPs differed in dosage, perceptions of factors that were important

General discussion | 111

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

for determining dosage were similar. Remarkably, similar factors seemed to lead to different choices in terms of PRP dosage in different centers. Chapter 6 describes a study in which factors related to the general status, defined as case complexity of patients eligible for PRP, were explored. The 10 most important factors were listed and weights of influence were explored, resulting in a case complexity index (CCI), which was tested for feasibility in clinical practice. Our hypothesis of a possible relationship between case complexity and dosage was consistent with the theme of “general status” from the interviews described in Chapter 5. Patients and rehabilitation professionals mentioned a relationship between dosage and the general status of a patient, which includes factors similar to the factors in the CCI. Therefore, it could be assumed that the CCI could be tested in clinical practice to determine case complexity and to relate this to the prediction of PRP dosage. The relationship between case complexity and dosage should be tested to confirm or reject this hypothesis.

Factors associated with PRP dosage While these five studies have not led to definitive conclusions about the optimum PRP dosage, it has contributed relevant insights regarding factors related to dosage and the effects of PRP. These can be divided into 1: factors related to the effect of PRP; 2: patient related factors; 3: PRP related factors; 4: common therapeutic factors and, 5: external factors. Factors related to the effect of PRP Literature and studies performed in this thesis showed a wide variety in PRP dosage 1,2. To determine the optimum PRP dosage, the question of “how much therapy is needed to establish the desired effect” should be answered. However, up until now, most studies have analyzed effects of PRP at a certain dosage by analyzing pre/post and follow-up outcomes 3-6, not knowing whether this effect could also be reached with a lower dosage or a greater effect could be reached at a higher dosage. Generally, studies do not explicitly target a certain effect, nor do they explicitly adjust dosages towards this target effect. Additionally, the question arises regarding which effect should be analyzed with respect to PRP: the effect of PRP on improving participation or the effect patients reach at a certain moment using better self-management skills acquired from PRP. For example, can PRP considered effective if a patient does not acquire a clinically important change on PDI, but shows good skills to self-manage his disabilities in the future? Moreover, multidimensional effects of multidisciplinary therapies, including psychological components, are more difficult to define, compared to the unidimensional effect of mono-disciplinary therapy, including mainly physical components. Additionally, it is easier to measure “physical condition” compared to measuring ”the ability to improve physical condition”. The current literature shows a wide variety

112 | Chapter 7

of outcome measures and domains addressing the effect of PRP in relation to participation, such as disability reduction, QoL and work participation 1,2,7. However, other outcome measures related to self-management skills should be added to assess the effect of PRP and could be used to assist with determining the optimum dosage. Being able to self-manage pain-related disabilities would presumably effect participation in daily life 8.Measuring self-management skills as an effect of PRP, in addition to rehabilitation programs in general, is in line with changes in the new concepts of health. Instead of describing health as “a state of complete physical, mental and social wellbeing and not merely the absence of disease or infirmity” 9, Huber et al. stated that health is “the ability to adapt and to self-manage, in the face of social, physical and emotional challenges” 10. Regarding the effects of PRP, it can be debated from which perspective the optimum effect should be targeted – optimum from the perspective of the patient, from the rehabilitation professionals or from an economic perspective. In the current system of healthcare finance in the Netherlands, and in some cases, from an economic perspective, a higher PRP dosage may provide more income for the rehabilitation provider, while from the perspective of a patient and/ or rehabilitation professional, the desired effect may have already been reached. If the optimum dosage is defined from this economic perspective, overtreatment could occur. For these reasons, more insight is needed regarding which factors are important to determine optimum PRP dosage. Additionally, scientific evidence and consensus regarding target effects of PRP are needed to improve comparability and the ability to determine optimum PRP dosage. Patient related factors Dosage of PRP could be related to the patients’ case complexity at the beginning and during PRP. Chapters 2, 3, 5 and 6 showed that several patient characteristics and patient-related factors, both physical and psychological, can, or are assumed to, influence the effect of PRP and the required dosage. Taking together the findings of this thesis and the questions regarding the effects of PRP, there could be an association between the stages of change

and PRP dosage.

The transtheoretical model of changing behavior presumes that the readiness of change can be described in different stages of change. During the precontemplation stage, patients believe that chronic pain is primarily a biomedical problem and rehabilitation professionals could help relieve the pain during PRP 12. Therefore, they are not willing to change behavior and are not motivated to take the biopsychosocial approach offered in PRP. At the contemplation stage, patients are aware of the biopsychosocial nature of their problem, but do not know how to manage this. Therefore at this stage, a patient needs professionals help to manage the pain and related disabilities. In the later stages (preparation and action stages), patients have learned how to self-manage their pain and disabilities and they need to maintain this healthy behavior. Being able to self-manage painrelated disabilities and to make plans for maintaining healthy behaviors will theoretically result in less need for professional help, and therefore, less, or even no, PRP. Self-management skills will

General discussion | 113

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

theoretically also result in late-treatment improvements in outcomes of reduction of experienced disability, QoL and work participation. This is in line with the results from previous studies A recent study

12,13

analyzed changes in precontemplation and action attitudes during cognitive-

behavioral therapy (CBT) for patients with chronic pain. From this study, it was concluded that there might be an association between changes in the stages of behavior modification and the following outcome improvements. However, at present, this association has not been proven regarding PRPs, there has been no documented association related to PRP dosage. In addition, measuring the stage of change of a patient at a certain moment is still difficult and arbitrary 14. PRP related factors PRP itself, as well as the rehabilitation professionals involved in PRP, could influence dosage. PRP is offered at a certain dosage – total duration in weeks, total number of contact hours, number of rehabilitation professionals involved, form of intervention and physical dosage of exercise (weight, repetition, duration, frequencies). These dosage variables differ across PRPs, nationally and internationally. From the systematic review and the interviews conducted in Chapters 2 and 5 respectively, it can be concluded that there is a lack of scientific evidence regarding on which factors the offered PRP dosage is based. Choices of dosages of PRPs included in Chapter 5 are mainly based on clinical expertise and historical grounds. Remarkably, both rehabilitation professionals and patients are generally satisfied with the offered and received PRP, regardless of the large differences in offered dosages. Patient satisfaction was studied in a systematic review within physical therapy care 15, concluding that therapist characteristics, the process of care, and the organization of care were consistently identified factors that influenced satisfaction with care, whereby treatment outcome and expectations of physical therapy were less-frequently reported factors. Dosage can be expressed in terms of total duration, in addition to expressing it in terms of total number of contact hours. For example, PRP can be offered for the same number of weeks (total duration), but with a greater or fewer total number of contact hours. These differences in the total duration and corresponding number of contact hours per week are based on different choices regarding the content of contact hours, the number of rehabilitation professionals, and the form of PRP. The choice of content of contact hours, for example, is based on the assumption of what content should be offered within the contact hours of PRP and which activities or homework should be performed by the patient himself outside the rehabilitation center. Apparently, rehabilitation teams have made different choices regarding offered content and related dosages based on different assumptions and the best available evidence 1. Several studies have been performed to analyze components of PRP, such as differences between outpatient and inpatient PRP graded activity versus graded exposure

and

. However, evidence for the optimum composition of

PRP and related optimum PRP dosages, including all components and factors, is lacking.

114 | Chapter 7

Common factors Common therapeutic factors 18 are described as possible explanations for equivalent effectiveness across different treatment approaches within psychotherapy. Rosenzweig wondered whether “the factors alleged to be operating in a given therapy are identical with the factors that actually are operating, and whether the factors that actually are operating in several different therapies may not have much more in common than have the factors alleged to be operating”. Examples of common factors are reassurance, feedback, insight, success experiences, therapeutic alliances, patient expectations, therapist expertness, and trust 19. Because of the biopsychosocial characteristics of PRP, common factors could probably inﬂuence the effects and PRP dosage. The results of the qualitative study in Chapter 5 showed that these common factors are similar but lead to different choices in PRP dosage. However, little evidence is available regarding speciﬁc and common factors and their relationship to outcomes of chronic pain treatment

14,20-22

and its

dosage. Overall, this thesis opened a black box in pain rehabilitation research. It is the ﬁrst step in unraveling the complexities of PRP dosages. Still, it is a huge challenge to answer the question of optimum PRP dosage, taking into account the relationship of dosage and content, including all of the above-mentioned factors. Adding the factor of dosage to what Burns

stated regarding

the working mechanisms for CBT, “it is not yet clear whether CBT worked for reasons specified by theory, through the inspiration of hope and the support and encouragement of a good therapist, or a combination of factors”. External factors The legislation of expenses and reimbursements in healthcare could lead to differences in PRP dosages across countries. Direct and indirect costs become more and more important. In general, PRPs have proven cost effectiveness 23,24, but it is a challenge to balance how much treatment is needed versus its costs. Because the effect of PRP is not strictly defined, costs of PRP can be the motivation for offering a certain PRP dosage in order to ensure cost benefits. In 2014, the Dutch government started, in cooperation with the Dutch Association of Physicians in Rehabilitation Medicine, to develop a new system of financing for rehabilitation medicine. Physicians across the Netherlands described the content of rehabilitation treatment according to different therapy components. Registering contact hours linked to these therapy components will lead to more transparency for all rehabilitation activities performed during rehabilitation programs across the Netherlands. Based on the combination of content and contact hours, a new finance system will be implemented before 2019. The present finance system is only based on the number of contact hours divided across categories, regardless of the content or effect of therapy. The adjustment of content in terms of therapy components related to contact hours of rehabilitation programs will probably draw attention and discussion regarding dosage of programs because of the link to

General discussion | 115

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

funding. Consequently, if dosages will be adapted, this will lead to more comparability of these programs across the Netherlands. This thesis showed that PRP dosage is associated with many implicit and un- or under-researched factors. Therefore, it is presently difﬁcult to draw robust conclusions about the optimum PRP dosage, and further research is needed to acquire more insight into, and scientiﬁc evidence of, differences and similarities between these factors and their interrelationships.

Implications Clinical implications In general, multidisciplinary PRPs have proven their effectiveness

1,6,7,25

. However, no distinction

can be made between different PRPs. This thesis resulted in a critical look at the different aspects of dosage, and it should encourage clinical practice and research to focus more on PRP dosage. Implications for patients Patients should be made aware of the different dosages. In the long term, research aimed at optimizing PRP dosage will benefit patients because healthcare providers will be better able to estimate which PRP dosage is required for subgroups of patients. Eventually, patients will receive the most effective and efficient PRP. Implications for therapists Therapists involved in PRP should reflect on the dosage of the offered program. They have to be aware of the differences of dosage within PRPs and they should be aware of the distinction between what is “nice to have or need to have”. Even though it seems like there is always something to practice or teach, there is room for improvement and patients have the right to receive the optimum care related to content and dosage of treatment. Over- or under-dosage of treatment will not benefit anyone. Therapists should integrate all factors related to the patients’ needs and preferences, as well as their professional experience, and they should be aware of the influence of common factors, costs and the available evidence to offer optimum PRP dosage. Implications for rehabilitation centers Insight into PRP dosage and being able to offer optimum PRP dosage will benefit the entire rehabilitation process. It will benefit the patient flow, and will reduce waiting lists and over- or underdosing treatment. As a result, centers will be able to treat the maximum number of patients suffering from chronic pain and offer cost-effective PRPs. The issue of dosage studied in this thesis, while studied within the pain rehabilitation context, is not unique for pain rehabilitation only. The

116 | Chapter 7

results of this thesis can be used as an example for other rehabilitation programs. Rehabilitation programs assist in improving health in terms of “the ability to adapt and self-manage,” and all rehabilitation programs can use this concept to offer more cost-effective programs. Scientiﬁc implications The results of this thesis, combined with the present developments regarding costs of healthcare, indicate that dosage of rehabilitation programs is important to include in analyzing these programs. Dosage of PRP has been a neglected topic in pain rehabilitation research, as well as in rehabilitation research in general. Previous studies analyzing the effect of PRPs could be confounded, and therefore, conclusions should be interpreted with caution. A theoretical example: a study comparing the effectiveness of physical therapy to PRP. The total duration of both programs is 10 weeks; however, physical therapy is given two hours a week and PRP is given 10 hours a week. PRP showed superior results, but it is questionable whether the content or the dosage is superior, or whether other common factors were responsible for differences in the effects. More research is needed to analyze what and how much works for whom. Societal implications Societal implications in general The results of this thesis contribute to the ability to optimize PRP dosage and other rehabilitation programs in general. Obtaining evidence of optimum dosage will lead to the reduction of overor underdosing treatment, thereby preventing the overdosage of treatment that leads to more “needs”. Over- or underdosing also leads to the opposite of self-management. Implications for (work)participation The effect of patients being able to self-manage their pain will subsequently optimize participation in daily life. Insight into “how much works for whom” will lead to the most desirable effects at the optimum PRP dosage for each individual. This will allow them to participate in daily life and return to work as soon as possible. Financial implications Because this thesis is an initial attempt at unraveling the complexities related to PRP dosage, at this stage, no ﬁnancial implications can be described since there is no solid evidence on the dose–effect relationship.

General discussion | 117

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Future research Future studies in rehabilitation research should report dosage of treatment. This will improve transparency, comparability and generalizability of rehabilitation treatment dosages and will give researchers the ability to analyze dose–effect relationships. Within pain rehabilitation research, future studies should also focus on the distinction between desired effect of PRP itself to be able to finish PRP and the desired long-term effect of the patient self-management skills regarding disability, QoL and (work) participation. Dosage of PRP should be analyzed together with all specific and aspecific factors, such as a core set of working mechanisms of effective PRPs. In the beginning, analyzing optimum PRP dosages appeared to be difficult. However, this thesis has yielded an overview of dosage-related factors. The results can be combined into new hypotheses regarding PRP dosages. The course of disability, which is reduced over time, could be related to the desired effect and the stages of change and needs of patients during PRP. This qualitative study confirmed the relationship between case complexity and dosage. The Delphi study gives substance to case complexity with the CCI, which can be used in clinical practice in relation to the required dosage, and can also be used to scientifically analyze this relation. Moreover, the RCT provides evidence that both arms with different dosages resulted in positive effects. This coincides with the positive experiences of the PRP at different dosages used in the qualitative study, as well as with the overview of effect sizes in the systematic review. These insights could be further analyzed to determine whether there is a trend towards shortening PRPs. Future research could result in more scientific evidence for optimum dosage, which will lead to better insight into what works for whom and how much therapy is needed. This will accomplish the optimum effects at the optimum dosage, with an addition focus on cost-effectiveness in terms of direct and indirect healthcare costs.

Strengths and limitations The research questions from this thesis evolved out of clinical practices, resulting in strengths and limitations since the methodological strictness is not always compatible with clinical practice. Because the studies in this thesis were the ﬁrst that analyzed PRP dosage, and no evidence was available beforehand mistakes were made. Some choices made became weaknesses. Most studies, except the systematic review, were performed in the Netherlands within the context of the Dutch healthcare system, which limits generalization. However, although the Dutch healthcare system differs from the healthcare systems in other countries, PRP dosage and rehabilitation in general is an important issue all over the world. This thesis could be used to perform similar studies within different countries to optimize dosage of therapy in each particular country. A strength of the thesis is the variety of methods used. Qualitative and quantitative methods were used to reveal

118 | Chapter 7

different factors related to PRP dosage, resulting in an overview of important factors that now need to be analyzed in future research to unravel and improve optimum PRP dosages.

Conclusion This thesis is the ďŹ rst that focuses on PRP dosage. Consequently, it should be regarded as a starting point for determining factors related to dosages and exploring optimum PRP dosages. This thesis analyzed dosage of PRP using different research methods and acquired different insights in aspects of dosage. Therefore, this thesis contributes to the body of research in the area of pain rehabilitation research aimed at analyzing the effects of all components and their dosages. Additionally, results from the studies within this thesis contribute to societal developments in transparency, comparability and cost effectiveness of PRPs and can be a basis for others to analyze the dosages of rehabilitation programs.

General discussion | 119

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

References 1.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidisciplinary rehabilitation for chronic low back pain: Systematic review. BMJ. 2001;322(7301):1511-1516.

McCracken LM, Turk DC. Behavioral and cognitive-behavioral treatment for chronic pain: Outcome, predictors of outcome, and treatment process. Spine (Phila Pa 1976). 2002;27(22):2564-2573.

Gatchel RJ, Okifuji A. Evidence-based scientiﬁc data documenting the treatment and cost-effectiveness of comprehensive pain programs for chronic nonmalignant pain. J Pain. 2006;7(11):779-793.

Scascighini L, Toma V, Dober-Spielmann S, Sprott H. Multidisciplinary treatment for chronic pain: A systematic review of interventions and outcomes. Rheumatology (Oxford). 2008;47(5):670-678.

van Geen JW, Edelaar MJ, Janssen M, van Eijk JT. The long-term effect of multidisciplinary back training: A systematic review. Spine (Phila Pa 1976). 2007;32(2):249-255.

Jensen MP, Nielson WR, Turner JA, Romano JM, Hill ML. Changes in readiness to self-manage pain are associated with improvement in multidisciplinary pain treatment and pain coping. Pain. 2004;111(12):84-95.

WHO. http://apps.who.int/gb/bd/PDF/bd47/EN/constitution-en.pdf?ua=1&ua=1. Updated 20052015.

10.

Huber M. Towards a new, dynamic concept of health. its operationalisation and use in public health and healthcare, and in evaluating health effects of food.; 2014.

11.

Prochaska JO, DiClemente CC. Transtheoretical therapy: Toward a more integrative model of change. Psychotherapy: Theory, Research & Practice. 1982;19(3):276-288.

12.

13.

Glenn B, Burns JW. Pain self-management in the process and outcome of multidisciplinary treatment of chronic pain: Evaluation of a stage of change model. J Behav Med. 2003;26(5):417-433.

14.

15.

Hush JM, Cameron K, Mackey M. Patient satisfaction with musculoskeletal physical therapy care: A systematic review. Phys Ther. 2011;91(1):25-36.

16.

Boonstra AM, Reneman MF, Schiphorst Preuper HR, Waaksma BR, Stewart RE. Differences between patients with chronic musculoskeletal pain treated in an inpatient or an outpatient multidisciplinary rehabilitation program. Int J Rehabil Res. 2014;37(2):187-191.

17.

Leeuw M, Goossens ME, van Breukelen GJ, et al. Exposure in vivo versus operant graded activity in chronic low back pain patients: Results of a randomized controlled trial. Pain. 2008;138(1):192-207.

120 | Chapter 7

18.

Rosenzweig S. Some implicit common factors in diverse methods of psychotherapy. Journal of Psychotherapy Integration. 2002;12(1):5-9.

19.

Lambert MJ, Ogles BM. Common factors: Post hoc explanation or empirically based therapy approach? Psychotherapy. 2014;51(4):500-504.

20.

Thorn BE, Burns JW. Common and speciﬁc treatment mechanisms in psychosocial pain interventions: The need for a new research agenda. Pain. 2011;152(4):705-706.

21.

Jensen MP. Psychosocial approaches to pain management: An organizational framework. Pain. 2011;152(4):717-725.

22.

Day MA, Jensen MP, Ehde DM, Thorn BE. Toward a theoretical model for mindfulness-based pain management. The Journal of Pain. 2014;15(7):691-703.

23.

Turk DC. Clinical effectiveness and cost-effectiveness of treatments for patients with chronic pain. Clin J Pain. 2002;18(6):355-365.

24.

Sletten CD, Kurklinsky S, Chinburapa V, Ghazi S. Economic analysis of a comprehensive pain rehabilitation program: A collaboration between ﬂorida blue and mayo clinic ﬂorida. Pain Med. 2015;16(5):898-904.

25.

Flor H, Fydrich T, Turk DC. Efﬁcacy of multidisciplinary pain treatment centers: A meta-analytic review. Pain. 1992;49(2):221-230.

General discussion | 121

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 124 | Summary

Pain rehabilitation programs (PRPs) consist of multidisciplinary treatments based on the biopsychosocial model. It can assist people with chronic musculoskeletal pain to manage their pain more effectively, and it can decrease pain related disabilities. Scientific research has shown that pain rehabilitation has a positive effect on disabilities, work participation and quality of life. However, there is a large amount of variation with regard to the content, form (inpatient or outpatient, individual or in a group), duration and intensity of treatments, and the number of disciplines involved in PRPs. In literature, very little is known about dosage of PRPs. Therefore, the aim of the studies in this thesis was to gain insight into the effect of the different aspects of dosage of PRPs. Various methods were used during the research and are described in the chapters of this thesis. In this thesis, dosage of PRP has been defined as the total duration of treatment in weeks, the total amount of contact hours, and the number of contact hours per week (intensity). In Chapter 2 a systematic review is described. During the review process, no studies were found that primarily dealt with the analysis of dosage of PRPs. Therefore, the aim of this review was to study randomized controlled trials that analyzed the effects of PRPs. A requirement was that these studies included dosage of treatments. The result is an overview of these studies, dosage of the treatments that were compared and the accompanying effects sizes regarding the domains: disabilities, work participation, and quality of life. Dosage of PRPs varied from 6.4 to 196.8 contact hours. In addition, the influence of dosage on the effect sizes was analyzed. A number of variables were found to influence effects. Because these variables were strongly related to dosage and it was not possible to isolate other factors, no robust conclusions could be drawn with regard to the influence of dosage on the effects.

Chapter 3 presents a prospective cohort study. Its objective was to analyze the course of disability reduction during PRP. The Pain Disability Index (PDI) was used as a measure of self-reported disability. Many studies are aimed at the effects of PRPs and use pre- and post-treatment outcome measures. With regard to dosage, however, the course of disability reduction during a treatment becomes important. The duration of a program could be shortened if the desired effect could be achieved before the end of the treatment. The analyses showed that the course of disability reduction occurred following a quadratic model. This means that disability reduction is greater in the beginning of the program and levels off as the treatment progresses. In addition, the measurement moment, initial PDI score, average pain score, and interaction between the initial PDI and measurement moment inﬂuenced the course of disability reduction. This means that disability reduction is faster for a patient with higher initial PDI and average pain score than for a patient with lower initial PDI and average pain score. The inﬂuence of the interaction between the measurement moment and the initial PDI makes the reduction in PDI per measurement moment become smaller at the end of the treatment. The results of this study can be used in clinical practice to predict the course of PDI, and to adjust dosage to the desired effect.

Summary | 125

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

The literature study, performed prior to the systematic review, resulted in no studies performed to analyze dosage of PRPs as a primary objective. In Chapter 4 a randomized controlled trial is described with regard to the use of different dosages of PRPs. The objective was to analyze the effect of PRP with a lower dosage compared to the current program at the Center of Rehabilitation UMCG. The treatment of the experimental group was 4 weeks shorter and had less contact hours than that of the control group. The content of the programs was similar. The trial is of a noninferiority design. This means that the effect on the experimental group must not be worse than the effect on the control group. The PDI was used as primary outcome measure. A non-inferiority margin of 4 points difference was applied. Disability reduction was observed in both groups. The difference between both groups was 2.5 points. This difference lies within the non-inferiority margin. However, the 95 percent confidence interval was -2.2 to 7.3. Because this interval is outside the margin between 0 and 4 points, non-inferiority cannot be established for certain. In the trial, treatments were often extended in the experimental group, making the contrast in dosage between both groups too small. Based on a priori stopping rules, the trial was stopped prematurely. Due to the lack of contrast in dosage between both groups and the wide margin with regard to the precision of the results, the conclusion that a shorter treatment is non-inferior cannot be drawn. Additionally, the difference in effects between the extended and non-extended treatments was analyzed. The analyses did not show a significant difference. This trial was the first to have the analysis of dosage as its primary goal. From the results, no robust conclusions can be drawn for now. However, the results do indicate possibilities to shorten PRPs and the need for further research into dosage of PRP. Besides quantitative data regarding dosage of PRPs, it is of added value to gather qualitative data. Chapter 5 presents a qualitative study with the aim of gaining insight into the perspectives of patients and professionals concerning dosage of PRPs. Patients who had finished a PRP and professionals of 3 rehabilitation centers in the Netherlands were interviewed, patients individually and professionals in focus groups. The treatments of the 3 rehabilitation centers differed in dosage. A total of 12 patients were interviewed, as well as a total of 17 professionals in 3 focus groups. Audio recordings were made of all interviews. These recordings were transcribed verbatim, and the data was analyzed and arranged thematically. In general, patients were satisfied with the dosage of treatment they had received. Although dosages differed, patients as well as professionals mentioned similar factors regarding dosage. All factors were categorized as patient-related, treatment-related, and external factors. Remarkably, the factors mentioned were the same to a large extent despite the difference in dosage. Furthermore, in the focus groups professionals reported that little is known about dosage in literature, and that is why, at the moment, the choice of dosage of PRP is based on clinical expertise and historical grounds.

126 | Summary

The objective of the study described in Chapter 6 was to explore case complexity of patients with chronic non-specific musculoskeletal pain (CMP). There could be a relationship between case complexity and the required dosage of PRP. However, the case complexity of this specific group was not transparent and difficult to measure, and that is why this relationship could not be analyzed. Because of this, a three-round Delphi study was performed among professionals involved in PRP in the Netherlands. These professionals were asked what they believed to be important factors that could influence the functioning of patients with CMP. In the first round, a total of 166 factors were identified. In the second round, the 10 most important factors were selected. In the third round, relative weights of each of these 10 factors were calculated. A scale of 0 (no weight) to 4 (very heavy weight) was used. The 10 most important factors were: psychiatric disorders, motivation, family system problems, mentalization capacity, somatization, financial and work related problems, features of complaints (duration, intensity), treatment-interfering personality traits, life events, and personal injury procedures. In daily clinical practice, establishing the influence of these factors on a patient’s disability is mainly based on clinical reasoning. The 10 factors were used to initiate the development of a Case Complexity Index (CCI). This index has a minimum score of 0 and a maximum score of 100, it shows the degree of case complexity, and is measured by establishing the presence of the 10 factors and the accompanying degree of influence on case complexity. Rehabilitation physicians and professionals tested the feasibility of the CCI on 14 patients. A reasonable similarity in CCI scores was found between the rehabilitation physicians and professionals. This study can help to initiate the assessment of case complexity. It may be used to gain transparency in case complexity, and to analyze the relationship between case complexity and dosage.

In the general discussion in Chapter 7, the main findings of this thesis are summarized. The factors related to dosage of PRPs are discussed, such as factors regarding the desired effect of treatment, patient- and PRP-related factors, and general and external factors. Clinical, scientific and societal implications are described, and divided using different perspectives. Because this thesis is a first step towards scientific research concerning dosage of PRP, recommendations are made for future research. The strengths and limitations of this thesis are described. The final conclusion in this chapter is that the results of this thesis offer a contribution towards: the research of PRP, making treatments transparent and comparable, and improving cost- effectiveness.

Summary | 127

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 130 | Samenvatting

Pijnrevalidatie kan mensen met chronische pijn aan bewegingsapparaat helpen om beter te leren omgaan met hun pijn en het kan de pijngerelateerde beperkingen verminderen. Deze revalidatie bestaat uit multidisciplinaire behandelingen gebaseerd op het biopsychosociale model. Wetenschappelijk onderzoek heeft aangetoond dat pijnrevalidatie een positief effect heeft op beperkingen, participatie en kwaliteit van leven bij mensen met chronische pijn. De praktijkvariatie in inhoud, vorm (klinisch - poliklinisch/ individueel - groepsbehandeling), duur en intensiteit van de behandeling en het aantal disciplines betrokken bij de pijnrevalidatie, is echter groot. Er is nog maar weinig bekend over de effecten van dosis van pijnrevalidatie op uitkomsten. Het doel van de studies in dit proefschrift was dan ook het analyseren van dosisaspecten van pijnrevalidatie en inzicht krijgen in de invloed van dosis op effecten van pijnrevalidatie. Dit is onderzocht middels verschillende onderzoeksmethoden, beschreven in de hoofdstukken in dit proefschrift. Onder dosis wordt in dit proefschrift verstaan: de totale duur van de behandeling in weken, het totale aantal contacturen en het aantal contacturen per week (intensiteit). In hoofdstuk 2 is een systematische review beschreven. De zoekstrategie in de wetenschappelijke literatuur naar studies die primair gericht waren op het analyseren van de dosis van pijnrevalidatie, leverde geen studies op. Daarom is gekozen om de review te richten op studies die de effecten van pijnrevalidatie hebben geanalyseerd middels een gerandomiseerde gecontroleerde studie. Een vereiste was dat de dosis van de onderzochte behandelingen was beschreven en het effect gemeten was voor minimaal 1 van de uitkomsten ervaren beperkingen, arbeidsparticipatie en kwaliteit van leven. Het resultaat is een overzicht van deze studies, de dosis van de behandelingen en de bijbehorende omvang van effect voor de genoemde uitkomsten. De dosis van de geĂŻncludeerde pijnrevalidatiebehandelingen varieerde in contacturen van 6,4 tot 196,8 uur. Aanvullend is de invloed van de dosis op het effect geanalyseerd. Een aantal variabelen bleek invloed te hebben op het effect. Maar die variabelen waren onderling ook sterk verbonden. Het was daarom niet mogelijk factoren te isoleren. Het was dan ook niet mogelijk om robuuste conclusies te trekken met betrekking tot de invloed van de dosis op uitkomsten. Hoofdstuk 3 is een prospectieve cohort studie met als doel het verloop van ervaren beperkingen, gemeten met de Pain Disability Index (PDI), gedurende pijnrevalidatie te analyseren. Veel studies onderzoeken het effect van pijnrevalidatie en meten uitkomstmaten vooraf en aan het eind van de behandeling. Echter in relatie tot dosis is het verloop van de uitkomst gedurende de behandeling interessant, omdat de behandeling verkort zou kunnen worden indien het effect al behaald is voor het einde van de behandeling. Uit de studie bleek dat de ervaren beperkingen afnamen volgens een kwadratisch tijd model: De mate van afname in ervaren beperkingen is in het begin van de behandeling groter en naarmate de behandeling vordert, vlakt de afname af. Daarbij bleek dat het meetmoment, de initiĂŤle score op de PDI, de gemiddeld pijnscore en de interactie tussen initiĂŤle PDI en meetmoment invloed hadden op het verloop van de ervaren beperkingen.

Samenvatting | 131

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Klinisch betekent deze uitkomst dat de afname van ervaren beperkingen sneller gaat voor een patiënt met een hogere initiële score op PDI en hogere pijnscore, dan voor een patiënt met een lagere initiële score op PDI en pijnscore. De invloed van de interactie tussen het meetmoment en de initiële PDI maakt dat de afname van PDI per meetmoment aan het einde van de behandeling kleiner wordt. De bevindingen uit deze studie kunnen gebruikt worden om in de klinische praktijk het verloop van de PDI te voorspellen en de dosis op het gewenste resultaat aan te passen. Uit de literatuurstudie voorafgaand aan de systematische review, bleek dat er geen studies waren, die primair gericht zijn op het vergelijken van pijnrevalidatie met verschillende doses. In hoofdstuk 4 is een gerandomiseerde gecontroleerde studie beschreven waarin pijnrevalidatie met verschillende doses is aangeboden. Het doel was het analyseren van het effect van een pijnrevalidatiebehandeling met een lagere dosis ten opzichte van de dosis van de huidige behandeling van het Centrum voor Revalidatie UMCG. De behandeling van de experimentele groep was 4 weken korter en bestond uit minder contacturen dan de controle groep. De inhoud bleef gelijkwaardig. De studie had een non-inferiority design. Dat betekent dat het effect van de experimentele groep niet slechter mocht zijn dan het effect van de controle groep. De PDI was gebruikt als primaire uitkomstmaat om het effect te meten, waarbij een marge van 4 punten verschil werd toegestaan als gelijkwaardig effect. Ervaren beperkingen namen af in beide groepen. Het verschil tussen beide groepen was gemiddeld 2,5 punten. Aangezien dit verschil binnen de marge van 4 punten ligt, wordt dit als gelijkwaardig geïnterpreteerd. De 95% betrouwbaarheidsintervallen lagen tussen -2.2 en 7.3. Aangezien de interval buiten de gestelde marge van 0 tot 4 punten ligt, is gelijkwaardigheid niet met zekerheid vast te stellen. In de praktijk werden de behandelingen in de experimentele groep vaak verlengd, waardoor het contrast in dosis tussen beide groepen te klein werd. De studie moest, vanwege vooraf opgestelde stopregels, voortijdig worden beëindigd. Het gebrek aan contrast in dosis tussen de groepen en de grote marge in betrouwbaarheid van de resultaten, betekent dat niet met zekerheid geconcludeerd kan worden dat het effect van een kortere behandeling gelijkwaardig is. Aanvullend zijn analyses gedaan om het verschil tussen effect van het al dan niet verlengen van behandelingen te analyseren. De effecten waren signiﬁcant verschillend. Deze studie was de eerste waarbij dosis als primair doel werd geanalyseerd. Uit deze resultaten kunnen nog geen robuuste conclusies getrokken worden. De resultaten geven echter wel een indicatie voor mogelijkheden voor het verkorten van pijnrevalidatiebehandelingen en daarmee voor de noodzaak voor verder onderzoek naar de dosis van pijnrevalidatie. Naast kwantitatieve gegevens omtrent de dosis van pijnrevalidatie, is het ook van toegevoegde waarde om kwalitatieve gegevens te verzamelen. Hoofdstuk 5 beschrijft een kwalitatieve studie met als doel het verkrijgen van inzicht in de perspectieven van de patiënten en professionals in relatie tot de dosis van pijnrevalidatie. Patiënten die een pijnrevalidatie behandeling hadden

132 | Samenvatting

afgerond en professionals werkzaam binnen de pijnrevalidatie in 3 revalidatiecentra in Nederland zijn geïnterviewd. Patiënten werden individueel geïnterviewd en professionals in een groep. De behandelingen van de 3 deelnemende revalidatiecentra verschilden in dosis. In totaal zijn 12 patiënten geïnterviewd en zijn 17 professionals in 3 groepen geïnterviewd. Van alle interviews waren geluidsopnamen gemaakt. Deze opnamen zijn daarna letterlijk uitgeschreven en de data zijn geanalyseerd en thematisch geordend. Over het algemeen waren patiënten tevreden met de verkregen dosis. Ondanks de verschillen in dosis, werden zowel door de patiënten als door de professionals soortgelijke factoren benoemd in relatie tot de dosis. Alle factoren zijn gecategoriseerd in factoren gerelateerd aan de patiënt, aan de behandeling en aan externe factoren. Opmerkelijk was, dat ondanks verschillen in dosis, de genoemde factoren grotendeels hetzelfde waren tussen centra. Verder werd in de focusgroepen met de professionals aangeven, dat er weinig bekend is over de dosis in de literatuur, waardoor de keuze van de dosis van pijnrevalidatie op dit moment gebaseerd is op klinische ervaringen en historische gronden. De studie beschreven in hoofdstuk 6, is opgezet om de complexiteit van de casus van de patiënten met chronische pijnklachten die in aanmerking komen voor pijnrevalidatie, te bestuderen. Er zou een relatie kunnen bestaan tussen de complexiteit van de casus en de benodigde dosis van pijnrevalidatie. “Complexiteit” van de casus bleek echter voor deze speciﬁeke doelgroep niet transparant en moeilijk meetbaar, waardoor de genoemde relatie niet geanalyseerd kon worden. Vanuit de klinische praktijk leek de complexiteit te bestaan uit invloed van verschillende factoren. In de studie namen professionals deel die werkzaam zijn binnen de pijnrevalidatie in heel Nederland. In 3 vragenrondes werd gevraagd welke factoren een relevante invloed kunnen hebben op het functioneren van patiënten met chronische pijnklachten. In de eerste ronde zijn 166 factoren benoemd. In de tweede ronde is hieruit een top 10 samengesteld. In de derde ronde zijn voor de factoren uit de top 10 de mate van invloed aangegeven op het functioneren van patiënten met chronische pijnklachten op een schaal van 0 tot 4 (0 geen invloed en 4 erg sterk van invloed). De 10 belangrijkste factoren waren: psychiatrische problematiek, lage motivatie, systeemproblematiek, gering mentaliserend vermogen, somatisatie, ﬁnanciële en werkgerelateerde problemen, klachtkenmerken (duur, intensiteit), persoonlijkheidskenmerken die de behandeling belemmeren, life events en letselschadeprocedure. Het vaststellen van (de mate van invloed van) deze factoren bleek in de huidige praktijk vooral gebaseerd op klinisch redeneren; niet op basis van uitkomsten van meetinstrumenten. De 10 factoren zijn gebruikt om een casus complexiteitsindex (CCI) op te stellen. Deze index, geeft de mate van complexiteit weer. De CCI wordt berekend op basis van de aanwezigheid van de 10 factoren en de bijbehorende mate van invloed op complexiteit. De score kan variëren van 0 tot 100, waarbij een hoge score duidt op een hoge casus complexiteit. De score kan bestaan uit aanwezigheid van meerdere factoren of een paar factoren met een hoge mate van invloed. Het gebruik van de CCI is getest in de huidige praktijk door revalidatieartsen en het behandelteam bij 14 patiënten. Er bleek een

Samenvatting | 133

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

redelijke overeenstemming van score van CCI tussen de revalidatieartsen en het behandelteam. Deze studie heeft een eerste aanzet gedaan naar het meetbaar maken van de complexiteit van de casus. De resultaten van deze studie kunnen mogelijk gebruikt worden om transparantie te krijgen in casuscomplexiteit en om de relatie tussen complexiteit en dosis te analyseren. In de algemene discussie in hoofdstuk 7 zijn de voornaamste resultaten van het proefschrift beschreven. De verschillende factoren die gerelateerd zijn aan de dosis van PRP, zoals factoren met betrekking tot het beoogde behandeleffect, patiĂŤnt en PRP gerelateerde factoren, algemene en externe factoren, worden benoemd en bediscussieerd. De klinische, wetenschappelijke en maatschappelijke implicaties worden vanuit verschillende perspectieven beschreven. Omdat dit proefschrift een eerste stap is geweest om de dosis van pijnrevalidatie wetenschappelijk te onderzoeken, zijn aanbevelingen gedaan voor vervolgonderzoek. De sterke en minder sterke punten van het proefschrift zijn beschreven. Het hoofdstuk wordt afgesloten met de conclusie dat de resultaten van dit proefschrift kunnen bijdragen aan onderzoek binnen de pijnrevalidatie, het transparant en vergelijkbaar maken van keuzes ten aanzien van de dosis en op termijn mogelijk tot het verbeteren van de kosteneffectiviteit.

134 | Samenvatting

Samenvatting | 135

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 138 | Dankwoord

In september 2009 ben ik gestart met mijn promotietraject. Een enorme uitdaging; dat was het zeker!!! Gelukkig heb ik vaak steun ervaren van de mensen om mij heen, met wie ik vele goede, en zeker ook de minder goede momenten heb mogen delen. Via deze weg wil ik iedereen bedanken en een aantal mensen in het bijzonder: Allereerst alle professionals en patiënten die deel hebben genomen aan de verschillende onderzoeken. Zonder deelnemers, geen onderzoek. Dank voor ieders inzet en bijdrage. Dank aan mijn promotieteam: prof. dr. JHB Geertzen, prof. dr. MF Reneman en prof. dr. PU Dijkstra. Voor mij was dit promotieteam de perfecte combinatie. Balans tussen actie en rust; inhoud en praktijk en de zuidelijke gezelligheid en de noordelijke stugheid. Ik dank jullie voor de prettige samenwerking en alles wat ik van jullie heb geleerd. Beste Jan, jij gaf mij de kans om als ergotherapeut te promoveren. De eerste jaren was jij mijn eerste promotor. Zodra Michiel benoemd werd tot professor, werd hij eerste promotor. Stiekem ben je, naast Michiel, ook altijd een beetje mijn eerste promotor gebleven! Dank voor al de kritische noten, de waardevolle aanvullingen en natuurlijk de Brabantse touch! Beste Michiel, ik weet nog dat je kort na de promotie van Remko tegen mij zei: “Wij hadden bedacht dat jij over een aantal jaren ook wel in die aula kon gaan staan”. Het duurde bij mij even voordat ik in de gaten had waar je het over had. Dank dat je mij deze kans hebt gegeven en dank voor de vele steun, het inkorten en afkappen van mijn ellenlange zinnen en discussies. Maar bovenal voor het vertrouwen dat je me gegeven hebt. Als ik even door de bomen het bos niet meer zag, kon je me in laten zien hoe mooi het project was en wat het onderzoek allemaal opleverde! Daardoor raakte ik, ondanks de obstakels die ik moest nemen, toch weer gemotiveerd om positief verder te gaan. Beste Pieter, jij kwam iets later bij de club. Wat ben ik vaak blij geweest met jouw expertise. Al die data en al die statistische formules waar ik niets van snapte en wat jij met alle geduld uitlegde. Samen zochten we naar de vertaalslag van de resultaten naar de betekenis voor klinische praktijk. Dank voor je geduld, je wijze raad en de goede en gezellige gesprekken.

Het nemen van de stap om als ergotherapeut wetenschappelijk onderzoek te gaan doen en uiteindelijk te gaan promoveren, heb ik voornamelijk te danken aan Nelleke Kombrink, Rita Schiphorst Preuper en Michiel Reneman. Zij hebben hier een belangrijke bijdrage aan geleverd. Nelleke, jij bent degene die mij heeft aangespoord om de wetenschappelijke kant op te gaan. Jij was teammanager van het pijnteam en tijdens mijn jaargesprek kwamen de onderwerpen uitdaging en wetenschap bij elkaar. Je stimuleerde me om mijzelf te oriënteren op de mogelijkheden een wetenschappelijke studie te gaan doen. Ik ben je dankbaar voor het advies wat je me toen gegeven hebt met dit proefschrift als resultaat!

Dankwoord | 139

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Rita, jij stimuleerde dit ook. Je dacht mee over de mogelijkheden om te kunnen promoveren en adviseerde mij tijdens het hele promotietraject. Jouw klinische en wetenschappelijke blik en de daaruit voortkomende zinvolle aanvullingen en ideeën hebben de uitvoer van het hele project goed gedaan! Dank hiervoor. Annemieke de Jong en Berry van Holland, wat geweldig dat jullie mijn paranimfen willen zijn! Annemieke: als collega’s in het pijnteam, als onderzoekers en zeker niet op de laatste plaats ook als dames die graag samen een hapje gaan eten om bij te kletsen, hebben we vele gesprekken gehad en heb ik veel aan je gehad. Dank voor alles: voor de gesprekken, adviezen en gezellige eetafspraken! Berry, je bent de beste kamergenoot die iemand zich kan wensen! Of het nu ging om de goede Engelse bewoording, de lastige formules in Excel, het vieren van de successen of het uiten van de frustraties, je had altijd een luisterend oor en was altijd behulpzaam! Heel, heel erg bedankt daarvoor! Alle collega’s en ex-collega’s van het pijnteam wil ik bedanken voor alle steun, acties, bijdragen, kritische blikken, discussies, interesses en ieders luisterend oor. De revalidatieartsen Maikel, Jitze, Rita en Wim: speciale dank voor het informeren en includeren van alle patiënten voor de RCT en ieders bijdrage aan het onderzoek en dit proefschrift. Alle ergotherapeuten, fysiotherapeuten, psychologen en maatschappelijk werkster bedankt! Zonder jullie inzet was het niet gelukt en zonder jullie steun was het mij nooit gelukt om dit tot een goed einde te brengen. Speciale dank voor mijn directe collega’s van de ergotherapie: Anne, Judith en Lucienne. Ik kon altijd bij jullie terecht, dat waardeer ik enorm. De secretaresses van het pijnteam en in het bijzonder Marleen Speller: dank voor alle hulp die jullie geboden hebben gedurende het hele traject en met name tijdens de uitvoer van de RCT. Wat kunnen jullie veel werk verzetten en ook nog overzicht houden! Dank hiervoor. Alle coauteurs wil ik bedanken voor hun waardevolle bijdrage aan het tot stand komen van de verschillende artikelen in dit proefschrift. Tijdens een promotietraject kruisen vele praktische en emotionele zaken je pad. Wat was ik dan ook altijd blij met de collega’s die dezelfde ervaringen hadden op het gebied van onderzoek doen en promoveren. De collega-onderzoekers van het pijnteam: Remko, Haitze, Berry en de collega’s van Oker: Sacha, Ilse, Feyuna, Christa, Ria, Anja en Ant. Dank voor alle adviezen en het delen van jullie (persoonlijke) ervaringen en expertise.

140 | Dankwoord

Tenslotte wil ik mijn lieve familie en vrienden bedanken voor alle steun en belangstelling. Er was altijd tijd en belangstelling voor mijn project. Lieve pap, mam en Ellen, dank dat jullie er altijd voor me zijn! Lieve Gino, jij hebt altijd geloofd en gezegd dat ik dit kon, ondanks dat ik daar zelf nog niet zo overtuigd van was. En je had gelijk: het is gelukt! Thanx laiverd!

Dankwoord | 141

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 144 | Over de auteur

Franka Waterschoot, geboren op 20 juni 1978 te Hoogeloon, groeide op in de buurt van Eindhoven. Na de middelbare school is ze de opleiding tot ergotherapeut gaan volgen in Heerlen, waar ze in 2002 afstudeerde. Nadien verhuisde ze van Noord Brabant naar Groningen. Het eerste half jaar heeft ze gewerkt als ergotherapeut in verpleegtehuis Dilgtoord in Haren en de daarbij horende locatie de Burcht in Hoogezand. Vanaf oktober 2002 is zij gaan werken bij het UMCG, Centrum voor Revalidatie (CvR), locatie Beatrixoord Haren. Binnen Beatrixoord heeft ze als ergotherapeut met name gewerkt binnen de pijnrevalidatie, maar ook binnen de diabetesen neurorevalidatie. Na een aantal jaren was Franka op zoek naar een nieuwe uitdaging. Deze uitdaging startte in september 2009 met een parttime promotietraject met dit proefschrift als eindresultaat. Sinds oktober 2014 heeft Franka de shift gemaakt van de patiĂŤntenzorg naar de kwaliteit van de zorg en zet ze haar kennis en expertise in voor projecten binnen het CvR. Momenteel is ze projectleider van de implementatie van de landelijke behandelmodules binnen het CvR en van de werkgroep die zich bezig houdt met de zorglogistiek ten behoeve van een gastvrij en patiĂŤntgericht CvR. Daarnaast is ze coĂśrdinator van innovaties in relatie tot ehealth en zorglogistieke processen binnen het pijnteam.

Over de auteur | 145

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 148 | Research Institute SHARE

Research Institute SHARE This thesis is published within the Research Institute SHARE (Science in Healthy Ageing and healthcaRE) of the University Medical Center Groningen / University of Groningen. Further information regarding the institute and its research can be obtained from our internetsite: http://www.share.umcg.nl/ More recent theses can be found in the list below. ((co-) supervisors are between brackets) Janse M The art of adjustment (prof AV Ranchor, prof MAG Sprangers, dr J Fleer) Nigatu YT Obesity and depression; an intertwined public health challenge (prof U B端ltmann, prof SA Reijneveld) Aris-Meijer JL Stormy clouds in seventh heaven; a study on anxiety and depression around childbirth (prof CLH Bockting, prof RP Stolk, dr H Burger) Kluitenberg B The NLstart2run study; running related injuries in novice runners (prof RL Diercks, dr H van der Worp, dr M van Middelkoop) Benjaminse A Motor learning in ACL injury prevention (Prof E Otten, prof KAPM Lemmink, prof RL Diecks) Potijk MR Moderate prematurity, socioeconomic status, and neurodevelopment in early childhood; a life course perspective (prof SA Reijneveld, prof AF Bos, dr AF de Winter)

Research Institute SHARE | 149

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Beyer AR The human dimension in the assessment of medicines; perception, preferences, and decision making in the European regulatory environment (prof JL Hillege, prof PA de Graeff, prof B Fasolo) Alferink M Psychological factors related to Buruli ulcer and tuberculosis in Sub-saharan Africa (prof AV Ranchor, prof TS van der Werf, dr Y Stienstra) Bennik EC Every dark cloud has a colored lining; the relation between positive and negative affect and reactivity to positive and negative events (prof AJ Oldehinkel, prof J Ormel, dr E Nederhof, dr JACJ Bastiaansen) Stallinga HA Human functioning in health care; application of the International ClassiďŹ cation of Functioning, Disability and Health (ICF) (prof RF Roodbol, prof PU Dijkstra, dr G Jansen) Papageorgiou A How is depression valued? (prof AV Ranchor, prof E Buskens, dr KM Vermeulen, dr MJ Schroevers) Spijkers W Parenting and child psychosocial problems; effectiveness of parenting support in preventive child healthcare (prof SA Reijneveld, dr DEMC Jansen) Ark M van Patellar tendinopathy; physical therapy and injection treatments (prof RL Diercks, prof JL Cook) Boven JFM van Enhancing adherence in patients with COPD: targets, interventions and costeffectiveness (prof MJ Postma, prof T van der Molen)

150 | Research Institute SHARE

Gerbers JG Computer assisted surgery in orthopedaedic oncology; indications, applications and surgical workďŹ&#x201A;ow (prof SK Bulstra, dr PC Jutte, dr M Stevens) Niet A van der Physical activity and cognition in children (prof C Visscher, dr E Hartman, dr J Smith)

For more 2015 and earlier theses visit our website

Research Institute SHARE | 151

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 154 | EXPAND

Wetenschappelijk onderzoek afdeling Revalidatiegeneeskunde – Centrum voor Revalidatie UMCG EXPAND Extremities, Pain and Disability Missie: EXPAND draagt bij aan participatie en kwaliteit van leven van mensen met aandoeningen en amputaties van de extremiteiten of met pijn aan het bewegingsapparaat. EXPAND omvat twee speerpunten: • onderzoek naar aandoeningen aan en amputaties van extremiteiten met nadruk op stoornissen, activiteiten en participatie, • onderzoek naar chronische pijn en arbeidsparticipatie. EXPAND draagt bij aan het UMCG-brede thema Healthy Ageing.

Research Department of Rehabilitation Medicine – Center for Rehabilitation UMCG EXPAND Extremities, Pain and Disability Mission: EXPAND contributes to participation and quality of life of people with health conditions and amputations of the extremities and musculoskeletal pain. EXPAND includes two spearheads: • research on the health conditions and amputations of the extremities with emphasis on body functions and structures, activities and participations, • chronic pain and work participation. EXPAND contributes to Healthy Aging, the focus of the UMCG.

E EXPAND | 155

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39