PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Karolinska Institute Library]On: 30 March 2009Access details: Access Details: [subscription number 909766720]Publisher Informa HealthcareInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Disability & RehabilitationPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713723807

What factors predict full or partial return to work among sickness absenteeswith spinal pain participating in rehabilitation?Britt Elfving ab; Malin Åsell c; Annina Ropponen ad; Kristina Alexanderson a

a Karolinska Institutet, Section of Personal Injury Prevention, Department of Clinical Neuroscience,Stockholm, Sweden b Division of Physiotherapy, Department of Neurobiology, Care Sciences and Society,Karolinska Institutet, Huddinge, Sweden c Centre for Musculoskeletal Research, University of Gävle, Umea,Sweden d Institute of Biomedicine, Department of Physiology/Ergonomics, University of Kuopio, Kuopio,Finland

First Published on: 17 March 2009

To cite this Article Elfving, Britt, Åsell, Malin, Ropponen, Annina and Alexanderson, Kristina(2009)'What factors predict full or partialreturn to work among sickness absentees with spinal pain participating in rehabilitation?',Disability & Rehabilitation,

To link to this Article: DOI: 10.1080/09638280802572965

URL: http://dx.doi.org/10.1080/09638280802572965

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

RESEARCH ARTICLE

What factors predict full or partial return to work among sicknessabsentees with spinal pain participating in rehabilitation?

BRITT ELFVING1,2, MALIN ASELL3, ANNINA ROPPONEN1,4 &

KRISTINA ALEXANDERSON1

1Section of Personal Injury Prevention, Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden,2Division of Physiotherapy, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet,

Huddinge, Sweden, 3Centre for Musculoskeletal Research, University of Gavle, Umea, Sweden, and 4Institute of

Biomedicine, Department of Physiology/Ergonomics, University of Kuopio, Kuopio, Finland

Accepted October 2008

AbstractPurpose. To identify the factors that predict full or partial return to work among long-term (�90 days) sickness absenteesdue to spinal pain who begin a multidisciplinary rehabilitation programme.Method. In a prospective cohort study, 312 patients with neck, thoracic and/or lumbar pain, aged 20–64, participated in a 4-week multidisciplinary rehabilitation programme in Sweden. Questionnaire data at inclusion were used. Factors included inlogistic regressions were as follows: age, gender, type of work, pain location, pain intensity (visual analogue scale), activitylimitations [Disability Rating Index (DRI)], health-related quality of life (SF-36), pain-related fear of movement (TampaScale of Kinesiophobia), motivation (Self Motivation Inventory), sickness absence at baseline and number of sick-leave daysduring the previous 2 years. Outcome factor was increased versus not increased working time at follow-up 6 months later.Results. Most patients (68%) reported two or three pain locations. At baseline, 56% were full-time sickness absent and 23%at follow-up; 61% had increased their working time. Predictors for increased working time were age below 40 years, lowactivity limitation (DRI5 50), low SF-36 bodily pain (430) and high SF-36 social functioning (460). Number of sick-leavedays during the previous 2 years (md 360; range 90–730) had no influence.Conclusions. Even patients with long previous sick leave can increase working time after a multidisciplinary rehabilitationprogramme, especially if they are younger, have lower levels of activity limitations and pain and better social functioning. Toinclude information on part-time work is useful when evaluating work ability following rehabilitation programmes.

Keywords: Sick leave, sickness absence, return to work, low back pain, neck pain, predictor variables, activity, participation,health-related quality of life

Background

Low back pain is very common in the population, is

often recurrent, and a pathological diagnosis can

seldom be established [1,2]. Instead, for non-specific

spinal pain, a symptom diagnosis is usually based on

the location of the pain, e.g. lower back or neck, and

on the duration of pain, e.g. chronic pain 43 months

[3,4]. In Sweden, as in many other countries, neck

and low back pain are the most common causes for

sick leave and disability pension [4,5]. With pro-

longed duration of disability and sick leave, return to

work (RTW) is more unlikely [3]. For example, after

6 months of sick leave, fewer than 50% will RTW,

and after 2 years of sick leave, the probability for

RTW is very small [6].

Costs associated with spinal pain are very high in

industrialised countries, especially indirect costs for

sick leave [7]. Therefore, rehabilitation and RTW as

soon as possible would be of high value from social

and financial points of view. In the last decades, the

so-called multidisciplinary rehabilitation pro-

grammes have been advocated to maintain work-

ability for those with chronic symptoms or to

Correspondence: Britt Elfving, Division of Physiotherapy 23100, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, 141 83

Huddinge, Sweden. E-mail: Britt.Elfving@ki.se

Disability and Rehabilitation, 2009; 1–10, iFirst article

ISSN 0963-8288 print/ISSN 1464-5165 online ª 2009 Informa Healthcare USA, Inc.

DOI: 10.1080/09638280802572965

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

promote RTW for sickness absentees. Such pro-

grammes include medical, physical, vocational and

behavioural components [3]. Not only better func-

tion but also RTW can be considered as the aim of

multidisciplinary rehabilitation of patients with

spinal pain, so as to maintain work participation,

income, quality of life and social contacts [8,9].

There is strong evidence that intensive multidisci-

plinary interventions do reduce pain and improve

function [10]. However, RTW would be a more

optimal outcome considering economic and social

effects of such programmes. Moderate scientific

evidence for RTW and reduced sickness absence

for those with chronic non-specific low back pain is

reported for brief educational interventions and

strong evidence for intensive multidisciplinary inter-

ventions [3] as well as for physical conditioning

programs that include a cognitive-behavioural ap-

proach [11]. However, there is a lack of studies on

whether specific patient groups could benefit more

from these programmes than others [10].

As regards psychosocial factors in low back pain,

pain-related fear of movement has been associated

with avoidance behaviour and increased disability.

The importance of pain intensity for the develop-

ment of avoidance behaviour has also been pointed

out recently [12]. Fear-avoidance beliefs may con-

tribute to the prediction of improvement after

multidisciplinary rehabilitation [13] but its role in

predicting RTW after multidisciplinary rehabilitation

is still not clear. Moreover, motivation is one of the

psychosocial factors that might influence RTW after

interventions among sickness absentees with low

back pain [14].

One problem in sickness absence research so far is

a large variety of the measures used for quantifying

RTW and sick leave, leading to difficulties in

comparing results between different studies [4].

Some studies use full RTW as outcome [15,16]

and some studies use return to same work [17,18].

However, the results may differ considerably if for

example also partial RTW is included. In line with

this, Braathen et al. [19] used work-related activity as

outcome, including being present at work, having

adjusted work tasks, new work, work training or

graded work activity, whereas education and un-

employment were not included in their measure

[19]. Other studies have also included part-time

work and work-related activities in their measure of

RTW [20–22]. Furthermore, a lower degree of

sickness absence has been used as an outcome [23].

In the literature of rehabilitation studies, factors

included as potential predictors vary considerably

between different studies. No core set of predictors

has been identified, except that the individual’s own

expectations of recovery is an important predictor for

RTW [8,14,24]. Therefore, additional research on

which factors that actually predict RTW after multi-

disciplinary rehabilitation for people with spinal pain

on long-term sickness absence is needed.

The aim of the present study was thus to gain more

knowledge about factors that predict full or partial

return to work among long-term (�90 days) sickness

absentees due to spinal pain who begin a multi-

disciplinary rehabilitation programme.

Methods

The design is a prospective cohort study (n¼ 312)

using questionnaire data at baseline and change in

working time at follow-up 6 months after a 4-week

multidisciplinary rehabilitation programme in

Sweden.

Patients and data collection

The study population was patients with spinal pain

who fulfilled a multidisciplinary rehabilitation pro-

gramme at Rygginstitutet (translated The Back

Institute). The patients had been referred to this

programme by either their Social Insurance Office,

by their employer, by an insurance company or by

their employment agency. The majority of the

patients had had recurrent sick-leave spells and

spinal pain problems since many years; usually

non-specific low back or neck pain. Some had

several years ago experienced a whiplash injury or a

spinal surgery, the latter in most cases due to disc

hernia. In addition, all referred patients had a

physician’s certificate documenting spinal pain and

stating no obstacles to this type of rehabilitation, e.g.

cardiovascular disorders. In this study, the inclusion

criteria were patients who:

1. started the rehabilitation programme at the

Rygginstitutet during the year 2005,

2. were aged 20–64,

3. were full- or part-time sickness absent due to

chronic spinal pain in the neck, thoracic and/

or lumbar spine region at the start of

rehabilitation

4. had had �90 sick-leave days during the 2 years

preceding inclusion.

Baseline data were obtained by comprehensive

questionnaires filled out at the start of the rehabilita-

tion programme. Outcome data about the work

situation were obtained from the participants at

follow-up 6 months later. The questionnaires in-

cluded demographic data and validated instruments

on health, activity limitations, fear of movement and

motivation.

2 B. Elfving et al.

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

During the inclusion period, 362 patients

fulfilled the inclusion criteria (of all the patients,

65 worked full-time and 25 had been on sick leave

less than 90 days and were thus not included). Of

the included, 44 patients denied to answer all the

questionnaires used for this study. However, those

patients did not significantly differ from the study

group in any variable, except for fewer patients

having experienced whiplash injury (there were

significantly more in the study group, 14 vs. 2%).

Data on the working situation at follow-up were

missing for six of the participants. Thus, the study

group was n¼ 312; the drop out was 14%. The

study was approved by the Regional Ethics Com-

mittee in Stockholm in 2004.

The multidisciplinary rehabilitation programme

The multidisciplinary rehabilitation was provided by

Rygginstitutet, which specialises in vocational reha-

bilitation for patients with spinal pain. Rygginstitutet

has five rehabilitation centres in different regions of

Sweden with essentially the same rehabilitation

programme. The programme has a certification

ISO 9001:2000, which means that the quality of

the organisation is assessed and approved by

independent assessors. The programme is an in-

tensive 4-week course, 8 h a day, including physical

training, education in anatomy, physiology, ergo-

nomics, pain, relaxation techniques, tests and

physical work techniques. The programme also

includes an ergonomic investigation of the patient’s

place of work [23]. Follow-up is made 6 months after

start.

Factors at baseline

The following data from the questionnaires were

used as baseline information. The model of the

International classification of functioning, disability

and health [25,26] is used in the presentation of the

factors in the method section and in the tables.

Body structures and functions.

. Previous spinal surgery (yes/no)

. Previous whiplash injury (yes/no)

. Pain location (neck, thoracic, lumbar) was

marked on a pain drawing and categorised as

number of pain locations; 1, 2 or 3.

. Pain intensity was rated on a visual analogue

scale (VAS 0–100 mm) for neck, thoracic and

lumbar pain, respectively. If a patient had more

than one pain location the highest pain level of

two or three locations was used.

Activity and participation, and health.

. Limitation in physical activities was assessed

with the Disability Rating Index (DRI), devel-

oped and tested in Sweden [27]. The question

‘How do you manage the following activities?’

was to be rated for 12 physical activities

(dressing, outdoor walks, climbing stairs, sit-

ting longer time, standing bent over a sink,

carrying a bag, making a bed, run, light work,

heavy work, lifting heavy objects, participating

in exercise/sports) on a VAS from ‘without

difficulty’ to ‘not at all’. A mean score was

calculated for the 12 activities (min 0; max

100). A high score indicates large activity

limitations.

. Health-related quality of life was assessed with

the Short Form-36 Health Survey (SF-36)

tested in Sweden for reliability and validity

[28]. The results are presented as scores for

eight subscales each including a different

number of questions: physical functioning, role

physical, bodily pain, general health, vitality,

social functioning, role emotional and mental

health. A high score means better functioning

and health or less pain (min 0; max 100).

. Type of work was classified into three cate-

gories (a, b, c) according to the first digit level

of the International Standard Classification of

Occupations (ISCO) in the following way: (a)

1–3, work requiring higher education or train-

ing, e.g. executives, physicists, civil engineers,

physicians and specialist nurses, teachers; (b)

4–5, office and service work, e.g. clerks, service

workers, nurses, engineers, shop and market

sales workers; (c) 6–9, manual work, e.g.

agricultural and fishery work, craft, construc-

tion and related trades work, plant and

machine operators, and work not requiring

special skills.

. Previous number of sick-leave days during

the 2 years before inclusion. The patient

was to contact the employer or the social

insurance office for exact information on the

number of sick-leave days during the preceding

2 years. Part-time sick-leave days were

converted to full-time days (part-time sick-

leave ratio6 number of days). Good agree-

ment has previously been shown between

self-reported and register data on sickness

absence [4,29].

. Percent of sickness absence of ordinary work-

ing time at inclusion: 25–100%. In Sweden, it

is possible to be part-time sickness absent for

25, 50 or 75% of ordinary working time. Work

was defined as any of the following types of

work-related activity: any paid work, training

Predictive factors for RTW 3

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

for work, unemployed but searching for work

or education.

Personal factors.

. Gender, age and body mass index (kg/m2).

. Pain-related fear of movement was assessed

with Tampa Scale of Kinesiophobia (TSK) in

the Swedish version, tested for reliability and

validity [30]. The 17 items were scored from 1

to 4, where 1 is ‘strongly disagree’ and 4 is

‘strongly agree’. Sum scores were calculated

(min 17; max 68) and a high score indicates

high degree of pain-related fear.

. Self motivation inventory (SMI) [31] assesses

motivation, meant as a perseverant goal-striv-

ing behaviour in this questionnaire. It consists

of 40 statements rated from 1 to 5, where 1 is

‘very unlike me’ and 5 is ‘very like me’. Some

examples are: ‘I can endure stressing tasks even

if it is physically tiresome and painful’, ‘I really

want to develop and keep self discipline’ and ‘I

am good at making decisions and stand for

them’. Sum scores were calculated (min 40;

max 200) where a high score means high

degree of motivation.

Outcome variable

Outcome variable was change in working time where

the working situation at the start of the rehabilitation

course (percent sickness absence at baseline) was

compared with the situation at follow-up (percent

sickness absence after 6 months). An increase in

percent working time meant consequently a reduc-

tion in percent sickness absence. Change was

dichotomised as either increase in working time or

as decrease or no change in working time.

Statistics

Correlations were checked with Spearman’s rank

correlation coefficient. In the logistic regressions,

factors were dichotomised according to the median.

Patients with exactly median scores were sorted into

the category with less disability. An exception to this

was previous sick-leave days and body mass index

where the dichotomisation limit was 365 days and

25 kg/m2, respectively. Age was categorised as 20–40

years/41–50 years/51–64 years. Simple logistic re-

gressions were made with increased versus decreased

or no change in working time as outcome variable to

obtain unadjusted odds ratios and 95% confidence

intervals. Then, the potential confounding effect of

previous sick-leave days was taken into account by

comparing the unadjusted and adjusted odds ratios

of the simple logistic regressions for each factor.

Also, the percent sickness absence at baseline was

checked for confounding effect in the same way, as

well as the two confounders together. Previous sick-

leave days and percent sickness absence at baseline

were selected to be tested for confounding effect due

to the fact they may influence the RTW rate [3,4,6].

Factors with p5 0.05 from the simple logistic

regressions were entered in a multiple backward

logistic regression. Remaining factors were those

with p5 0.05. Statistical analyses were made in

SPSS 15.0 for Windows (SPSS, USA).

Results

Descriptive data of the cohort of patients are shown

in Table I. The majority of patients (68%) had spinal

pain in two or three locations. During the 2 years

before the start of the rehabilitation, 16% of the

patients had been on sick leave for 90–180 days, 40%

for 181–365 days and 44% for 366–730 days.

The percentages of patients with different working

time at baseline and at follow-up are shown in

Figure 1. A detailed survey of the change in working

time is shown in Table II. At baseline, 56% of the

patients were on full-time sick leave and 23% at

follow-up. At follow-up, 61% of the patients had

increased their working time. Results from simple

logistic regressions (unadjusted) are shown in

Table III and from multiple regression (adjusted)

in Table IV. DRI (score5 50), SF-36 bodily pain

(score 430), SF-36 social functioning (score 462)

and age below 40 years remained significant as

predictors for increased working time at follow-up.

These factors remained significant after adjusting for

previous sick-leave days and for percent of sickness

absence at baseline.

Discussion

We have studied sickness absentees with spinal pain

undergoing a multidisciplinary rehabilitation pro-

gramme and the focus was on baseline factors that

predicted increased working time 6 months later. We

found four such factors: less activity limitations, less

bodily pain, better social functioning and age below

40 years. The proportion of patients on full-time sick

leave had decreased from 56 to 23%. Furthermore,

we recommend that more detailed measures of RTW

than full-time RTW is used as outcome measure in

this type of studies.

Similar predictors to our findings, as well as

not similar, have been reported in previous long-

itudinal studies evaluating predictors for RTW after

4 B. Elfving et al.

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

multidisciplinary rehabilitation, such as age, gender,

pain, sick-leave days, disability, general health,

vitality, locus of control and pain-related fear of

movement [15–17,23,32,33].

In future studies, we recommend that also factors

such as availability of vacant jobs and modified

working tasks are included when reporting RTW

outcomes, because RTW can be affected by this and

by the labour market situation [3] and presumably

by political decisions. Therefore, in our outcome

measure, we included work-related activities such

as job searching. The ability to work, rather than

employment, is then considered, e.g. being

unemployed but capable to perform work or in

training for work. Moreover, we assessed change in

working time as outcome, considering it to be a more

sensitive measure than return to full-time work. For

example, in this study, 61% increased their working

time, whereas only 23% returned to full-time work.

If only full RTW had been considered much of this

positive change would have remained undetected.

Our rate of increased working time (61%), i.e.

patients decreasing their degree of sickness absence,

is comparable with the result of Selander et al. [23],

who showed a decreased degree of sickness absence

for 55% of the patients, with a data set similar to

Table I. Descriptive data of the cohort of patients on sickness absence due to neck, thoracic and/or low back pain and participating in a

multidisciplinary rehabilitation programme (n¼ 312).

ICF domains/factors Missing data

Body structures and functions

Spinal surgery, n (%) 36 (11.5)

Whiplash, n (%) 43 (13.8) 1

Pain location

One pain location, n (%) 101 (32.4)

Neck, n (%) 18 (5.8)

Thoracic, n (%) 3 (1.0)

Lumbar, n (%) 80 (25.6)

Two pain locations, n (%) 95 (30.4)

Neck and thoracic, n (%) 26 (8.3)

Neck and lumbar, n (%) 43 (13.8)

Thoracic and lumbar, n (%) 26 (8.3)

All three locations, n (%) 116 (37.2)

Pain intensity (VAS; mm) median (quartile; range) 63 (47–76; 0–100) 1

Activity and participation, and health

Disability rating index (DRI; 0–100) median (quartile; range) 50 (40–61; 0–100) 10

SF-36* Physical functioning (0–100) median (quartile; range) 60 (45–70; 15–100) 2

SF-36 Role physical (0–100) median (quartile; range) 0 (0–25; 0–100) 2

SF-36 Bodily pain (0–100) median (quartile; range) 31 (22–41; 0–74) 2

SF-36 General health (0–100) median (quartile; range) 55 (40–70; 0–100) 2

SF-36 Vitality (0–100) median (quartile; range) 40 (25–55; 0–100) 2

SF-36 Social functioning (0–100) median (quartile; range) 63 (38–75; 0–100) 2

SF 36 Role emotional (0–100) median (quartile; range) 67 (0–100; 0–100) 2

SF 36 Mental health (0–100) median (quartile; range) 68 (52–80; 0–100) 2

Type of work

Work requiring higher education or training n (%) 79 (25.5) 2

Office and service work n (%) 97 (31.3)

Manual and transport work n (%) 134 (43.2)

Previous no. of sick-leave days{ 360 (231–574; 90–730)

Percent of sickness absence 100%, n (%){ 176 (56.4)

Percent of sickness absence 75%, n (%) 34 (10.9)

Percent of sickness absence 50%, n (%) 81 (26.0)

Percent of sickness absence 25%, n (%) 21 (6.7)

Personal factors

Gender; women, n (%) 168 (53.8)

Age (year) median (quartile; range) 43 (36–50; 22–63)

Body mass index (kg/m2) median (quartile; range) 27 (24–30; 19–47) 1

Tampa Scale of Kinesiophobia (TSK; 17–68) median (quartile; range) 39 (33–43; 19–65) 9

Self Motivation Inventory (SMI; 40–200) median (quartile; range) 156 (147–169; 96–196) 12

*Higher scores are better in SF-36.{Number of days during the 2 years before inclusion. Full-time days or part-time sick-leave ratio6number of days.{Percent of sickness absence of ordinary working time at inclusion; in Sweden it is possible to be part-time sickness absent.

Predictive factors for RTW 5

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

ours. Other studies have found that the rate of return

to same work for patients with low back pain, sick-

listed 6–12 weeks, was 67% [18] and to full-time

work 79% [16].

Many studies find that lower age predicts higher

probability of RTW [17,23,34] and our results are in

line with that. Another known factor of importance

for RTW in chronic low back pain patients on long-

term sick leave is the previous length of sickness

absence [34,35]. However, we did not find a

previous length of sick leave less than 365 days to

be associated with increased working time. This

result concurs with the findings of Selander et al.

[23]. This could be interpreted as a positive finding

in that patients also with previous sickness absence

more than 12 months can profit from such rehabi-

litation measures. Other studies on patients with

musculoskeletal pain and long previous sick leave, 11

months and 22 months, respectively, have reported

that 70% had returned to work-related activity [21],

and 63% to full-time or part-time work [20] after

1 year.

In our study, better social functioning was found

to predict increased working time after the rehabili-

tation programme. General health, vitality, internal

locus of control and age have been predictive for

RTW among patients on long-term sick leave [23].

We found that general health, vitality, social func-

tioning and mental health correlated moderately

(r¼ 0.45–0.64), which indicates the relevance of

health and social factors in predicting RTW. Lower

level of self-reported activity limitation was a

predictor for increased working time in our study

as well as for RTW in other studies [16,22,32]. We

chose the DRI questionnaire, instead of neck- or low

back pain specific questionnaires, because our cohort

included patients with neck, thoracic and lumbar

pain. The other measure of activity limitation, the

Figure 1. Distribution of patients (%) and their working ability at baseline and at follow-up after 6 months (n¼ 312).

Table II. Number of patients with different working time at baseline and the same patients distributed over different working times at the 6

month follow-up (n¼312).

At baselineAt 6-month follow-up

Number Work 0% Work 25% Work 50% Work 75% Work 100%

Work 0% 176 64 40 37 4 31

Work 25% 34 4 7 15 2 6

Work 50% 81 3 0 34 19 25

Work 75% 21 0 1 3 7 10

Work 100% 0 0 0 0 0 0

Total at follow-up 312 71 48 89 32 72

The shaded area is the increased working time. Work 0% means full-time sickness absence.

6 B. Elfving et al.

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

SF-36 physical functioning, includes relatively simi-

lar questions of activities as the DRI but was not

significant in the simple regression, neither in the

study by Selander et al. [23].

Lower back pain intensity as a predictor for RTW

has been shown in studies evaluating rehabilitation

[15,34]. However, also patients with high pain levels

can benefit from multidisciplinary rehabilitation

[18]. We used two measures for pain, pain intensity

at baseline and SF-36 bodily pain; the latter gives a

combined score for pain during the previous 4 weeks

and pain interfering with work. These two variables

were only moderately correlated (r¼ 0.43) and only

bodily pain was significant. Furthermore, we ex-

pected that those having pain in several locations

would be less likely to increase their working time

than those having pain solely in one spinal location.

We did not, however, find any association between

number of pain locations and increased working

time.

The number of men and women were about equal

in the present study. We did not find any gender

effect on increased working time. As a matter of fact,

the significant gender differences in the baseline

factors (Table I) were few: the men had lower SF-36

role physical, higher fear of movement and more

manual and transport work. However, these factors

turned out to be not significant for increased working

time. Previously, a faster rate of RTW has been

shown for women after multidisciplinary rehabilita-

tion [36] but also no effect of gender [37].

Neither did we find that the type of work

influenced increase in working time. A reason for

this absence of influence might have been our way in

classifying work. We chose to classify by means

of the International Standard Classification of

Occupations, because this method should have the

Table III. Simple logistic regressions of factors at baseline

predicting increased working time; n¼ 189 (vs. no change or

decreased working time; n¼ 123) 6 months later.

ICF domains/factors n OR 95% CI

Body functions

Pain location

One 101 1

Two 95 0.67 0.38–1.19

Three 116 0.81 0.46–1.40

Pain intensity (VAS)

0–63 157 1.47 0.93–2.32

64–100 154 1

Activity and participation, and health

Disability rating index

0–50 148 2.17 1.36–3.48

50.1–100 154 1

SF-36 Physical functioning

60–100 177 1.37 0.86–2.16

0–59 133 1

SF-36 Role physical

11–100 102 0.89 0.55–1.45

0–10 208 1

SF-36 Bodily pain

31–100 172 2.12 1.34–3.38

0–30 138 1

SF-36 General health

67–100 158 1.48 0.93–2.33

0–66 152 1

SF-36 Vitality

40–100 171 1.41 0.89–2.23

0–39 139 1

SF-36 Social functioning

63–100 175 2.27 1.43–3.62

0–62 135 1

SF 36 Role emotional

67–100 165 1.05 0.66–1.66

0–66 145 1

SF 36 Mental health

68–100 165 1.17 0.74–1.85

0–67 145 1

Type of work

Work requiring higher

education or training

79 1

Office and service work 97 1.54 0.84–2.84

Manual and transport work 134 1.18 0.67–2.07

Previous no. of sick-leave days*

�365 days 173 1.40 0.89–2.21

366–730 days 139 1

Percent of sickness absence 100%{ 176 1

Percent of sickness absence 75% 34 1.20 0.55–2.61

Percent of sickness absence 50% 81 0.68 0.40–1.16

Percent of sickness absence 25% 21 0.52 0.21–1.29

Personal factors

Gender

men 144 0.89 0.56–1.40

women 168 1

Age

20–40 years 129 2.46 1.36–4.44

41–50 years 110 1.79 0.98–3.25

51–64 years 73 1

(continued)

Table III. (Continued).

ICF domains/factors n OR 95% CI

Tampa Scale of Kinesiophobia

17–39 169 0.99 0.62–1.58

40–68 134 1

Self motivation inventory

156–200 144 1.15 0.73–1.82

40–155 156 1

Dichotomisation of the factors was done according to the median,

shown in the table for each factor. For body function and activity

and participation the category with the highest disability was used

as the reference category. Unadjusted odds ratios (OR) and 95%

confidence intervals (CI) are shown.

*Number of sick-leave days during the 2 years before inclusion.

Full-time days or part-time sick-leave ratio6number of days with

benefits.{Percent sick-leave of ordinary working time at inclusion; in

Sweden it is possible to be part-time sickness absent.

Predictive factors for RTW 7

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

advantage of being highly reproducible. However,

this way to classify might not mirror the physical

demands in the work, which should be more

important for working ability. A better measure might

be self-reports on physical work loads. Aittomaki

et al. [38] used this method and showed all the same

that occupational class (similar classification as ours)

and physical demands were associated.

Psychosocial factors that were included in this

study were motivation and pain-related fear of

movement. We did not include any question about

the patient’s own prediction regarding RTW,

because it might be expected that a patient starting

rehabilitation may consider it prudent to answer

positively. Instead, we chose the self-motivation

inventory [31], but this was not found to be a

significant predictor. Neither did we find that pain-

related fear of movement was predictive for

increased working time, contrary to what we had

expected. However, in other intervention studies

lower pain-related fear of movement has been

predictive of better work resumption [15,16,22].

The reasons for the different results may be due to

the samples used; only low back pain [16,22],

shorter previous sick listing [16] or all types of

musculoskeletal disorders [15]. Other reasons

could be the method used; Fear-Avoidance Beliefs

Questionnaire [16,22] or the outcome; full RTW

[15,16].

The cohort in this study included sickness absent

patients participating in the rehabilitation at the

Rygginstitutet due to spinal pain. Because

Rygginstitutet provides rehabilitation in four differ-

ent cities across Sweden, the cohort in this study

should represent Swedish inhabitants from various

regions of the country and from cities differing in

size. In the cohort, there were not only patients with

non-specific spinal pain but also a smaller number of

patients with previous whiplash and spinal surgery.

As the majority of the patients had a combination of

pain sites we refrained from doing further subgroup

analyses. The strength of the current study is the

large cohort size including patients with neck,

thoracic, and/or low back disorders from all over

Sweden. Another advantage is that we included data

on sick leave as long as 2 years before the start of the

rehabilitation. Thus, generalisation could be made to

patients with spinal pain (on one, two or three

locations), on long-term sick leave (3 months or

more) and participating in multidisciplinary rehabi-

litation. Other strengths are the more comprehensive

measure of RTW. A limitation of the study is the

single follow-up at 6 months. A further limitation

may be suggested, namely the absence of a control

group. However, the aim of this study was not to

analyse the effect of a multidisciplinary rehabilitation

programme compared to not going through such a

programme. There is evidence for the benefits of this

kind of programmes as mentioned in the back-

ground. Here we aimed to get more specific knowl-

edge about factors predicting increased working time

following multidisciplinary rehabilitation. For this

purpose, a prospective cohort design has more often

been used [13,15,17,21,23,32,35] than, for example,

controlled trials [19,34].

As regards clinical implications, we suggest that

the DRI and SF-36, especially the subscales bodily

pain and social functioning, might give useful

information for therapists at the start of the

rehabilitation. A subgroup with decreased function

in these factors might need some kind of extra

support. However, future research might investigate

this more closely.

We conclude that age below 40 years, low level

of self reported activity limitations, low pain and

high social functioning were factors predicting an

increase in working time after rehabilitation in this

cohort of patients, where the majority had spinal

pain in two or three locations and had been on

sick leave for 180–730 days during the 2 years

before rehabilitation. More research is needed on

how to support these patients in their rehabilita-

tion. However, gender, length of previous sick

leave, number of pain locations, type of work and

pain-related fear of movement seemed not to

influence an increase in working time following

rehabilitation. Including part-time work provides

additional information when evaluating the change

in work ability.

Table IV. Multiple logistic regression for increased working time

(vs. no change or decreased working time) for factors, which were

significant in the crude logistic regressions (n¼300) (Table III).

Factors OR* 95% CI

Disability rating index

0–50 1.93 1.15–3.23

50.1–100 1

SF-36 Bodily pain

31–100 1.70 1.00–2.91

0–30 1

SF-36 Social functioning

63–100 1.82 1.07–3.10

0–62

Age

20–40 years 3.17 1.66–6.03

41–50 years 1.72 0.91–3.27

51–64 years 1

Dichotomisation of the factors was done according to the median,

shown in the table for each factor. The category with the highest

disability was used as the reference category. Adjusted odds ratios

(OR) and 95% confidence intervals (CI) are shown.

*Adjusted for previous number of sick-leave days and percent

sickness absence at baseline.

8 B. Elfving et al.

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

Acknowledgement

The study was financed by the Centre for Health

Care Science, Karolinska Institutet, Stockholm, and

the Swedish Council for Working Life and Social

Research.

References

1. Nachemson A, Jonson E. Neck and back pain. The scientific

evidence of courses, diagnoses, and treatment. Philadelphia:

Lippincot, Williams & Wilkins; 2000.

2. Krismer M, van Tulder M. Low back pain (non-specific). Best

Pract Res Clin Rheumatol 2007;21:77–91.

3. Airaksinen O, Brox JI, Cedraschi C, Hildebrandt J,

Klaber-Moffett J, Kovacs F, Mannion AF, Reis S, Staal JB,

Ursin H, Zanoli G, COST B13 Working Group on Guidelines

for Chronic Low Back Pain. European guidelines for the

management of chronic nonspecific low back pain. Eur Spine

J 2006;15(Suppl 2):192–300.

4. Alexanderson K, Norlund AE. Sickness absence – causes,

consequences, and physicians’ sickness certification practice.

A systematic literature review by the Swedish Council on

Technology Assessment in Health Care. Scand J Public

Health 2004;32(Suppl 63):1–263.

5. Waddell G. Preventing incapacity in people with musculoske-

letal disorders. Br Med Bull 2006;77-78:55–69.

6. Waddell G. The clinical course of low back pain. In: Waddell

G, editor. The Back Pain Revolution. Edinburgh: Churchill

Livingstone; 2004.

7. Dagensis S, Caro J, Haldeman S. A systematic review of low

back pain cost of illness studies in the United states and

internationally. Spine J 2008;8:8–20.

8. Poiraudeau S, Rannou F, Revel M. Functional restoration

programs for low back pain: a systematic review. Ann Readapt

Med Phys 2007;50:425–429.

9. van Geen JW, Edelaar MJA, Janssen M, van Eijk J. The Long-

term effect of multidisciplinary back training: a systematic

review. Spine 2007;32:249–255.

10. Lundberg D, Axelsson S. Metoder for behandling av

langvarig smarta (in Swedish). Treatment of chronic pain –

a systematic literature review. Stockholm: The Swedish

Council on Technology Assessment in Health Care, SBU;

2006.

11. Schonstein E, Kenny D, Keating J, Koes B. Work condition-

ing, work hardening and functional restoration for workers

with back and neck pain. Cochrane Database Syst Rev

2003;3:CD001822.

12. Leeuw M, Goossens MEJB, Linton SJ, Crombez G, Boersma

K, Vlaeyen JWS. The Fear-avoidance model of musculo-

skeletal pain: Current state of scientific evidence. J Behav Med

2007;30:77–94.

13. van der Hulst M, Vollenbroek-Hutten M, Groothuis-

Oudshoorn K, Hermens H. Multidisciplinary rehabilitation

treatment of patients with chronic low back pain: a

prognostic model for its outcome. Clin J Pain 2008;24:

421–430.

14. Selander J, Marnetoft SU, Bergroth A, Ekholm J. Return to

work following vocational rehabilitation for neck, back and

shoulder problems: risk factors reviewed. Disabil Rehabil

2002;24:704–712.

15. Sullivan MJ, Ward LC, Tripp D, French DJ, Adams H, Stanish

WD. Secondary prevention of work disability: community-

based psychosocial intervention for musculoskeletal disorders.

J Occup Rehabil 2005;15: 377–392.

16. Storheim K, Brox JI, Holm I, Bor K. Predictors of return

to work in patients sick listed for sub-acute low back pain:

a 12-month follow-up study. J Rehabil Med 2005;37:

365–371.

17. Buchner M, Neubauer E, Zahlten-Hinguranage A,

Schiltenwolf M. Age as a predicting factor in the therapy

outcome of multidisciplinary treatment of patients with

chronic low back pain – a prospective longitudinal clinical

study in 405 patients. Clin Rheumatol 2007;26:385–392.

18. Buchner M, Neubauer E, Zahlten-Hinguranage A. The

influence of the grade of chronicity on the outcome of

multidisciplinary therapy for chronic low back pain. Spine

2007;32:3060–3066.

19. Braathen TN, Veiersted KB, Heggenes J. Improved work

ability and return to work following vocational multidisciplin-

ary rehabilitation of subjects on long-term sick leave. J Rehabil

Med 2007;39:493–499.

20. Norrefalk JR, Svensson O, Ekholm J, Borg K. Can the back-

to-work rate of patients with long-term non-malignant pain be

predicted? Int J Rehabil Res 2005;28:9–16.

21. Øyeflaten I, Hysing M, Eriksen H. Prognostic factors associated

with return to work following multidisciplinary vocational

rehabilitaion. J Rehabil Med 2008;40:548–554.

22. Heymans M, Ford J, McMeeken J, Chan A, de Vet H,

van Mechelen W. Exploring the contribution of patient-

reported and clinician based variables for the prediction of low

back work status. J Occup Rehabil 2007;17:383–397.

23. Selander J, Marnetoft S-U, Asell M. Predictors for successful

vocational rehabilitation for clients with back pain problems.

Disabil Rehabil 2007;29:215–220.

24. Kuijer W, Groothoff JW, Brouwer S, Geertzen JHB, Dijkstra

PU. Prediction of sickness absence in patients with chronic

low back pain: a systematic review. J Occup Rehabil 2006;

16:439–467.

25. World Health Organization (WHO). International Classifica-

tion of Functioning, Disability and Health, ICF. Geneva;

2001.

26. Weigl M, Cieza A. Physical disability due to musculoskeletal

conditions. Best Pract Res Clin Rheumatol 2007;21:167–190.

27. Sahlen B, Spangfort SV, Nygren AL, Nordemar R. The

disability rating index: an instrument for the assessment

of disability in clinical settings. J Clin Epidemiol 1994;47:

1423–1445.

28. Sullivan M, Karlsson J, Ware JE. The Swedish SF-36 Health

Survey–I. Evaluation of data quality, scaling assumptions,

reliability and construct validity across general populations in

Sweden. Soc Sci Med 1995;41:1349–1358.

29. Voss M, Stark S, Alfredsson L, Vingard E, Josephson M.

Comparisons of self-reported and register data on sickness

absence among public employees in Sweden. Occup Environ

Med 2008;65:61–67.

30. Lundberg MKE, Styf J, Carlsson SG. A psychometric

evaluation of the Tampa Scale for Kinesiophobia – from a

physiotherapeutic perspective. Physiother Theory Pract 2004;

20:121–133.

31. Dishman RK, Ickes W. Self-motivation and adherence to

therapeutic exercise. J Behav Med 1981;4:421–438.

32. Koopman FS, Edelaar MJA, Slikker R, Reynders K, van der

Woude LHV, Hoozemans MJM. Effectiveness of a multi-

disciplinary occupational training program for chronic low

back pain: a prospective cohort study. Am J Phys Med Rehabil

2004;83:94–103.

33. Mayer T, Gatchel RJ, Evans T. Effect of age on outcomes of

tertiary rehabilitation for chronic disabling spinal disorders.

Spine 2001;26:1378–1384.

34. Bendix AF, Bendix T, Haestrup C. Can it be predicted which

patients with chronic low back pain should be offered tertiary

rehabilitation in a functional restoration program? A search for

Predictive factors for RTW 9

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009

demographic, socioeconomic, and physical predictors. Spine

1998;23:1775–1783.

35. Pfingsten M, Hildebrandt J, Leibing E, Franz C, Saur P.

Effectiveness of a multimodal treatment program for chronic

low-back pain. Pain 1997;73:77–85.

36. Jensen IB, Bergstrom G, Ljungquist T, Bodin L. A 3-year

follow-up of a multidisciplinary rehabilitation programme for

back and neck pain. Pain 2005;115:273–283.

37. Gatchel RJ, Mayer TG, Kidner CL, McGeary DD. Are

gender, marital status or parenthood risk factors for outcome

of treatment for chronic disabling spinal disorders? J Occup

Rehabil 2005;15:191–201.

38. Aittomaki A, Lahelma E, Rahkonen O, Paivi L-A, Martikainen

P. The contribution of musculoskeletal disorders and physical

workload to socioeconomic inequalities in health. Eur J Public

Health 2006;17:145–150.

10 B. Elfving et al.

Downloaded By: [Karolinska Institute Library] At: 10:35 30 March 2009