12
1 23 Journal of Occupational Rehabilitation ISSN 1053-0487 Volume 21 Number 3 J Occup Rehabil (2011) 21:421-430 DOI 10.1007/s10926-011-9285-5 Self-Efficacy and Health Locus of Control: Relationship to Occupational Disability Among Workers with Back Pain Sylvie Richard, Clermont E. Dionne & Arie Nouwen

Self-Efficacy and Health Locus of Control: Relationship to Occupational Disability Among Workers with Back Pain

Embed Size (px)

Citation preview

1 23

Journal of OccupationalRehabilitation ISSN 1053-0487Volume 21Number 3 J Occup Rehabil (2011) 21:421-430DOI 10.1007/s10926-011-9285-5

Self-Efficacy and Health Locus of Control:Relationship to Occupational DisabilityAmong Workers with Back Pain

Sylvie Richard, Clermont E. Dionne &Arie Nouwen

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer

Science+Business Media, LLC. This e-offprint

is for personal use only and shall not be self-

archived in electronic repositories. If you

wish to self-archive your work, please use the

accepted author’s version for posting to your

own website or your institution’s repository.

You may further deposit the accepted author’s

version on a funder’s repository at a funder’s

request, provided it is not made publicly

available until 12 months after publication.

Self-Efficacy and Health Locus of Control: Relationshipto Occupational Disability Among Workers with Back Pain

Sylvie Richard • Clermont E. Dionne •

Arie Nouwen

Published online: 29 January 2011

� Springer Science+Business Media, LLC 2011

Abstract Objectives Although self-efficacy and health

locus of control (HLC) have been extensively studied in

health research, little is known about their contribution to

occupational disability among workers with back pain.

This 2 year prospective study examined the association

between these control belief constructs and ‘‘return to

work in good health’’ (RWGH), a four-category, composite

index of back pain outcome. Methods The participants

(n = 1,007, participation = 68.4%, follow-up = 86%)

were workers with occupational disruptions who sought a

medical consultation for non specific back pain in primary

care and emergency settings in the Quebec City area,

Canada. Information about self-efficacy for return to work

(SERW) and HLC, as well as potential confounders, was

collected during a telephone interview about 3 weeks after

the baseline medical consultation. Polytomous logistic

regression was used to assess the relationship between the

baseline control variables and RWGH at 2 year. Odds

ratios (OR) and their 95% confidence intervals were used

to quantify the strength of associations. For all analyses,

the ‘‘success’’ category was considered the reference

group. Results Although bivariate analyses showed a sig-

nificant association between external HLC and RWGH at

2 year, this relationship was not significant in multivariate

analyses. Higher scores on the self-efficacy questionnaire

were however protective of ‘‘failure to return to work after

attempt(s)’’ (OR: 0.28; 95% CI: 0.14–0.57) and of ‘‘failure

to return to work’’ (OR: 0.19; 95% CI: 0.07–0.48) in multi-

variate analyses. Conclusion Self-efficacy is an important

determinant of the occupational outcome of back pain.

Keywords Self-efficacy � Health locus of control �Control beliefs � Back pain � Occupational disability

Introduction

Back pain is one of the leading causes of chronic disability

and incapacity for work in the western world [1, 2]. For-

tunately, for the vast majority of cases, return to previous

activities occurs within the first 3 months following the

onset of the episode [3]. However, in some individuals,

back pain progresses into chronic disability, even though

no important physical changes may be detected [4].

Because the socioeconomic burden of back pain is deter-

mined mostly by this minority of cases [3], the under-

standing of this phenomenon appears important.

It is now widely recognized that chronic pain cannot be

accurately understood as simply the product of noxious

peripheral inputs [5]. A more comprehensive view should

include its sensory, affective and cognitive dimensions

[5, 6]. The literature indicates that cognitions pertaining to

perceived control play a major role in the adjustment of

S. Richard � C. E. Dionne (&)

URESP du Centre de recherche FRSQ du CHA universitaire de

Quebec, Hopital du Saint-Sacrement, 1050, chemin Ste-Foy,

Quebec, QC G1S 4L8, Canada

e-mail: [email protected]

S. Richard

Department of Social and Preventive Medicine,

Laval University, Quebec, Canada

C. E. Dionne

Department of Rehabilitation, Laval University, Quebec,

Canada

A. Nouwen

School of Psychology, University of Birmingham,

Edgbaston, Birmingham, UK

123

J Occup Rehabil (2011) 21:421–430

DOI 10.1007/s10926-011-9285-5

Author's personal copy

individuals with chronic pain [7–9]. Two control related

constructs, self-efficacy and health locus of control, have

been widely used in health research. Self-efficacy refers to

one’s confidence in the ability to perform a particular

behaviour and to overcome barriers to that behaviour [10].

This construct was found to be an important factor in the

control of pain, adaptive psychological functioning, phys-

ical performance and disability [7, 11–17].

Health locus of control (HLC) refers to beliefs relevant

to the location where control resides regarding one’s health

[18]. Its corresponding main measure, the Multidimen-

sional HLC (MHLC) scale [19] assesses the extent to

which individuals believe that their health is related to their

own behaviours (internal HLC) or to some external causes

(external HLC) like chance or fate (chance HLC) or

powerful others such as doctors and family members

(powerful others HLC). In general, ‘‘internals’’ are likely to

demonstrate a better adjustment to illness [20–22], whereas

those who do not feel able to control their own health

(‘‘externals’’) are less likely to do so [23, 24]. Not all

studies have however confirmed this hypothesis [25, 26].

Few studies have looked at control beliefs to explain

return to work among workers with back pain. Moreover,

to the authors’ knowledge, no study has examined self-

efficacy concurrently with HLC using a conceptual

framework such as Bandura’s self-efficacy theory. Bandura

[10] underscores the importance to differentiate these

control beliefs: believing that one’s health status is affected

by one’s own behaviours (internal HLC) is conceptually

different than believing in one’s ability to perform a given

behaviour (self-efficacy) (Fig. 1). Based on Bandura’s

theory [10], self-efficacy would be a better determinant of

health behaviours than HLC, although this assumption

needs to be verified among workers with back pain. Clar-

ifying which of these constructs is most influential to

explain occupational outcome among workers with back

pain, and integrating this knowledge into clinical practice,

has the potential to have an important impact on the quality

of care offered to this population.

The aim of the current study was to investigate the

association between HLC and ‘‘self-efficacy for return to

work’’ (SERW) on the one hand, and ‘‘return to work in

good health’’ (RWGH) 2 years later, on the other hand,

among workers with occupational disruptions consulting in

primary care and emergency settings for non specific back

pain. Drawing from the self-efficacy theory, it was

hypothesized that: (1) control perception (SERW and

HLC) would be associated with ‘‘return to work in good

health’’ (RWGH) 2 years later among workers who con-

sulted in primary care settings for back pain; and (2)

SERW would prove a better determinant than HLC to

explain RWGH 2 years later among workers who con-

sulted in primary care settings for back pain.

Materials and Methods

Study Design

This study used data collected as part of the RAMS-

Prognosis Study [25, 27], a 2 year prospective study with

repeated measurements on long-term occupational out-

come of back pain. The current study is based on data

collected at baseline and at the 2 year follow-up.

Study Settings and Selection of Participants

Subjects were recruited in 1999–2000 in 7 major medical

settings of the Quebec City area: 4 emergency room

departments and 3 family medicine practices. Eligible

subjects were workers aged 18–64 years who consulted for

back pain (including the cervico-thoracic, thoracic, lumbar

and lumbo-sacral areas) that had affected their capacity to

accomplish their regular work for at least 1 day (by self-

report). Participants had to be fluent in French. Exclusion

criteria were pregnancy, cancer, spinal infection, vertebral

fracture, systemic disease, cauda equina syndrome, referred

visceral pain and any major medical illness that could

affect work status (e.g. severe heart disease and psychiatric

illness). This study was approved by the ethics committees

of all the institutions involved.

Data Collection

Data were collected through structured telephone inter-

views and to a lesser extent from medical records. Partic-

ipants were contacted about 3 weeks after the medical

consultation. Information was gathered on the following

topics: back pain history, sociodemographic characteristics,

as well as SERW and HLC beliefs. Others questions were

Own behaviours (Internal) Powerful Others (External) Chance or Fate (External)

Health Locus of Control (HLC)

Self-efficacy (SERW)

Return to regular work

Performance Agent

Worker

Results

Health Pain Functional limitations

Fig. 1 Illustration of the theoretical constructs of the study (adapted

from Bandura’s self-efficacy theory). Self-efficacy refers to the

worker’s (agent) confidence to return to his or her regular work

(performance). The results represent the possible outcomes of a return

to work. The HLC construct refers to the worker’s beliefs that the

health-related results/outcomes are due to his or her behaviour/

performance or to some external causes like chance, fate or powerful

others [10]

422 J Occup Rehabil (2011) 21:421–430

123

Author's personal copy

also asked about factors that were considered as potential

confounders of the association between perception of

control and long-term occupational outcome: (1) Fear-

avoidance beliefs related to work were measured with 3

items of the ‘‘Work’’ scale of the Fear-Avoidance Beliefs

Questionnaire (FABQ) [5] (My work aggravated my pain;

My work makes or would make my pain worst; My work

might harm my back). It had been decided a priori to

exclude items of the scale that were, to the authors’ judg-

ment, more closely related to self-efficacy (e.g. I cannot do

my normal work with my present pain), (2) perception of

general health status (one question), (3) average back pain

intensity in the past 6 months measured with an 11-point

Numerical Rating Scale [28], and (4) physical demands of

job (from self-report): an index made from answers to the

questions How would you rate the physical demands of

your job? on an 1–11 scale multiplied by How often do you

have to lift heavy objects like boxes or pieces of furniture?

measured on an 1–4 scale (range: 1–44).

The MHLC Scale (Form A), developed by Wallston

et al. [19], and translated into Canadian French by Talbot

et al. [29], was used to measure HLC. This 18-item

instrument consists of three subscales, each measuring a

different dimension of HLC: (1) Internal HLC (IHLC), the

belief that one’s behaviour will have an effect on one’s

health status; (2) Powerful Others HLC (PHLC), the belief

that powerful other people, such as doctors and family

members, have control over one’s health status; and (3)

Chance HLC (CHLC), the belief that one’s health condi-

tion is a matter of fate, luck or chance. Each subscale

contains 6 items answered on a 6-point Likert scale from

‘‘Strongly disagree’’ to ‘‘Strongly agree’’. The internal

consistency (Cronbach’s alpha) of the three subscales of

the translated version is: Internal (0.64), Powerful others

(0.69), and Chance (0.67). Overall, internal consistency of

the Canadian French translation is comparable to that

reported for the original English version. Test–retest reli-

ability (Pearson’s correlation coefficient) of the Canadian

French version over a 6 week interval was 0.58, 0.65, and

0.63 for IHLC, PHLC and CHLC, respectively [29].

The Self-efficacy for return to work questionnaire is

presented in Appendix. This scale was developed by

Dionne and Nouwen [25, 27] according to Bandura’s

guidelines [10]. The authors recognize that formal adap-

tation and validation of this English translation remain to

be done. This 8-item instrument assesses the worker’s

confidence in his or her ability to return to previous work

activities despite the presence of specific difficulties or

obstacles (e.g. intensity of pain, tension at work, etc.).

These obstacles were identified in a previous qualitative

study [27]. The response format of this questionnaire is a

101-point Numerical Rating Scale where 0% = ‘‘Not at all

confident’’ and 100% = ‘‘Very confident’’. The total range

is 0–800, with higher scores indicating higher perceived

self-efficacy for return to work. The total score of this scale

was transformed to a scale of 0–100 by dividing by 8. The

internal consistency of this measure was evaluated [27] and

was found to be good (Cronbach’s alpha = 0.88).

Outcome Measure

‘‘Return to Work in Good Health’’ (RWGH), developed by

Dionne et al. [25, 27], is a composite index of the occu-

pational outcome of workers suffering from back pain. This

variable has four categories (Table 1) (Success, Partial

success, Failure after attempt and Failure) and takes into

account: (1) the occupational status (return vs no return to

previous work activities), (2) back-related functional lim-

itations (measured with the Roland-Morris Disability

Questionnaire), (3) number of days of work absence due to

back pain and (4) the presence or not of attempts to return

to work. It has been described in details elsewhere [27].

Statistical Analyses

All analyses were performed using the SAS computer

program, version 9.1 [30]. Descriptive analyses were first

conducted to identify the baseline characteristics of the

participants. The associations between control belief con-

structs and some other continuous variables were measured

with the Spearman’s correlation coefficient. Explicative

analyses were then performed to assess which of the

variables measured at baseline were associated with

RWGH at 2 years. Given the negatively skewed distribu-

tion observed for the internal HLC subscale, the score of

each HLC subscale was dichotomized (High, Low) using

the median as the cut-off point. Scores of self-efficacy were

divided into 3 categories (High, Moderate, Low) along

tertiles. Because the outcome measure in this study was a

variable in four categories, polytomous logistic regression

was conducted. For all analyses, the ‘‘Success’’ category

was considered the reference group. Odds ratios (OR) and

their 95% confidence intervals (95% CI) were used to

quantify the strength of the associations identified.

Bivariate analyses were conducted first to assess the

crude association between each independent variable and

RWGH at 2 years, using a 0.05 significance threshold. The

potentially modifying effect of gender on the association

between each control beliefs variable and RWGH was

assessed by adding an interaction term to the models. Since

no actual effect modification was detected (P-value [ 0.1),

both genders were treated simultaneously in analyses.

Multivariate analyses were then conducted with all the

control-related variables in the model. Potential con-

founders were identified from the literature according to

specific criteria: the variable had to be (1) a known

J Occup Rehabil (2011) 21:421–430 423

123

Author's personal copy

determinant of the outcome, (2) associated with at least one

of the independent variables considered and (3) not a

mediating variable in the causal pathway between the

control beliefs variables and the occupational outcome.

The potential confounders considered in this study were:

age, gender, average pain over the past 6 months, previous

episodes of back-pain in the past 10 years having affected

work capacity, fear-avoidance beliefs about work, fear-

avoidance beliefs about activity, perceived health status

and perceived physical demands of job.

Each of the potential confounding variables was put in

the basic model one at a time. When the odds ratio of at

least one independent variable in the model was changed

by at least 10%, the potential confounding variable was

identified as an actual confounder and included in the final

model [31]. In order to be able to compare the results of

this study with previous work, it was decided a priori to

adjust for age and gender even if these variables were not

found to be actual confounders. Finally, the absence of

collinearity among variables of the final model was tested

with the ‘‘COLLIN’’ option of the SAS software.

Results

A total of 1,007 individuals (68.4% of those eligible)

agreed to participate in this study and completed the tele-

phone interview at baseline. Of these, 86% (n = 867)

completed the 2 year follow-up interview and 860 had

complete information. Some differences were noted

between the patients who completed the study and those

who did not. Subjects lost to follow-up were more often

male (P = 0.009), had a lower level of education

(P = 0.04), had lower annual personal income (P = 0.03),

reported higher average pain level over the past 6 months

(P = 0.003) and had been working for a shorter period in

their actual job (P = 0.04) than those who completed the

study. In addition, individuals lost to follow-up reported a

lower level of SERW than participants (P = 0.02). Anal-

yses comparing participants with eligible individuals who

did not participate showed only a significant difference in

gender: participants were more often female (P = 0.006).

Table 2 presents selected baseline characteristics of the

participants. The mean age of the subjects in this study was

38.7 years (SD = 10.6). The majority of the participants

were male (58.5%), married or living as married (70.3%)

and had at least finished secondary school (81.1%). The

vast majority of them reported a recurrent or a persistent

back pain problem (77.8%).The mean number of years at

the job was 7.5 (SD = 8.3) and most of the participants

reported a full-time work schedule (80.2%). Compared

with the Quebec City area workers’ population, the study

participants were less educated [32].

At the 2 year follow-up, 80.8% of participants were

back to their regular job (‘‘partial success’’ or ‘‘success’’),

while 7.4% had not even attempted to return to work.

Health locus of control variables were only weakly

correlated with SERW. However, correlations between

SERW, and fear avoidance beliefs and pain intensity were

moderate and statistically significant (Table 3).

In bivariate analyses, all of the control beliefs variables

considered, except internal HLC, were statistically associ-

ated with RWGH at 2 years and the associations were in

the expected direction. Individuals who demonstrated a

higher level of Powerful Others HLC were more likely to

be found in the ‘‘Failure after attempt’’ (OR: 1.84, 95% CI:

1.19–2.87) or in the ‘‘Failure’’ categories (OR: 1.87, 95%

CI: 1.10–3.19). Similar results were observed between

Chance HLC and RWGH: higher level of Chance HLC was

associated with ‘‘Failure’’ (OR: 1.89, 95% CI: 1.10–3.27).

Participants who presented a high or moderate level of

SERW at baseline were less likely to be found in the

‘‘Failure after attempt’’ (high SERW–OR: 0.18, 95% CI:

0.09–0.36; moderate SERW–OR: 0.61, 95% CI: 0.37–1.00)

or in the ‘‘Failure’’ (high SERW–OR: 0.11, 95% CI:

0.04–0.27; moderate SERW–OR: 0.35, 95% CI: 0.19–0.67)

groups at 2 years (Table 4).

In multivariate analyses, only SERW remained statisti-

cally associated with RWGH after controlling for age,

gender, average pain over the past 6 months and fear-

avoidance beliefs about work. Higher levels of SERW were

Table 1 Definition of ‘‘return to work in good health’’

Category Definition

Success At regular work at time of interview and functional

limitations B30% and number of days of absence

over the past year B7

OR

Other situation unrelated to back pain (e.g. became

homemaker by choice)

Partial success PS1: At regular work at time of interview and

functional limitations B30% and number of days

of absence over the past year [7

OR

PS2: At regular work at time of interview and

functional limitations [30% and number of days

of absence over the past year B7

OR

PS3: At regular work at time of interview and

functional limitations [30% and number of days

of absence over the past year [7

Failure after

attempt

Absent from regular work at time of interview

because of back pain and at least one attempt to

return to regular work during the 2 years period

Failure Absent of regular work at time of interview because

of back pain and no attempt to return to work yet

424 J Occup Rehabil (2011) 21:421–430

123

Author's personal copy

protective of ‘‘Failure after attempt’’ (OR: 0.28, 95% CI:

0.14–0.57) and of ‘‘Failure’’ (OR: 0.19, 95% CI:

0.07–0.48). A moderate level of SERW was protective for

‘‘Failure’’ (OR: 0.51, 95% CI: 0.26–1.00) (Table 5).

Discussion

This study analyzed simultaneously the role of self-efficacy

and health locus of control beliefs to explain occupational

outcome among workers with non specific back pain.

Based on Bandura’s self-efficacy theory, we expected that

SERW would be a more powerful cognitive determinant of

RWGH than HLC among these workers. The results of this

study support this hypothesis. Although bivariate analyses

yielded significant associations between external HLC

(PHLC and CHLC) and RWGH at 2 years, HLC did not

emerge as a significant determinant of RWGH in multi-

variate analyses when entered simultaneously with SERW.

These results suggest that, among this specific population,

beliefs about one’s ability to return to work is a better

indicator than one’s perceived control over health to

explain RWGH at 2 years.

To our knowledge, no previous work has looked at these

control-related constructs simultaneously in relation to

occupational outcome among patient with non specific

back pain. Therefore, comparison of the present results

with others is not possible. However, a review of studies

that have looked at these constructs separately to explain

work disability among back pain patients revealed note-

worthy information. Haldorsen et al. [33] conducted a

1 year follow-up study of back pain patients treated with a

Table 2 Selected baseline

characteristics of participants

(n = 1,007)

SERW self-efficacy for return to

work, FABQ fear-avoidance

beliefs questionnaire,

IHLC internal health locus of

control, PHLC powerful others

health locus of control,

CHLC chance health locus of

controla Single, separated, divorced

or widowedb CND $ = 0.93 US $ = 0.64

€ (Aug 4 2009)c Measured with a Numerical

Rating Scale (0–10)

Variables [Missing values] Mean (SD) n (%)

Age (years) 38.7 (10.6)

Gender

Male 589 (58.5)

Marital status [24]

Married or living as married 691 (70.3)

Living alonea 292 (29.7)

Education completed [24]

Primary 186 (18.9)

Secondary 327 (33.3)

College 305 (31.0)

University 165 (16.8)

Annual personal income (CND $) [77]b 19,249 (12,773)

Duration of work in same job (years) [19] 7.5 (8.3)

Work schedule [15]

Full-time 796 (80.2)

Part-time 196 (19.8)

Type of episode [5]

First-time 222 (22.2)

Recurrent 458 (45.7)

Persistent 322 (32.1)

Average back pain in past 6 months [49]c 5.2 (2.1)

Self-reported health status [17]

Excellent or very good 494 (49.9)

Good 312 (31.5)

Fair 135 (13.6)

Poor 49 (5.0)

SERW (0–100) [89] 49.5 (23.3)

FABQ activity (0–6) [12] 3.9 (1.8)

Modified FABQ work (0–6) [33] 3.7 (1.9)

IHLC (0–36) [24] 26.5 (5.0)

PHLC (0–36) [23] 20.2 (6.5)

CHLC (0–36) [24] 20.1 (6.2)

J Occup Rehabil (2011) 21:421–430 425

123

Author's personal copy

light mobilization program. They found that low scores on

the internal HLC scale were associated with non-return to

work at the 1 year follow-up. This association was still

statistically significant at the 5 year follow-up. These

results are inconsistent with those of the current study,

where internal HLC was not associated with RWGH in

bivariate analyses [33]. A second analysis of these data was

carried by Hagen et al. [34]. This time, both the control and

the intervention groups were considered simultaneously in

the analyses. Chance HLC was associated with non-return

to work at the 1 year follow-up (OR: 2.3, 95% CI: 1.2–4.7).

In another study by Harkapaa and her colleagues, no sig-

nificant associations were found between HLC and early

retirement in patients with chronic low back pain [23].

Several studies on back pain have found an association

between a high level of self-efficacy and various favour-

able outcomes [7, 12, 16, 35]. However, very few studies

have looked at the role of self-efficacy to explain occupa-

tional outcome among patients with back pain. Although

not defined as self-efficacy in their study, Reiso and his

colleagues [36] studied a variable called ‘‘self-assessed

work ability’’ to predict time until return to work among

patients with back disorders. Subjects were asked to assess

their work ability (5-point scale) by answering the fol-

lowing question: ‘‘To what degree does your back disorder

reduce your ability to perform your ordinary work today?’’

This variable was associated with a longer period until

return to work (Hazard Ratio: 0.43, 95% CI: 0.25–0.73). In

a recent study, Lotters et al. [37] have looked at the

prognostic value of depressive symptoms, fear-avoidance

beliefs and self-efficacy for duration of lost-time benefits in

workers with musculoskeletal disorders (58% back and

42% neck and upper extremities). Although self-efficacy

was measured with a different instrument, the construct

used was similar to SERW used in the current study. No

significant associations were found between this construct

and the number of days on benefit among this specific

population. This finding seems to be inconsistent with the

results obtained in the current study. However, one must

consider that the populations under study as well as the

outcomes measured in each of the study were different. In

addition, another possible explanation for the differences

observed is that in the present study it had been decided

a priori not to adjust for depressive symptoms because of

the probable intermediate role of this variable between self-

efficacy and occupational outcome. The inclusion of

depressive symptoms in the Lotter et al. regression model

may have reduced the importance of self-efficacy to

explain duration of lost-time benefits.

The results of the current study provide support for the

application of the theoretical framework of self-efficacy to

low back disorders. According to Bandura, self-efficacy

determines whether a person attempts behaviours or tasks,

how long they will persist in the face of obstacles and the

level of success eventually achieved [10]. Workers who

suffer from back pain are confronted with a multitude of

factors that have the potential to affect their return to

previous work activities: fluctuating level of pain, conflicts

with employer, fear that work will exacerbate their back

condition and so on. These stressors can generate a great

deal of anxiety, especially if the worker perceives having

no control over them. In light of the self-efficacy theory, an

individual with a low sense of self-efficacy would not even

attempt to return to work or if he does he would be more at

risk to give up rapidly if confronted with difficulties. The

findings of this study support this theory since high levels

of self-efficacy were protective for the ‘‘Failure after

attempt’’ and ‘‘Failure’’ categories of RWGH.

The value of a scientific theory is judged not only by its

explanatory or predictive power but also by its capacity to

provide some possible mechanisms underlying the observed

phenomena. Bandura’s self-efficacy theory provides three

possible mechanisms by which a high sense of self-efficacy

could influence occupational outcome among workers with

back pain. Self-efficacy beliefs could influence occupa-

tional outcome by their effects on (1) anxiety and its

Table 3 Spearman’s correlation coefficients between control belief constructs and other baseline measures

SERW IHLC PHLC CHLC FABA FABW VAS

SERW 1.00** 0.04 -0.11** -0.01 -0.31** -0.38** -0.19**

IHLC 1.00** 0.15** 0.12** -0.08* 0.01 -0.04

PHLC 1.00** 0.33** 0.15** 0.10* 0.20**

CHLC 1.00** 0.02 0.04 0.05

FABA 1.00** 0.34** 0.23**

FABW 1.00** 0.19**

VAS 1.00**

SERW self-efficacy for return to work, IHLC internal health locus of control, PHLC powerful others health locus of control, CHLC chance

health locus of control, FABA fear-avoidance beliefs-activity, modified FABW fear-avoidance beliefs-work, VAS average pain intensity of the

past 6 months

* P \ 0.05; ** P \ 0.01

426 J Occup Rehabil (2011) 21:421–430

123

Author's personal copy

concomitant physiological arousal, (2) depression, and (3)

the use of maladaptive coping strategies [10]. A large body

of research indicates that these variables are involved in the

transition from acute to chronic disability among patients

with back pain problems [25, 38–43]. Furthermore, in the

pain literature, several studies have confirmed the role of

self-efficacy as a determinant of anxiety, depression and the

use of specific coping mechanisms [13, 14, 44, 45]. How-

ever, these pathways need to be further explored among

workers with non specific back pain.

The results of the current study should be interpreted in

light of its limitations. Some differences were noted

Table 4 Bivariate associations of baseline control beliefs variables with return to work in good health at 2 year (n = 860)

Crude odds ratios (95% CI)

Variables Success (n = 487)a Partial success (n = 208) Failure after attempt (n = 101) Failure (n = 64)

SERW (tertiles)

High (0–41.25) 0.72 (0.47–1.09) 0.18 (0.09–0.36)* 0.11 (0.04–0.27)*

Moderate (41.26–60.63) 0.76 (0.49–1.16) 0.61 (0.37–1.00)* 0.35 (0.19–0.67)*

Low (60.64–100) 1.00 1.00 1.00

IHLC (median)

High (28–36) 0.98 (0.71–1.35) 1.11 (0.72–1.72) 1.32 (0.78–2.25)

Low (0–27) 1.00 1.00 1.00

PHLC (median)

High (21–36) 1.18 (0.85–1.63) 1.84 (1.19–2.87)* 1.87 (1.10–3.19)*

Low (6–20) 1.00 1.00 1.00

CHLC (median)

High (21–36) 1.17 (1.25–2.42) 1.29 (0.84–2.00) 1.89 (1.10–3.27)*

Low (6–20) 1.00 1.00 1.00

SERW self-efficacy for return to work, IHLC internal health locus of control, PHLC powerful others health locus of control, CHLC chance

health locus of control

* Significant at the 5% levela Reference category—the other categories are compared to the reference category in a single model, but one at a time

Table 5 Results of multivariate analyses of baseline control beliefs variables with return to work in good health at 2 year (n = 860)

Adjusted odds ratios (95% CI)

Variables Success (n = 487)a Partial success (n = 208) Failure after attempt (n = 101) Failure (n = 64)

SERW (tertiles)

High 0.83 (0.52–1.32) 0.28 (0.14–0.57)* 0.19 (0.07–0.48)*

Moderate 0.95 (0.60–1.50) 0.82 (0.47–1.40)* 0.51 (0.26–1.00)*

Low 1.00 1.00 1.00

IHLC (median)

High 1.05 (0.73–1.50) 1.12 (0.69–1.82) 1.58 (0.85–2.92)

Low 1.00 1.00 1.00

PHLC (median)

High 0.79 (0.54–1.15) 1.38 (0.83–2.31) 1.31 (0.69–2.47)

Low 1.00 1.00 1.00

CHLC (median)

High 1.20 (0.83–1.73) 1.16 (0.70–1.90) 1.37 (0.73–2.57)

Low 1.00 1.00 1.00

All odds ratios have been adjusted for age, gender, average pain over the past 6 months and fear avoidance beliefs about work

SERW self-efficacy for return to work, IHLC internal health locus of control, PHLC powerful others health locus of control, CHLC chance

health locus of control

* Significant at the 5% levela Reference category—the other categories are compared to the reference category in a single model, but one at a time

J Occup Rehabil (2011) 21:421–430 427

123

Author's personal copy

between the subjects who completed the study and those

who did not, which suggests that a selection bias may have

been introduced. Since the participants who were lost to

follow-up had lower levels of self-efficacy, we believe that

the effect of this bias would have been to underestimate the

strength on the relation observed between SERW and

RWGH.

Another possible source of bias is related to the rela-

tively high number of missing data for the variable SERW

measured at the baseline interview (89 missing data for

1,007 subjects, which represents 8.8%). An analysis of

these missing data revealed a uniform distribution across

the different categories of RWGH at 2 year. Thus, the

effect of missing data on the association observed is likely

to be negligible.

SERW was measured with an instrument recently

developed by Dionne and Nouwen [27] according to

Bandura’s guidelines. Although this questionnaire has

demonstrated good construct validity, its complete psy-

chometric properties still need to be addressed.

Given the negatively skewed distribution observed for

one of the HLC subscale, the predictor variables were

categorized, which may lead to overfitting [46]. However,

as HLC was not significantly associated with RWGH in

multivariate analyses, the probability of getting overly

optimistic results with this variable was deemed negligible.

Moreover, with a correlation of 0.33 between chance and

powerful others HLC, we would have increased the prob-

ability of type 1 error in our results. Therefore, we believe

that statistical overfitting is not a serious concern in the

present study.

We examined the effects of several potential con-

founders. Despite this effort, residual confounding is

always possible, considering the numerous variables

involved in back pain disability.

A final limitation of the current study relates to the

generalization of our findings. Although the evolution of

return to work seen in our study is a bit more complicated

to compare with the occupational outcome of previous

studies because of our definition of ‘‘return to work in good

health’’, our results as to the frequency of return to work at

2 years (81%) are coherent with those of several other

studies [47–53]. Also, the subjects of this study were

workers who sought medical consultation in first line care

settings for back pain with occupational disruptions.

Compared to the general population of Quebec’s workers,

participants of the study were somewhat less educated [32].

While access to care may differ according to socio-eco-

nomic status (SES) [54, 55], we believe the impact of this

difference on our results to be limited because people with

low SES (well represented in our study) tend to consult

less, and because of universal access to health care in

Canada. In any case, workers who decide not to consult or

to consult in a different setting may have different char-

acteristics. The interpretation of our findings must therefore

be limited to similar populations.

Despite the aforementioned limitations, this study pro-

vides an important insight into some of the cognitive

variables involved in workers’ adjustment to back pain. It

underscores the importance to distinguish conceptually

control beliefs. As predicted by Bandura’s theory, self-

efficacy was found to be more important than health locus

of control to explain RWGH. Because self-efficacy is

potentially modifiable, intervention focusing on enhancing

self-efficacy may provide an innovative approach for the

management of patients with back pain. Future studies are

needed to determine whether interventions targeting this

modifiable determinant can actually decrease work dis-

ability and other unfavourable outcomes among this

population.

Acknowledgments The authors thank all the study participants,

the staff of the Population Health Research Unit, and all the

research assistants who worked on this study. Thanks also to the

following physicians for their help with the recruitment of subjects:

Dr. Stephane Bergeron, Dr. Alexandra Dansereau, Dr. Georges

Dufresne, Dr. Louis Larue, Dr. Natalie Le Sage, Dr. Jean Maziade,

and Dr. Jean Ouellet. Special thanks to Isabelle Larocque, MSc and

Julie Soucy, PhD, for the coordination of the study and Eric De-

mers, MSc, for statistical analyses. This study was supported by a

grant (#97-061) from the Quebec Institute for Occupational Safety

and Health (Institut de recherche Robert-Sauve en sante et en

securite du travail du Quebec – IRSST) to CE Dionne, R Bourb-

onnais, P Fremont, M Rossignol and S Stock. IRSST did not

interfere in any way in the scientific and publication processes. CE

Dionne is a Senior Quebec Health Research Fund (FRSQ) Scholar.

This article reports part of the Master’s thesis in Community Health

of S Richard conducted under the supervision of CE Dionne, and

has benefited from a collaboration with A Nouwen. Both authors are

guarantors for the paper.

Appendix: Self-Efficacy for Return to Work

Questionnaire

Although most individuals on sick leave want to work,

some situations make return to work difficult. In order to

better understand what kind of context makes return to

work difficult for individuals with back pain, I will read

you a list of specific situations. For each statement, I will

ask you to rate how confident you are that you can do your

work at present when that specific situation is present.

Please rate your confidence in percentage from 0 to 100

where 0% = not at all confident, 50% = moderately con-

fident and 100% = completely confident.

1. How confident are you to do your job:

When you have a lot of pain

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

428 J Occup Rehabil (2011) 21:421–430

123

Author's personal copy

2. How confident are you to do your job:

When your pain level is low

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

3. How confident are you to do your job:

Without the risk of aggravating your existing back pain

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

4. How confident are you to do your job:

When your employer shows no or minimal effort to

adapt your working conditions to your back pain

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

5. How confident are you to do your job:

When your employer lacks understanding as to your

back pain

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

6. How confident are you to do your job:

When tension is present at work

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

7. How confident are you to do your job:

When you must work under pressure

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

8. How confident are you to do your job:

When you have difficulty commuting to work because

of your pain

0% = not at all confident, 50% = moderately confi-

dent and 100% = completely confident

References

1. Rapoport J, Jacobs P, Bell NR, Klarenbach S. Refining the

measurement of the economic burden of chronic diseases in

Canada. Chronic Dis Can. 2004;25(1):13–21.

2. Waddell G, Aylward M, Sawney P. Back pain, incapacity for

work and social security benefits: an international literature

review and analysis. London: Royal Society of Medecine Press;

2002.

3. Hashemi L, Webster BS, Clancy EA, Volinn E. Length of dis-

ability and cost of workers’ compensation low back pain claims.

J Occup Environ Med. 1997;39(10):937–45.

4. Deyo RA, Weinstein JN. Low back pain. N Engl J Med. 2001;

344(5):363–70.

5. Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A

fear-avoidance beliefs questionnaire (FABQ) and the role of fear-

avoidance beliefs in chronic low back pain and disability. Pain.

1993;52(2):157–68.

6. Boureau F. Pratique du traitement de la douleur. Paris: Doin;

1988.

7. Lackner JM, Carosella AM. The relative influence of perceived

pain control, anxiety, and functional self efficacy on spinal

function among patients with chronic low back pain. Spine (Phila

Pa 1976). 1999;24(21):2254–60; discussion 2260–61.

8. Schiaffino KM, Revenson TA, Gibofsky A. Assessing the impact

of self-efficacy beliefs on adaptation to rheumatoid arthritis.

Arthritis Care Res. 1991;4(4):150–7.

9. Feuerstein M, Beattie P. Biobehavioral factors affecting pain and

disability in low back pain: mechanisms and assessment. Phys

Ther. 1995;75(4):267–80.

10. Bandura A. Self-efficacy: the exercise of control. New York:

Freeman; 1997.

11. Reid MC, Williams CS, Gill TM. The relationship between psy-

chological factors and disabling musculoskeletal pain in commu-

nity-dwelling older persons. J Am Geriatr Soc. 2003;51(8):

1092–8.

12. Lackner JM, Carosella AM, Feuerstein M. Pain expectancies,

pain, and functional self-efficacy expectancies as determinants of

disability in patients with chronic low back disorders. J Consult

Clin Psychol. 1996;64(1):212–20.

13. Turner JA, Ersek M, Kemp C. Self-efficacy for managing pain is

associated with disability, depression, and pain coping among

retirement community residents with chronic pain. J Pain. 2005;

6(7):471–9.

14. Brister H, Turner JA, Aaron LA, Mancl L. Self-efficacy is

associated with pain, functioning, and coping in patients with

chronic temporomandibular disorder pain. J Orofac Pain. 2006;

20(2):115–24.

15. Woby SR, Watson PJ, Roach NK, Urmston M. Adjustment to

chronic low back pain—the relative influence of fear-avoidance

beliefs, catastrophizing, and appraisals of control. Behav Res

Ther. 2004;42(7):761–74.

16. Council JR, Ahern DK, Follick MJ, Kline CL. Expectancies and

functional impairment in chronic low back pain. Pain. 1988;

33(3):323–31.

17. Keefe FJ, Lefebvre JC, Maixner W, Salley AN Jr, Caldwell DS.

Self-efficacy for arthritis pain: relationship to perception of

thermal laboratory pain stimuli. Arthritis Care Res. 1997;10(3):

177–84.

18. Luszczynska A, Schwarzer R. Multidimensional health locus of

control: comments on the construct and its measurement. J Health

Psychol. 2005;10(5):633–42.

19. Wallston KA, Wallston BS, DeVellis R. Development of the

multidimensional health locus of control (MHLC) scales. Health

Educ Monogr. 1978;6(2):160–70.

20. Burker EJ, Evon DM, Galanko J, Egan T. Health locus of control

predicts survival after lung transplant. J Health Psychol. 2005;

10(5):695–704.

21. Pucheu S, Consoli SM, D’Auzac C, Francais P, Issad B. Do

health causal attributions and coping strategies act as moderators

of quality of life in peritoneal dialysis patients? J Psychosom Res.

2004;56(3):317–22.

22. Nyland J, Johnson DL, Caborn DN, Brindle T. Internal health

status belief and lower perceived functional deficit are related

among anterior cruciate ligament-deficient patients. Arthroscopy.

2002;18(5):515–8.

23. Harkapaa K. Psychosocial factors as predictors for early retire-

ment in patients with chronic low back pain. J Psychosom Res.

1992;36(6):553–9.

24. Crisson JE, Keefe FJ. The relationship of locus of control to pain

coping strategies and psychological distress in chronic pain

patients. Pain. 1988;35(2):147–54.

25. Dionne CE. Psychological distress confirmed as predictor of

long-term back-related functional limitations in primary care

settings. J Clin Epidemiol. 2005;58(7):714–8.

26. Cvengros JA, Christensen AJ, Lawton WJ. Health locus of con-

trol and depression in chronic kidney disease: a dynamic per-

spective. J Health Psychol. 2005;10(5):677–86.

J Occup Rehabil (2011) 21:421–430 429

123

Author's personal copy

27. Dionne CE, Bourbonnais R, Fremont P, Rossignol M, Stock SR,

Nouwen A, et al. Determinants of ‘‘return to work in good

health’’ among workers with back pain who consult in primary

care settings: a 2-year prospective study. Eur Spine J. 2007;16(5):

641–55.

28. Jensen MP, Turner JA, Romano JM. What is the maximum

number of levels needed in pain intensity measurement? Pain.

1994;58(3):387–92.

29. Talbot F, Nouwen A, Gauthier J. Is health locus of control a

3-factor or a 2-factor construct? J Clin Psychol. 1996;52(5):

559–68.

30. SAS Institute Inc. The SAS system for Sun OS. Cary, NC: SAS

Institute Inc; 2000.

31. Rothman KJ, Greenland SE. Modern epidemiology. 2nd ed.

Philadelphia: Lippincott Williams and Wilkins; 1998.

32. Daveluy C, Pica L, Audet N, et al. Enquete sociale et de sante

1998. Quebec: Institut de la statistique du Quebec; 2000.

33. Haldorsen EM, Indahl A, Ursin H. Patients with low back pain

not returning to work. A 12-month follow-up study. Spine (Phila

Pa 1976). 1998;23(11):1202–7; discussion 1208.

34. Hagen EM, Svensen E, Eriksen HR. Predictors and modifiers of

treatment effect influencing sick leave in subacute low back pain

patients. Spine (Phila Pa 1976). 2005;30(24):2717–23.

35. Kaplan GM, Wurtele SK, Gillis D. Maximal effort during func-

tional capacity evaluations: an examination of psychological

factors. Arch Phys Med Rehabil. 1996;77(2):161–4.

36. Reiso H, Nygard JF, Jorgensen GS, Holanger R, Soldal D,

Bruusgaard D. Back to work: predictors of return to work among

patients with back disorders certified as sick: a two-year follow-

up study. Spine (Phila Pa 1976). 2003;28(13):1468–73; discus-

sion 1473–1474.

37. Lotters F, Franche RL, Hogg-Johnson S, Burdorf A, Pole JD. The

prognostic value of depressive symptoms, fear-avoidance, and

self-efficacy for duration of lost-time benefits in workers with

musculoskeletal disorders. Occup Environ Med. 2006;63(12):

794–801.

38. Fransen M, Woodward M, Norton R, Coggan C, Dawe M,

Sheridan N. Risk factors associated with the transition from acute

to chronic occupational back pain. Spine (Phila Pa 1976).

2002;27(1):92–8.

39. Oleinick A, Gluck JV, Guire K. Factors affecting first return to

work following a compensable occupational back injury. Am J

Ind Med. 1996;30(5):540–55.

40. Pincus T, Burton AK, Vogel S, Field AP. A systematic review of

psychological factors as predictors of chronicity/disability in

prospective cohorts of low back pain. Spine (Phila Pa 1976).

2002;27(5):E109–20.

41. Burton AK, Tillotson KM, Main CJ, Hollis S. Psychosocial

predictors of outcome in acute and subchronic low back trouble.

Spine (Phila Pa 1976). 1995;20(6):722–8.

42. Sullivan MJ, Stanish W, Waite H, Sullivan M, Tripp DA. Ca-

tastrophizing, pain, and disability in patients with soft-tissue

injuries. Pain. 1998;77(3):253–60.

43. Thomas E, Silman AJ, Croft PR, Papageorgiou AC, Jayson MI,

Macfarlane GJ. Predicting who develops chronic low back pain in

primary care: a prospective study. BMJ. 1999;318(7199):1662–7.

44. Lombardo ER, Tan G, Jensen MP, Anderson KO. Anger man-

agement style and associations with self-efficacy and pain in male

veterans. J Pain. 2005;6(11):765–70.

45. Jensen MP, Turner JA, Romano JM. Self-efficacy and outcome

expectancies: relationship to chronic pain coping strategies and

adjustment. Pain. 1991;44(3):263–9.

46. Babyak MA. What you see may not be what you get: a brief,

nontechnical introduction to overfitting in regression-type mod-

els. Psychosom Med. 2004;66(3):411–21.

47. Von Korff M, Saunders K. The course of back pain in primary

care. Spine. 1996;21(24):2833–7; discussion 2838–2839.

48. Croft PR, Macfarlane GJ, Papageorgiou AC, Thomas E, Silman

AJ. Outcome of low back pain in general practice: a prospective

study. BMJ. 1998;316(7141):1356–9.

49. Macfarlane GJ, Thomas E, Croft PR, Papageorgiou AC, Jayson

MI, Silman AJ. Predictors of early improvement in low back pain

amongst consulters to general practice: the influence of pre-

morbid and episode-related factors. Pain. 1999;80(1–2):113–9.

50. Miedema HS, Chorus AM, Wevers CW, van der Linden S.

Chronicity of back problems during working life. Spine. 1998;

23(18):2021–8; discussion 2028–2029.

51. Schiottz-Christensen B, Nielsen GL, Hansen VK, Schodt T,

Sorensen HT, Olesen F. Long-term prognosis of acute low back

pain in patients seen in general practice: a 1-year prospective

follow-up study. Fam Pract. 1999;16(3):223–32.

52. Vingard E, Mortimer M, Wiktorin C, Pernold RPTG, Fredriksson

K, Nemeth G, et al. Seeking care for low back pain in the general

population: a two-year follow-up study: results from the MUSIC-

Norrtalje Study. Spine. 2002;27(19):2159–65.

53. Rossignol M, Suissa S, Abenhaim L. The evolution of compen-

sated occupational spinal injuries. A three-year follow-up study.

Spine. 1992;17(9):1043–7.

54. Lasser KE, Himmelstein DU, Woolhandler S. Access to care,

health status, and health disparities in the United States and

Canada: results of a cross-national population-based survey. Am J

Public Health. 2006;96(7):1300–7.

55. Baltzan MA. Access to health care, socioeconomic status, and

health. Ann Intern Med. 1999;130(5):452–3.

430 J Occup Rehabil (2011) 21:421–430

123

Author's personal copy