work-neck pain by Qualified Physio

PAIN 151 (2010) 162–173

www.elsevier.com/locate/pain

Work and neck pain: A prospective study of psychological, social, and mechanical risk factors Jan Olav Christensen a,b,*,1, Stein Knardahl a,1 a b

The National Institute of Occupational Health, Norway Department of Psychology, University of Oslo, Norway

a r t i c l e

i n f o

Article history: Received 19 November 2009 Received in revised form 20 May 2010 Accepted 1 July 2010

Keywords: Neck pain Occupational Psychosocial Prospective Mechanical Exposure

a b s t r a c t To determine the impact of occupational psychological/social and mechanical factors on neck pain, a prospective cohort study with a follow-up period of 2 years was conducted with a sample of Norwegian employees. The following designs were tested: (i) cross-sectional analyses at baseline (n = 4569) and follow-up (n = 4122), (ii) prospective analyses with baseline predictors, (iii) prospective analyses with average exposure over time [(T1 + T2)/2] as predictor, and (iv) prospective analyses with measures of change in exposure from T1 to T2 as predictors. A total of 2419 employees responded to both the baseline and follow-up questionnaire. Data were analyzed using ordinal logistic regression. After adjustment for age, sex, neck pain at T1, and other exposure factors that had been estimated to be confounders, the most consistent risk factors were role conﬂict (highest OR 2.97, 99% CI: 1.29–6.74) and working with arms raised to or above shoulder level (highest OR 1.37, 99% CI: 1.05–1.78). The most consistent protective factors were empowering leadership (lowest OR 0.53, 99% CI: 0.35–0.81) and decision control (lowest OR 0.60, 99% CI: 0.36–1.00). Hence, psychological and social factors are important precursors of neck pain, along with mechanical factors. Although traditional factors such as quantitative demands and decision control play a part in the etiology of neck pain at work, in this study several new factors emerged as more important. Ó 2010 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

1. Introduction Neck pain is a widespread health problem in the working population. The annual prevalence has been reported to be above 40% [1,3,24]. Neck pain may severely disrupt the well being and productivity of the individual. Hence, it is a major public health issue, even if the medical consequences for each individual are not severe. Traditionally, biomechanical loads have been assigned the main role in explaining relationships between work and musculoskeletal disorders (MSDs) [6]. However, MSDs are common in ofﬁce workers and remain as prevalent conditions even if biomechanical loads have been reduced in most occupations. Hence, psychological factors are acknowledged as contributors to the pathogenesis of MSD. Nevertheless, the pathogenetic mechanisms of neck pain remain obscure. Several hypotheses have maintained that sustained muscle contraction with metabolic demands in muscle ﬁbers exceeding the energy supply leads to energy crisis which could produce pain [19,25]. What actually activates the nociceptive neurons is still un-

* Corresponding author. Address: The National Institute of Occupational Health, P.O. Box 8149 Dep, N-0033 Oslo, Norway. Tel.: +47 23 19 52 58; fax: +47 23 19 52 00. E-mail address: joc@stami.no (J.O. Christensen). 1 Tel.: +47 23 19 51 00; fax:+47 23 19 52 00.

known, and several studies have reported very low levels of trapezius muscle activity during office work (e.g. [45]). During psychological challenge, active coping behavior includes a cardiovascular response pattern in which blood flow to muscles increases [45]. It has been hypothesized that factors involved in the regulation of muscle blood vessels may activate nociceptors [30,31]. However, it is also possible that the pain may originate from tendons or periostea or other components outside the muscle tissue. Systematic reviews have concluded that several occupational psychological and social factors have been found to be related to neck and upper limb pain, such as lack of supervisory support [2,38,49], low social support [2,7,47], low job control [6,7,47,49], high job strain, conflicts at work, low job security [2], and, most consistently, job demands [2,7,38,47,49]. However, consistency of findings across studies is low, and several methodological shortcomings have been pointed out [6,7,16,22,32]. Perhaps most severe are ill-defined conceptualizations of psychological and social factors [37] and a paucity of studies with longitudinal prospective design. In their critical systematic review, Hartvigsen et al. [21] found that evidence of associations between psychosocial factors at work and low back pain was lacking. Variation between studies, they claimed, indicated spurious findings due to widespread use of non-standardized, non-validated instruments for measuring psychosocial factors. Furthermore, they suggested that this tendency was reinforced by multiple testing within and across studies. Davis

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

and Heaney [11] reported that only one third of the studies in their review employed well-validated multiple-item measures. Also, psychological demands may be confounded by physical demands, yet there has been few studies that take both of these factors into account [3,37]. Almost all previous studies have measured only a small selection of psychological and social work factors [32]. The objective of the present study was to determine which occupational psychological and social factors contribute to risk of neck pain, taking into account several factors not previously reported as well as aspects of biomechanical exposure. With a prospective study design we aimed to take into account changes in exposure over time as well as temporal associations between exposures and subsequent neck pain.

2. Methods 2.1. Design The study design was longitudinal and prospective, with all variables measured at baseline and at follow-up 2 years later. Since the question of the mechanisms of work-related musculoskeletal symptoms is unresolved, the required time period before exposure produces a health effect (neck pain) is unknown [15]. Therefore, several statistical designs testing associations over time were performed. In order to evaluate variance in patterns of association between psychological and social factors and neck pain with different time periods, data were analyzed cross-sectionally at both time points as well as prospectively. The time period was considered the same for all respondents within each company, since the electronic forms were closed for all respondents in each company simultaneously, with the opportunity of responding until that point in time. The average time period from the end of baseline to the end of follow-up for respondents was 23 months (range 17–36). 2.2. Subjects Subjects were recruited from organizations in Norway that were contacted and offered to participate in the study. Baseline data were gathered from November 2004 until November 2006, and follow-up data were gathered from September 2006 until February 2009. In return for participating and making the data available for research, companies received written reports and oral presentations of results as a tool for organizational development and an aid for monitoring the organizational work environment. The survey was web-based, although the participants with limited access to computers at work were given the option of filling out a paper version. Employees and management in the companies were informed at the organizational level first. Subsequently, all employees excluding those on sick leave were mailed a letter with information about the survey. This letter contained either a personalized code for logging into the web-questionnaire or a paper version of the questionnaire with a pre-stamped return envelope, in addition to information about the survey. The written information explained the aims of the study and assured that responses would be treated confidentially, in strict accordance with the general guidelines and specific license from the Norwegian Data Inspectorate. Employees were given the opportunity of filling out the questionnaire at work, but it was also possible to fill it out from home or any other location. Each subject had the opportunity to log into the web-questionnaire an unlimited number of times to change or complete their answers during the survey period. The questionnaire gathered data about background information, work organization, psychological and social factors at work,

163

organizational change, attitudes towards work, personality, physical activity, smoking, alcohol use, mental health, mastery of work, working ability, and health complaints. This study is based on parts of this information. In addition, all companies except one supplied lists containing information about address, age, sex, department, and classification of occupation for all employees prior to the survey. Occupation was classified according to the standard classification of occupations (STYRK), developed by Statistics Norway (www.ssb.no) based on the International Standard Classification of Occupation (ISCO-88). By the time of data analysis, 8928 employees had been invited to participate at baseline, of which 4569 responded (51.2%), forming cross-sectional sample 1. A responder was defined as a person who had completed the neck pain item and at least one predictor. At follow-up, a total of 8836 employees were invited, and 4122 responded (46.7%), forming cross-sectional sample 2. Paper-andpencil versions were returned by 1082 participants (12.1% of possible respondents, 23.7% of actual respondents) at baseline, and 388 at follow-up (4.4% of possible, and 9.4% of actual respondents at follow-up, including those not participating at baseline). Respondents were recruited from 20 organizations, 10 private (2703 invited and 1060 responded at baseline, 2434 invited and 791 responded at follow-up) and 10 public (6225 invited and 3509 responded at baseline, 6402 invited and 3331 responded at follow-up), with a range of 13–2316 invited employees at baseline and 14–2036 at follow-up. The participating organizations represented a wide variety of job types, with four municipalities (1315, 1060, 1113, and 434 invited at baseline), one insurance company (2317 invited at baseline), and two public organizations (297 and 1754 invited at baseline) comprising 8290 of 8928 (92.9%) invited employees at baseline. Eight companies were unable to participate twice, and thus were not included in this study. Of the 8928 employees that were invited to participate at baseline, 6361 (71.2%) were also invited at follow-up. Of these, 3749 (58.9%) responded at baseline, 3093 (48.6%) responded at followup, and 2419 (38.0%) responded at both times. Responding at both time points was defined as completing neck pain items at both baseline and follow-up and at least one predictor at baseline. 2.3. Outcome measure The outcome measure was the reported intensity of neck pain during 4 weeks prior to answering the questionnaire (see [44]). This item was part of a questionnaire measuring 21 different health complaints. Health complaints were measured by asking ‘‘have you been bothered by .... (i.e. neck pain) the last 4 weeks”, with optional answers ‘‘not bothered” (1), ‘‘a little bothered” (2), ‘‘rather intensely bothered” (3), and ‘‘very intensely bothered” (4). If the subject answered ‘‘a little bothered”, ‘‘rather intensely bothered”, or ‘‘very intensely bothered”, the question was followed by items reflecting the duration of the health complaint, whether it had been experienced at work, and whether it was believed to be caused by work. 2.4. Psychological and social work factors The psychological and social factors were assessed by the General Nordic Questionnaire for Psychological and Social Factors at Work (QPSNordic) [10]. The following scales were studied (a coefficients are calculated from the total baseline and follow-up samples, respectively): quantitative demands (four items, a = 0.76 and 0.76), decision demands (three items, a = 0.63 and 0.64), decision control (five items, a = 0.69 and 0.69), control over work intensity (four items, a = 0.80 and 0.82), role conflict (three items, a = 0.68 and 0.68), role clarity (three items, a = 0.82 and 0.82), support from immediate superior (three items, a = 0.85 and 0.86),

164

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

empowering leadership (three items, a = 0.88 and 0.88), fair leadership (three items, a = 0.81 and 0.81), predictability during the next month (three items, a = 0.65 and 0.64), commitment to organization (three items, a = 0.87 and 0.86), social climate (three items, a = 0.72 and 0.71), and positive challenge at work (three items, a = 0.75 and 0.75). The scales were constructed on the basis of the following frequency scoring: ‘‘1 = very seldom or never”, ‘‘2 = somewhat seldom”, ‘‘3 = sometimes”, ‘‘4 = somewhat often”, and ‘‘5 = very often or always”. All relationships were not necessarily expected to be linear or monotonic (see [13,14]). Therefore, scale scores were categorized so that the effect of different levels of exposure could be investigated separately. Scores between 1.00 and 1.80 were assigned the value ‘‘1”, scores between 1.81 and 2.60 the value ‘‘2”, scores between 2.61 and 3.40 the value ‘‘3”, scores between 3.41 and 4.20 the value ‘‘4”, and scores between 4.21 and 5.00 the value ‘‘5”. In order to estimate the effect of exposure over time on subsequent neck pain, the baseline and follow-up measures of exposure were averaged. The resulting variables measure exposure on a scale of 1–5. For regression analyses, the scale was categorized into five categories according to the procedure described in the previous paragraph. Due to a low number of responses in some categories, ‘‘1” and ‘‘2” were combined for role clarity, social climate and positive challenge in analyses of cross-sectional samples, and for decision demands, role clarity, predictability during the next month, social climate, and positive challenge in analyses of the prospective sample. To investigate the effect of increasing and decreasing exposure or different levels of unchanged exposure across time, analyses were also performed with change in exposure from baseline to follow-up as predictor. Hence, the categorized measures were trichotomized to reflect ‘‘low” (1 and 2), ‘‘middle” (3), and ‘‘high” (4 and 5) levels. Any change across these categories from baseline to follow-up was considered a change. Possibly, this operationalization leaves too little contrast between categories, i.e. not enough change to detect effects. Also, changes in exposure may interact with baseline levels of exposure. For instance, increases from high levels could be quite different from increases from low levels. However, more specific classifications were not studied because of statistical power issues: for the different psychological/social exposures an average of only 41 employees changed from ‘‘low” to ‘‘high” and 40 employees changed from ‘‘high” to ‘‘low”. It is important to keep in mind that classifying ‘‘change” according to direction alone may have affected effect estimates. The resulting categories of change were ‘‘constant low”, ‘‘decrease”, ‘‘constant middle”, ‘‘increase”, and ‘‘constant high”. Due to few responses in the ‘‘constant low” category when multivariate regressions were run, ‘‘constant low” and ‘‘constant middle” were combined for ‘‘role clarity”, ‘‘predictability during the next month”, ‘‘social climate”, and ‘‘positive challenge”. 2.5. Mechanical exposure factors Perceived physical workload was measured with three items under the heading ‘‘to what degree do your work tasks consist of. . .”. The items were ‘‘lifting or handling objects that weigh approximately 1–5 kg with your own muscular strength”, ‘‘lifting or handling objects that weigh approximately 6–15 kg with your own muscular strength”, ‘‘lifting or handling objects that weigh more than approximately 15 kg with your own muscular strength”. The response categories for these questions were ‘‘1 = seldom or never”, ‘‘2 = sometimes”, ‘‘3 = daily”, and ‘‘4 = many times pr. day”. These items were combined into one scale (a = 0.90 at both total baseline sample and total follow-up sample). This scale was weakly negatively correlated with the scale quantitative demands

(Spearman’s q = 0.13, p = 0.00 and q = 0.11, p = 0.00, in the baseline and follow-up samples, respectively), thereby displaying discriminant validity. For the regression analyses, the variable was recoded: ‘‘1.00–1.50” set to ‘‘1”, ‘‘1.51–2.50” set to ‘‘2”, ‘‘2.51–3.50” set to ‘‘3”, and ‘‘3.51–4.00” set to ‘‘4”. Working with arms raised to or above shoulder level was measured with the item ‘‘work positions where you have to raise your arms to or above shoulder level”, using the same response categories as the physical work load items. The average exposure across time points was also estimated for physical workload and working with arms raised to or above shoulder level. For regression analyses, the variables were recoded: ‘‘1.00–1.50” was set to ‘‘1”, ‘‘1.51–2.50” was set to ‘‘2”, ‘‘2.51–3.50” was set to ‘‘3”, and ‘‘3.51–4.00” was set to ‘‘4”. The change in physical workload was estimated by dichotomizing the categorized factor into ‘‘low” (1) and ‘‘high” (2–4), and then combining baseline and follow-up measures to form the categories ‘‘constant low”, ‘‘decrease”, ‘‘increase”, and ‘‘constant high”. The change in working with arms raised to or above shoulder level was estimated the same way with the single item at baseline and follow-up. 2.6. Statistical analyses All statistical analyses were conducted with PASW Statistics (formerly SPSS), version 17.0. 2.6.1. Non-response analysis In order to estimate the potential confounding effects of non-response, a multivariate binary logistic regression model was computed to calculate the odds of being a responder at baseline based on sex and age group. All invited employees at baseline were included in this analysis. 2.6.2. Attrition analysis To assess the potential presence of attrition bias, univariate binary logistic regression analyses were conducted to identify psychological, social, and mechanical factors associated with responding at both time points as opposed to responding only at baseline. All subjects that had been invited at both baseline and follow-up and had responded at baseline were included in these analyses. Factors found to predict response at follow-up were then entered into a multivariate binary logistic regression model to compute the odds of being a responder at follow-up controlling for the other related factors as well as age, sex, and neck pain intensity at baseline. 2.6.3. Main analyses Because of the ordinal nature and non-normality of the outcome scores, ordinal logistic regressions, by Generalized Linear Models (GZLMs) in PASW, were conducted to predict the outcome from psychological, social, and mechanical factors. Cumulative odds ratios were calculated for each category of each predictor. Cumulative odds ratios compare the odds of outcome being above vs. below each of the possible cutpoints (i.e. pain intensity of 2–4 vs. 1, 3–4 vs. 1–2, and 4 vs. 1–3) for all levels of the predictor variable with the odds for the reference category [29]. An assumption of proportional odds across all categories of the outcome is necessary for this model to be valid and this assumption was tested. Age, sex, and baseline level of neck pain were selected a priori for inclusion in all multivariate analyses. Educational level was also measured, but was omitted from multivariate analyses due to a high rate of missing data (39% among baseline responders). However, the confounding inﬂuence of educational level was estimated according to the procedure described below, and such confounding was indicated only for the change measure of physical workload.

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Categorizing predictors helps detect nonlinear or threshold effects, but may decrease statistical power [51]. Therefore, analyses were also conducted with the original scale scores of the psychological and social factors entered as continuous variables. Odds ratios for continuous predictors are interpreted as the change in odds per unit change of the predictor. Multiple testing was necessary to investigate a comprehensive set of factors. Since this inflates the risk of type I error, a strict approach was adopted by using 99% confidence intervals in main analyses. In addition a Bonferroni-corrected significance level was estimated based on the number of factors tested. This was done by dividing the overall significance level by the number of factors tested (i.e. 0.01/15 = 0.0007). Exposures were modeled separately, not adjusted for all other independent variables. While entering all factors into one model would account for the maximum amount of variance and adjust for confounding, it would partial out all explained variance that is shared between predictors. Confounding is not the only possible reason for shared variance, and it is statistically indistinguishable from mediation [4]. Unless the aim is to identify the strongest predictor or establish evidence of mediation, entering a mediator in a multivariate model constitutes overadjustment since it obscures the impact of the preceding factor [9,28]. Hence, it may motivate the incorrect conclusion that the antecedent factor is theoretically irrelevant. One contributing reason for conflicting results or lack of results in epidemiologic research is the adjustment for mediators [9,12]. The aim of the present study was to investigate a comprehensive range of factors in order to identify predictors of neck pain. These factors correlate to some degree and may be causally related [10]. Thus, in order to limit potential overadjustment each predictor was adjusted only for other predictors that were first estimated to exert an influence above a certain threshold level, according to the strategy recommended by Rothman [43]. These assessments were based on the baseline sample data. Univariate regressions were first conducted for each exposure variable. Then each variable was included in bivariate regressions with all other exposure variables. Subsequently, the crude odds ratios from the univariate regressions were compared with the adjusted odds ratios from the bivariate regressions. When the inclusion of a potential confounder resulted in a change in odds ratio of 10% or more that variable was treated as a real confounder. These confounders were included in all subsequent multivariate analyses with the relevant exposure. For the exposure change variables the confounder estimation procedures were repeated with change measures as potential confounders. 2.6.4.1. Potentially etiologic factors – onset of pain. Factors that influence the onset of neck pain (etiologic factors) may differ from factors that influence the course of pre-existing pain (prognostic factors). Musculoskeletal pain exhibits substantial variation over time [34], hence a single report of no pain does not establish a status of no previous pain. However, in order to elucidate factors that may determine onset of neck pain, a separate analysis was performed on the 1915 subjects who reported of no pain (1 = ‘‘not bothered”) at baseline in the prospective sample. Multivariate binary logistic regressions (with continuous baseline predictors) were run to estimate the risk of neck pain at follow-up. 2.6.4.2. Potentially prognostic factors in subjects with pain. Multivariate binary logistic regressions (with continuous baseline predictors) were performed on the 1838 subjects who reported (2 = ‘‘a little bothered”/‘‘rather intensely bothered”/‘‘very intensely bothered”) at baseline to estimate the odds of recovery, i.e. reporting of no neck pain at follow-up.

165

3. Results 3.1. Subject characteristics According to the STYRK classification of occupation, the three largest groups among all employees that responded at baseline were technicians and associated professionals (N = 2075, 45.4%), service workers and shop and market sales workers (N = 844, 18.5%), and professionals (N = 750, 16.4%) (Table 1). The rest of the categories on the classification were legislators, senior officials and managers, skilled agricultural and fishery workers, craft and related trades workers, plant and machine operators and assemblers, elementary occupations, and armed forces and unspecified (817, 17.9%). Baseline distributions of employees who were invited twice but responded at baseline only and employees who responded twice were similar (see Table 1). The proportion of responders that reported that they were ‘‘a little bothered” or more severely bothered by neck pain at baseline was 49.0% (2239 employees) (Table 1). The prevalence of severe neck pain (‘‘rather intensely bothered” or ‘‘very intensely bothered”) was 19.5%. For those who responded at both time points, these prevalence rates were 47.9 and 18.4, respectively. 3.2. Non-response and attrition analyses Binary logistic regression with age and sex as predictors revealed that being in the three middle age groups (30–39, 40–49, and 50–59) significantly increased the odds of responding to the neck pain item and at least one predictor at baseline compared to being in the lowest age group (<30) (p < 0.05, analyses not shown). Sex did not predict responding at baseline. Multivariate binary logistic regression analysis showed that quantitative demands, decision control, role conflict, social climate, age, and pain intensity at baseline had independent and significant relationships with responding at follow-up for those who had already responded at baseline (p < 0.05, analyses not shown). Increased odds of responding at follow-up were observed for category 3 of quantitative demands, category 5 of decision control, categories 3 and 4 of role conflict, category 2 of social climate, and all three age groups between <30 and >59. Scoring 3 on pain intensity at baseline decreased the odds of responding at followup. 3.3. Cross-sectional analyses 3.3.1. Baseline sample Univariate regressions showed that associations for all psychological/social and mechanical variables were statistically significant except decision demands (p > 0.05) (analyses not shown). All factors except role clarity and physical workload remained statistically significant after Bonferroni-correction. ORs with p < 0.01 ranged from 0.34 (decision control) to 2.06 (working with arms raised). Neck pain was also univariately regressed on sex and age for the baseline sample (analyses not shown). Being a woman was associated with higher neck pain intensity (OR 1.81, 99% CI 1.55–2.12), and being in the age groups 40–49 (OR 1.62, 99% CI 1.20–2.17) and 50–59 (OR 1.76, 99% CI 1.30– 2.38) were associated with higher neck pain intensity when compared with being <30. Multivariate regressions showed statistically significant associations for quantitative demands, decision control, control over work intensity, role conflict, support from immediate superior, empowering leadership, fair leadership, predictability during the next month, commitment to organization, social climate, and working with arms raised to or above shoulder level (Tables 2a

166

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Table 1 Baseline characteristics of respondersa, non-responders, and responders to both time pointsb.

a b

All workers invited at baseline (N = 8928)

Workers who were invited at both baseline and follow-up (N = 6361)

Baseline responders (n = 4569)

Baseline non-responders (n = 4359)

Baseline responders (n = 3749)

Responded at both times (n = 2419)

Age <30 30–39 40–49 50–59 >59 Missing data

364 1223 1468 1202 305 7

(8.0%) (26.8%) (32.1%) (26.3%) (6.7%) (0.2%)

389 924 1041 935 379 691

(8.9%) (21.2%) (23.9%) (21.4%) (8.7%) (15.9%)

243 945 1242 1090 229 0

(6.5%) (25.2%) (33.1%) (29.1%) (6.1%) (0.0%)

126 609 821 722 141 0

(5.2%) (25.2%) (33.9%) (29.8%) (5.8%) (0.0%)

Sex Female Male Missing data

2984 1575 10

(65.3%) (34.5%) (0.2%)

2390 1278 691

(54.8%) (29.3%) (15.9%)

2449 1300 0

(65.3%) (34.7%) (0.0%)

1542 877 0

(63.7%) (36.3%) (0.0%)

Marital status Not married Married Cohabiting Widowed Divorced Separated Missing data

362 2117 501 42 205 57 1285

(7.9%) (46.3%) (11.0%) (0.9%) (4.5%) (1.2%) (28.1%)

– – – – – – –

263 1776 381 34 184 42 1069

(7.0%) (47.4%) (10.2%) (0.9%) (4.9%) (1.1%) (28.5%)

167 1138 229 17 114 29 725

(6.9%) (47.0%) (9.5%) (0.7%) (4.7%) (1.2%) (30.0%)

Classification of occupation Legislators, senior officials and managers Professionals Technicians and associate professionals Clerks Service workers and shop and market sales workers Skilled agricultural and fishery workers Craft and related trades workers Plant and machine operators and assemblers Elementary occupations Armed forces and unspecified Missing data

348 750 2075 344 844 5 39 9 72 0 83

(7.6%) (16.4%) (45.4%) (7.5%) (18.5%) (0.1%) (0.9%) (0.2%) (1.6%) (0.0%) (1.8%)

– – – – – – – – – – –

292 568 1711 284 713 2 32 7 64 0 76

(7.8%) (15.2%) (45.6%) (7.6%) (19.0%) (0.1%) (0.9%) (0.2%) (1.7%) (0,0%) (2.0%)

225 383 1100 172 429 1 26 3 40 0 40

(9.3%) (15.8%) (45.5%) (7.1%) (17.7%) (0.0%) (1.1%) (0.1%) (1.7%) (0.0%) (1.7%)

Neck pain intensity Not bothered A little bothered Rather intensely bothered Very intensely bothered

2330 1348 754 137

(51.0%) (29.5%) (16.5%) (3.0%)

– – – –

1911 1101 631 106

(51.0%) (29.4%) (16.8%) (2.8%)

1260 715 378 66

(52.1%) (29.6%) (15.6%) (2.7%)

Response at baseline is deﬁned as having completed the neck pain item and at least one predictor. Response to both time points is deﬁned as having completed the neck pain item at both baseline and follow-up, and at least one predictor at baseline.

and b). Associations for all factors remained statistically signiﬁcant after Bonferroni-correction and ORs ranged from 0.36 (commitment to organization) to 2.55 (role conﬂict).

3.3.2. Follow-up sample Univariate regressions showed statistically significant associations for all psychological/social and mechanical variables except role clarity with neck pain intensity (analyses not shown). All factors except decision demands remained statistically significant after Bonferroni-correction. ORs ranged from 0.47 (commitment to organization) to 2.24 (role conflict). Multivariate analyses showed statistically significant associations for quantitative demands, decision demands, decision control, control over work intensity, role conflict, empowering leadership, fair leadership, predictability during the next month, commitment to organization, positive challenge, and working with arms raised to or above shoulder level (Tables 2a and b). All factors remained statistically significant after Bonferroni-correction, except decision demands, control over work intensity, and positive challenge. ORs with p < 0.01 ranged from 0.40 (commitment to organization) to 3.67 (role conflict).

3.4. Prospective analyses 3.4.1. Baseline exposure as predictor Univariate regressions showed statistically significant associations for all baseline predictors except predictability during the next month, decision demands, and role clarity (analyses not shown) with neck pain intensity at follow-up. All associations remained statistically significant after Bonferroni-correction, except for quantitative demands, control over work intensity, positive challenge, and physical workload. ORs with p < 0.01 ranged from 0.37 (decision control) to 3.02 (role conflict). In the multivariate analyses decision control, role conflict, empowering leadership, and positive challenge remained statistically significant (Tables 3a and b). Statistical significance after Bonferroni-correction was maintained for role conflict and empowering leadership. ORs with p < 0.01 ranged from 0.48 (positive challenge) to 2.97 (role conflict). 3.4.2. Average exposure as predictor Univariate regressions showed statistically significant associations for all predictors except decision demands and role clarity (analyses not shown). All associations remained statistically

167

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Table 2a Cross-sectional analyses: multivariate ordinal logistic regressions with psychological, social, and mechanical factors as independent variables and neck pain intensity as outcome. Separate regressions were run with independent variables entered as categorical and continuous. Psychological, social, and mechanical factors

Quantitative demands

Category

Confounders included in multivariate analysesa

Category

99% CI

332 1113 1635 858 261 4199

1.00 1.08 1.19 1.41 1.99 1.25

[ref] [0.79–1.48] [0.88–1.62] [1.02–1.96]* [1.32–3.01]** [1.12–1.38]**

228 720 1102 615 173 2838

1.00 1.18 1.34 1.49 2.58 1.25

[ref] [0.80–1.75] [0.93–1.95] [1.01–2.22]* [1.56–4.28]** [1.10–1.42]**

1 2 3 4 5

85 235 1682 1560 747 4309

1.00 1.08 0.97 1.03 1.15 1.09

[ref] [0.59–2.01] [0.57–1.68] [0.60–1.78] [0.66–2.03] [0.98–1.21]

55 153 1165 1032 510 2915

1.00 0.69 0.84 0.98 1.07 1.14

[ref] [0.32–1.50] [0.43–1.66] [0.50–1.94] [0.54–2.16] [1.00–1.30]*

1 2 3 4 5

298 1180 1665 857 184 4184

1.00 0.65 0.62 0.50 0.37 0.78

[ref] [0.47–0.89]** [0.46–0.84]** [0.36–0.69]** [0.23–0.60]** [0.71–0.87]**

176 835 1148 614 125 2898

1.00 0.80 0.59 0.49 0.41 0.73

[ref] [0.54–1.20] [0.40–0.87]** [0.32–0.75]** [0.22–0.74]** [0.64–0.83]**

1 2 3 4 5

638 832 922 1184 878 4454

1.00 0.78 0.96 0.75 0.74 0.92

[ref] [0.61–1.01] [0.75–1.24] [0.58–0.95]** [0.57–0.96]** [0.86–0.99]*

505 541 609 797 577 3029

1.00 0.90 1.02 0.83 0.75 0.91

[ref] [0.67–1.22] [0.76–1.36] [0.62–1.10] [0.55–1.03] [0.84–0.99]*

1 2 3 4 5

931 1334 1795 304 99 4463

1.00 1.20 1.57 2.15 2.55 1.37

[ref] [0.97–1.49] [1.28–1.92]** [1.55–2.97]** [1.52–4.27]** [1.25–1.51]**

653 958 1192 201 43 3047

1.00 1.39 1.83 2.82 3.67 1.51

[ref] [1.07–1.80]* [1.43–2.35]** [1.89–4.21]** [1.68–7.95]** [1.34–1.71]**

Fair leadership, social climate

Continuous Decision control

Category

Continuous Control over work intensity

Category

Role conﬂict

Category

Continuous

Continuous Role clarity

Category

1 and 2 3 4 5

Support from immediate superior, fair leadership, commitment to organization

102 370 1094 2540 4106

1.00 1.31 1.16 1.12 0.98

[ref] [0.86–2.01] [0.78–1.73] [0.75–1.66] [0.87–1.11]

69 228 817 1773 2887

1.00 1.12 1.01 1.07 1.06

[ref] [0.57–2.23] [0.54–1.92] [0.57–2.02] [0.91–1.23]

1 2 3 4 5

Fair leadership

145 228 928 1225 1830 4356

1.00 0.61 0.58 0.54 0.50 0.87

[ref] [0.36–1.05] [0.35–0.94]* [0.33–0.89]** [0.30–0.83]** [0.78–0.97]*

85 157 631 853 1218 2944

1.00 1.30 1.03 0.90 0.82 0.92

[ref] [0.65–2.60] [0.54–1.95] [0.47–1.75] [0.42–1.60] [0.80–1.05]

Continuous Support from immediate superior

Category

Continuous a * **

Follow-up sample

N 1 2 3 4 5

Continuous Decision demands

Baseline sample

Age and sex were included as confounders in all multivariate models. p < 0.01. p < 0.0007, which is the Bonferroni-corrected threshold based on the number of factors tested (0.01/15).

significant after Bonferroni-correction. ORs with p < 0.01 ranged from 0.32 (decision control) to 3.90 (role conflict). The multivariate regressions showed statistically significant associations for decision control, role conflict, role clarity, empowering leadership, fair leadership, predictability during the next month, commitment to organization, and working with arms raised to or above shoulder level (Tables 3a and b). Role conflict and empowering leadership remained statistically significant after Bonferroni-correction. ORs with p < 0.01 ranged from 0.53 (empowering leadership) to 3.01 (role clarity). 3.4.3. Change in exposure as predictor Univariate regressions showed statistically significant associations for all predictors except decision demands, role clarity, predictability during the next month, and positive challenge (analyses not shown). Statistical significance was maintained for all factors after Bonferroni-correction except for control over work intensity. ORs with p < 0.01 ranged from 0.40 (constant high fair leadership) to 2.69 (constant high role conflict).

The confounder estimation procedure revealed that none of the psychological/social and mechanical factors acted as confounders. Hence, all multivariate regressions were conducted only with age, sex, and baseline pain as covariates. Multivariate regressions showed statistically significant associations for decision control, role conflict, empowering leadership, social climate, physical workload, and working with arms raised (Tables 4a and b). The association remained statistically significant after Bonferroni-correction for constant high social climate. ORs ranged from 0.58 (constant high social climate) to 2.16 (constant high role conflict). 3.4.4. 3.4.4.1. Potentially etiologic factors – onset of pain. Of the employees that reported of no pain at baseline, 283 subjects reported of neck pain at follow-up (incidence rate 22.4%). Among those employees that were pain free at baseline, empowering leadership was a statistically significant predictor of neck pain at follow-up (OR 0.83, p < 0.01, analyses not shown).

168

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Table 2b Cross-sectional analyses: multivariate ordinal logistic regressions with psychological, social, and mechanical factors as independent variables and neck pain intensity as outcome. Separate regressions were run with independent variables entered as categorical and continuous. Psychological, social, and mechanical factors

Empowering leadership

Category

Confounders included in multivariate analysesa

Baseline sample N

99% CI

495 684 1603 999 700 4481

1.00 0.76 0.64 0.54 0.47 0.81

[ref] [0.57–1.02] [0.50–0.83]** [0.41–0.71]** [0.35–0.63]** [0.75–0.87]**

314 440 1101 699 501 3055

1.00 0.85 0.62 0.52 0.56 0.83

[ref] [0.60–1.21] [0.45–0.84]** [0.38–0.73]** [0.40–0.80]** [0.76–0.91]**

98 192 831 1131 1968 4220

1.00 1.15 0.77 0.76 0.61 0.85

[ref] [0.63–2.11] [0.45–1.33] [0.45–1.30] [0.36–1.04] [0.77–0.94]**

59 149 573 797 1396 2974

1.00 1.73 1.01 0.84 0.73 0.79

[ref] [0.83–3.65] [0.52–2.00] [0.43–1.66] [0.37–1.44] [0.70–0.90]**

1 2 3 4 5

62 107 408 1161 2749 4487

1.00 0.80 0.82 0.69 0.54 0.81

[ref] [0.37–1.73] [0.42–1.59] [0.37–1.31] [0.29–1.02] 0.73–0.90**

28 63 266 756 1926 3039

1.00 1.78 1.31 0.98 0.80 0.76

[ref] [0.58–5.59] [0.50–3.58] [0.39–2.59] [0.32–2.10] [0.67–0.87]**

1 2 3 4 5

192 337 1146 1277 1355 4307

1.00 0.87 0.62 0.47 0.36 0.72

[ref] [0.56–1.35] [0.42–0.90]* [0.32–0.69]** [0.25–0.53]** [0.66–0.78]**

114 203 834 894 1009 3054

1.00 0.93 0.62 0.53 0.40 0.74

[ref] [0.53–1.63] [0.39–1.01] [0.33–0.86]** [0.25–0.64]** [0.67–0.82]**

1 2 3 4 5

Continuous Fair leadership

Category

1 2 3 4 5

Commitment to organization

Continuous Predictability during the next month

Category

Continuous Commitment to organization

Category

Continuous Social climate

Category

1 and 2 3 4 5

Fair leadership, commitment to organization

203 1054 1452 1414 4123

1.00 0.65 0.59 0.54 0.86

[ref] [0.44–0.96]* [0.39–0.88]** [0.35–0.82]** [0.75–0.98]*

109 774 1001 1023 2907

1.00 1.37 0.91 0.88 0.86

[ref] [0.81–2.35] [0.53–1.58] [0.50–1.56] [0.73–1.01]

1 and 2 3 4 5

Commitment to organization

140 630 1396 1885 4051

1.00 1.08 0.98 0.92 0.96

[ref] [0.68–1.72] [0.63–1.55] [0.58–1.44] [0.85–1.08]

76 403 990 1334 2803

1.00 0.83 0.68 0.63 0.85

[ref] [0.45–1.55] [0.37–1.25] [0.35–1.16] [0.73–0.99]*

1 2 3 4

Working with arms raised to or above shoulder level

2668 1061 438 218 4385

1.00 1.08 1.08 1.03 1.03

[ref] [0.89–1.31] [0.81–1.43] [0.70–1.52] [0.93–1.14]

1756 733 326 142 2957

1.00 1.22 1.16 1.14 1.09

[ref] [0.96–1.54] [0.83–1.63] [0.71–1.83] [0.95–1.25]

2961 990 374 177 4502

1.00 1.21 1.28 1.73 1.17

[ref] [1.01–1.44]* [0.98–1.67] [1.18–2.54]** [1.07–1.28]**

1980 693 263 112 3048

1.00 1.22 1.42 2.08 1.23

[ref] [0.98–1.52] [1.03–1.96]* [1.31–3.28]** [1.10–1.38]**

Continuous Positive challenge

Category

Physical workload

Category

Continuous

Continuous Working with arms raised to or above shoulder level

Category

1 2 3 4

Continuous a * **

Follow-up sample

Age and sex were included as confounders in all multivariate models. p < 0.01. p < 0.0007, which is the Bonferroni-corrected threshold based on the number of factors tested (0.01/15).

3.4.4.2. Potentially prognostic factors in subjects with pain. Of the employees that reported of pain at baseline, 284 subjects reported of no pain at follow-up (incidence rate 15.5%). Role conflict (OR 0.71, p < 0.0007) and commitment to organization (OR 1.31, p < 0.01) were statistically significant predictors of being pain free at follow-up (analyses not shown). 4. Discussion This study clearly demonstrated the role of both psychological/ social and mechanical factors for the experience of neck pain. A strength was the comprehensive set of exposures measured. The most robust predictors were role conflict and empowering leadership. These factors were statistically significant in all regressions, with p < 0.0007 in 4 of 5 regressions. Decision control was statistically significant in all analyses. Mechanical factors were not confounders.

We have not found studies relating role conflict or empowering leadership to neck pain. Role conflict has been linked to strain [8], well-being [42], and sickness absence [36]. Empowering leadership has been reported not to predict sick leave due to MSD [5]. Hence, the current findings represent important new information that warrants further investigations. Working with arms raised is an established risk factor for neck and shoulder pain [18,20,33,48], and was both cross-sectionally and prospectively related to outcome in this study. Physical workload was only statistically significant across time, for increased and constant high levels. Systematic reviews have concluded that job demands is the psychological factor most consistently associated with neck/shoulder pain [2,6,7,37,38,47,49]. The current results partly supported the role of quantitative demands. Decision demand was not a risk factor. This seems to be at odds with the previous research. One common measure of job demands in Karasek’s Job Content

169

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Table 3a Prospective analyses: multivariate ordinal logistic regressions with psychological, social, and mechanical factors measured at baseline (T1) and averaged across time [(T1 + T2)/2] as independent variables, and neck pain intensity at follow-up as outcome. Separate regressions were run with predictors entered as categorical and continuous. Psychological, social, and mechanical factors

Quantitative demands

Category

Confounders included in multivariate analysesa

Baseline exposure as predictor

Average exposure as predictor

99% CI

150 589 932 455 125 2251

1.00 0.84 0.94 0.86 1.19 0.99

[ref] [0.51–1.38] [0.59–1.51] [0.52–1.44] [0.62–2.28] [0.85–1.17]

98 490 1041 408 65 2102

1.00 0.76 0.90 1.00 1.40 1.10

[ref] [0.42–1.41] [0.51–1.61] [0.55–1.86] [0.60–3.28] [0.92–1.32]

1 and 2 3 4 5

160 898 846 385 2289

1.00 0.91 0.82 0.93 0.97

[ref] [0.58–1.44] [0.52–1.29] [0.57–1.54] [0.82–1.14]

148 758 1012 249 2167

1.00 0.90 0.95 0.99 1.05

[ref] [0.56–1.47] [0.60–1.53] [0.56–1.73] [0.87–1.26]

1 2 3 4 5

137 605 886 485 122 2235

1.00 0.69 0.70 0.60 0.66 0.89

[ref] [0.42–1.13] [0.43–1.12] [0.36–1.00]* [0.33–1.30] [0.76–1.04]

74 546 913 469 93 2095

1.00 0.86 0.70 0.64 0.61 0.82

[ref] [0.45–1.66] [0.37–1.33] [0.33–1.26] [0.26–1.43] [0.69–0.98]*

1 2 3 4 5

340 431 479 620 511 2381

1.00 1.09 1.26 0.97 1.01 0.98

[ref] [0.74–1.62] [0.86–1.85] [0.67–1.42] [0.69–1.49] [0.88–1.09]

282 436 543 629 446 2336

1.00 1.16 1.31 1.01 0.97 0.96

[ref] [0.77–1.75] [0.89–1.95] [0.68–1.50] [0.64–1.48] [0.85–1.07]

1 2 3 4 5

477 717 980 164 40 2378

1.00 1.01 1.24 1.48 2.97 1.25

[ref] [0.73–1.40] [0.91–1.69] [0.91–2.39] [1.29–6.74]** [1.08–1.45]**

358 1006 816 156 15 2351

1.00 1.32 1.58 2.89 2.10 1.41

[ref] [0.94–1.88] [1.11–2.27]* [1.72–4.87]** [0.48–8.32] [1.19–1.67]**

1 2 3 4 5

Fair leadership, social climate

Continuous Decision demands

Category

Continuous Decision control

Category

Control over work intensity

Category

Continuous

Continuous Role conﬂict

Category

Role clarity

Category

Continuous 1 and 2 3 4 5

Support from immediate superior, fair leadership, commitment

52 188 601 1332 2173

1.00 1.28 1.56 1.42 1.00

[ref] [0.54–3.14] [0.69–3.67] [0.63–3.32] [0.83–1.20]

38 134 656 1228 2056

1.00 3.01 2.29 2.33 1.04

[ref] [1.08–8.90]* [0.88–6.39] [0.89–6.49] [0.84–1.29]

1 2 3 4 5

Fair leadership

71 114 494 642 999 2320

1.00 1.16 1.13 0.84 1.05 1.00

[ref] [0.50–2.74] [0.53–2.45] [0.39–1.87] [0.48–2.33] [0.84–1.17]

28 104 450 799 836 2217

1.00 0.96 0.98 0.73 0.82 0.99

[ref] [0.26–3.68] [0.28–3.66] [0.20–2.77] [0.22–3.13] [0.81–1.22]

Continuous Support from immediate superior

Category

Continuous a * **

Baseline measures of age, sex, and neck pain (T1) were included as confounders in all multivariate models. p < 0.01. p < 0.0007, which is the Bonferroni-corrected threshold based on the number of factors tested (0.01/15).

Questionnaire (JCQ) [26] incorporates ‘‘intense concentration” [26]. This appears to correspond to decision demands of the QPSNordic (e.g. ‘‘maximum attention”). Job demands of the JCQ also incorporate ‘‘conﬂicting demands”. This appears to correspond to role conﬂict (QPSNordic) (e.g. ‘‘receive opposing requests from two or more persons”), which was the strongest predictor in our study. Hence, the ‘‘job demands” concept of previous studies may include factors discriminately related to neck pain, which may explain inconsistent results. Systematic reviews have designated decision control an important, but less consistent, predictor [2,6,7,38,47,49]. The present study supported the role of decision control. Inconsistency is perhaps due to conceptual imprecision. ‘‘Decision latitude” (JCQ) encompasses ‘‘decision authority” and ‘‘skill discretion” [26]. The former is closest to decision control of the QPSNordic. The latter includes the items ‘‘learn new things”, ‘‘requires creativity”, ‘‘high skill level”, and ‘‘develop own abilities”. This seems conceptually related to the less predictive positive challenge (QPSNordic) (e.g. ‘‘are your special knowledge and skills useful in your work”, ‘‘is your work challenging in a positive way”).

Supervisory support has received limited support as a predictor of neck pain [2,38,49]. In this study no statistically significant prospective ORs were detected for support from immediate superior. However, support was adjusted for fair leadership, which may be conceptually related [10] or a consequence of support from the superior. This is a general concern in multivariate analyses. The number of predictors in each model was limited to avoid overadjustment. However, overadjustment cannot be ruled out. For instance, commitment to organization was included in several analyses. Commitment could be an outcome of other work factors (e.g. positive challenge) rather than a confounder. Commitment to the organization was cross-sectionally related to pain at both time points. Prospectively, only average commitment as a continuous variable reached statistical significance. Organizational commitment (QPSNordic) has recently been associated with subjective health after 3.5 [35] and 4 years [23]. Hence, 2 years may not be sufficient. Prolonged exposure to demanding work situations with sustained high levels of physiological arousal and insufficient recovery may induce poor health [16]. For decision control, role conflict,

170

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173

Table 3b Prospective analyses: multivariate ordinal logistic regressions with psychological, social, and mechanical factors measured at baseline (T1) and averaged across time [(T1 + T2)/2] as predictors, and neck pain intensity at follow-up as outcome. Separate regressions were run with predictors entered as categorical and continuous. Psychological, social, and mechanical factors

Empowering leadership

Category

Fair leadership

Category

Confounders included in multivariate analysesa

Baseline exposure as predictor

Average exposure as predictor

99% CI

235 363 866 531 392 2387

1.00 0.74 0.63 0.53 0.64 0.88

[ref] [0.48–1.14] [0.43–0.93]* [0.35–0.81]** [0.41–0.99]* [0.79–0.99]*

168 447 776 665 310 2366

1.00 0.79 0.71 0.54 0.73 0.87

[ref] [0.50–1.26] [0.46–1.10] [0.34–0.84]** [0.44–1.21] [0.77–0.98]*

43 93 444 600 1052 2232

1.00 1.26 1.06 0.91 0.82 0.89

[ref] [0.48–3.39] [0.44–2.57] [0.38–2.20] [0.35–1.98] [0.76–1.03]

22 78 348 872 850 2170

1.00 0.72 0.68 0.59 0.47 0.82

[ref] [0.21–2.53] [0.22–2.23] [0.19–1.90] [0.15–1.54] [0.68–0.99]*

1 and 2 3 4 5

66 191 612 1523 2392

1.00 0.95 1.05 0.88 0.90

[ref] [0.44–2.08] [0.53–2.15] [0.45–1.77] [0.77–1.06]

36 173 732 1419 2360

1.00 0.95 0.79 0.65 0.82

[ref] [0.35–2.67] [0.31–2.07] [0.26–1.71] [0.68–1.00]*

1 2 3 4 5

88 170 577 697 741 2273

1.00 0.76 0.68 0.62 0.67 0.91

[ref] [0.38–1.51] [0.38–1.26] [0.34–1.14] [0.37–1.22] [0.80–1.04]

42 188 530 872 621 2253

1.00 1.36 0.98 0.99 0.85 0.87

[ref] [0.56–3.42] [0.42–2.35] [0.43–2.35] [0.37–2.04] [0.75–1.00]*

1 2 3 4 5

Continuous 1 2 3 4 5

Commitment to organization

Continuous Predictability during the next month

Category

Continuous Commitment to organization

Category

Continuous Social climate

Category

1 and 2 3 4 5

Fair leadership, commitment to organization

102 531 783 769 2185

1.00 1.12 0.94 0.92 0.89

[ref] [0.61–2.07] [0.51–1.76] [0.48–1.76] [0.73–1.09]

67 421 1036 567 2091

1.00 1.23 1.00 0.92 0.85

[ref] [0.60–2.58] [0.48–2.12] [0.42–2.05] [0.66–1.10]

1 and 2 3 4 5

Commitment to organization

57 311 712 1055 2135

1.00 0.56 0.51 0.48 0.90

[ref] [0.27–1.19] [0.25–1.06] [0.23–0.99]* [0.75–1.07]

41 218 852 860 1971

1.00 0.86 0.64 0.62 0.85

[ref] [0.34–2.18] [0.26–1.58] [0.25–1.56] [0.68–1.07]

1 2 3 4

Working with arms raised to or above shoulder level

1390 611 229 120 2350

1.00 1.25 1.24 1.04 1.12

[ref] [0.94–1.65] [0.81–1.90] [0.58–1.82] [0.95–1.32]

1416 544 245 60 2265

1.00 1.27 1.06 1.15 1.16

[ref] [0.95–1.69] [0.68–1.64] [0.53–2.42] [0.96–1.41]

1539 563 201 91 2394

1.00 1.19 1.40 1.17 1.12

[ref] [0.91–1.55] [0.95–2.07] [0.66–2.05] [0.98–1.29]

1680 496 169 20 2365

1.00 1.23 1.45 1.49 1.20

[ref] [0.93–1.61] [0.95–2.18] [0.47–4.63] [1.03–1.41]*

Continuous Positive challenge

Category

Continuous Physical workload

Category

Continuous Working with arms raised to or above shoulder level

Category

1 2 3 4

Continuous a * **

support from immediate superior, empowering leadership, fair leadership, commitment to organization, social climate and working with arms raised, this was indicated by constant high levels displaying the strongest ORs among levels of the change/constancy factors. For some factors the risk associated with constant high exposure was comparable to that of increasing exposure. Hence, prolonged exposure cannot be concluded to have an effect beyond that of novel exposure at follow-up. For some factors, decreasing exposure seemed to pose a risk comparable to that of increasing exposure. This could signify long-term effects of high risk factors or low protective factors at baseline. If pain processes were initiated at baseline, mechanisms of chroniﬁcation could result in heightened pain levels in absence of prolonged exposure to high risk factors or low protective factors [31]. Most previous studies assessed only baseline exposure. Thus, the current study adds information and strengthens assumptions of causality by demonstrating that exposure at T1 predicts neck pain at T2 when exposure at T2 is lowered.

4.1. Methodological considerations The baseline response rate was 51.2%. Possibly not all employees received enough information to motivate them to answer the comprehensive questionnaire. Self-selection is a threat to internal validity if it is related to the exposure and, independently of exposure, to the disease/outcome [40]. Prevalences of a range of health complaints, including neck pain, have been shown not to differ between responders and non-responders [17,46]. Prevalences in the present study were consistent with those commonly reported [17]. The survey involved a comprehensive range of work and health factors apart from those included in the current study. Adverse working conditions and illness was not a speciﬁc focus. At baseline 56.9% of the invited employees responded to one or more items and the response rate declined gradually throughout the questionnaire. Hence, there is little reason to suspect self-selection to be based on a combination of high exposure and high neck pain intensity.

171

J.O. Christensen, S. Knardahl / PAIN 151 (2010) 162–173 Table 4a Prospective analyses: multivariate ordinal logistic regressions with the change in psychological, social, and mechanical factors as predictors and neck pain intensity at follow-up as outcome.

a * **

Psychological, social, and mechanical factors