Work-related interventions for people on long-term sick leave: a systematic review
Systematic review
|Published
The Norwegian Institute of Public Health (NIPH) was commissioned by the Norwegian Labour and Welfare Administration (NAV) to conduct a systematic review on the effect of work-related interventions on return to work for people on long-term sick leave and people at risk for long-term sick leave.
Key message
The Norwegian Institute of Public Health (NIPH) was commissioned by the Norwegian Labour and Welfare Administration (NAV) to conduct a systematic review on the effect of work-related interventions on return to work for people on long-term sick leave and people at risk for long-term sick leave.
In May 2020, we conducted a systematic search in electronic databases. We screened the references, extracted data, and analysed the included studies. We included 20 randomized controlled trials (RCTs), and 13 systematic reviews of high methodological quality. The RCTs assessed eight different interventions.
The main findings are:
- It is uncertain whether work-related interventions, compared to other interventions or usual care, makes a difference on return to work after 12 months.
- There is probably no difference between multidisciplinary rehabilitation and other active interventions in return to work after 24 months.
- It is uncertain whether work-related interventions, compared to other interventions or usual care, makes a difference on health-related outcomes.
There seems to be no difference between work-related interventions and other active interventions or usual care in return to work for people on long-term sick leave. Our conclusion is consistent with earlier systematic reviews. Further studies may change the conclusion.
Summary
Background
Long-term sick leave is a serious concern in developed countries, including Norway. The most common reasons for sick leave are musculoskeletal disorders, common mental disorders, and respiratory disorders. Yearly, about 10 billion NOK are used for different work-related measures, aiming to prevent sick leave and falling out of working life. The current IA-agreement (Agreement for a more inclusive working life 2019-2022) has two key objectives: to reduce sick leave and prevent falling out of working life. The latter relates particularly to people on long-term sick leave who might not return to work (RTW). For this group, the Norwegian Labour and Welfare Administration (NAV) and the health services offer various occupational rehabilitation interventions. However, the effects of these interventions remain uncertain. NAV commissioned the Norwegian Institute of Public Health to carry out a systematic review about the effects of work-related interventions on return to work among people on long-term sick leave or at risk of long-term sick leave.
Objective
The aim of this review was to answer the question: What is the effect of work-related interventions for people on long-term sick leave and people at risk of long-term sick leave?
Method
We conducted a systematic review of primary research. In May 2020, we conducted a systematic search in relevant electronic databases and a search for grey literature. We included randomized controlled trials (RCTs). In addition, we listed the abstracts of systematic reviews of high methodological quality. We included interventions with an active health component that aimed to promote RTW or to remain in work. Eligible outcomes were sick leave, symptom reduction, physical/social/cognitive function and cost-effectiveness. Two researchers screened all identified titles and abstracts, and relevant full texts, and they assessed the risk of bias. When possible, we performed meta-analyses, otherwise we presented the results narratively. We used the GRADE approach (Grading of Recommendations Assessment, Development, and Evaluation) to assess the certainty of the evidence. With GRADE we provide standardized expressions of the degree of certainty to which the results show the ‘truth’ or the ‘actual’ effect of the intervention. There are four levels of certainty: high, moderate, low, very low. Certainty is affected by several methodological factors, such as how many studies there are. In general, the more and better studies there are, the higher the degree of certainty that the results show the ‘true’ effect of the intervention. When there is weak or limited evidence, e.g. because there are few studies or only studies with a high degree of bias, we state that the effect is uncertain.
Results
We screened 7186 abstracts and 200 full-texts. We included 20 studies (reported in 25 publications) and 13 systematic reviews of high methodological quality. The studies were conducted in Norway, Denmark, Sweden and the Netherlands. They were published between 2001 and 2020. The studies comprised 5753 participants between 16 and 65 years, with time on sick leave from five to 52 weeks. Most were on sick leave due to common mental disorders and musculoskeletal disorders. We identified eight different interventions with ten different comparisons. Most studies assessed the effect of multidisciplinary rehabilitation (10 studies). Four studies assessed cognitive therapy. The remaining six studies examined other unique interventions. The control groups received usual care or other active interventions. We summarize the key findings below.
Multidisciplinary rehabilitation vs. Usual care. We included five studies that assessed the effect of multidisciplinary rehabilitation versus usual care. We analyzed two studies narratively and performed a meta-analysis for three of the studies. The results showed no difference in effect between the intervention and usual care on the number of people who returned to work (relative risk (RR) 1.01, 95% confidence interval (CI) 0.70 to 1.48). Stated in GRADE terms, we have very low certainty in the effect estimate. The evidence is weak, due to methodological limitations of the studies and high statistical heterogeneity, which means that it is uncertain whether multidisciplinary rehabilitation has an effect on RTW.
Most studies also reported health related outcomes, such as physical function and pain. We have very low certainty in these effect estimates, due to methodological limitations of the studies and imprecise effect estimates. That means that the effect of multidisciplinary rehabilitation on health-related outcomes is uncertain. Three studies assessed cost effectiveness. Two studies reported the intervention to be cost-effective compared to usual care, and the third found no difference. However, the evidence is weak, and it is uncertain whether multidisciplinary rehabilitation is cost effective.
Multidisciplinary rehabilitation vs. other active intervention. Five studies (six articles) presented the effect of multidisciplinary rehabilitation versus other active intervention. The results of the meta-analysis showed no difference at 12 months follow up (four studies; RR 1.04; 95 % CI 0.86 to 1.25) or at 24 months follow up (two studies; RR 0.94; 95 % CI 0.84 to 1.05). However, the evidence is weak, and we have very low certainty in the effect estimate at 12 months follow up. We have moderate certainty in the effect estimate at 24 months follow up (the findings at 24 months follow up are less heterogeneous than at 12 months follow up). The studies also reported results for symptom reduction for depression, pain, health related quality of life, health problems and fear-avoidance behaviour. Because the evidence is weak, we have very low certainty in the effect estimates.
Work-related cognitive therapy vs. usual care. Two of the three included studies reported that work-related cognitive therapy led to shorter sick leaves compared to usual care. The third study reported no difference between the groups. We have low certainty in this finding due to the studies’ methodological limitations. One study assessed symptom reduction among participants with musculoskeletal disorders after six months, and found no difference between the groups in pain, depression and physical function. Because the evidence is weak, we have very low certainty in the effect estimates.
Additional results. We identified seven comparisons where only one study presented the relevant comparison. The results showed effect on return to work for a stress reducing intervention, as did one study using an electronic health intervention with education and assistance on the internet. The remaining five comparisons found no clear effects of the interventions for faster return to work. Because the evidence is weak, we have very low certainty in the effect estimates.
In three of the included systematic reviews, a small effect on return to work was reported for certain subgroups or secondary outcomes, however, none of the reviews found any clear effects of work-related interventions on return to work.
Discussion
Overall, the included studies show no effect of work-related interventions compared to other active interventions or usual care. The studies had various methodological limitations concerning randomization and blinding, and the studies reported imprecise effect estimates. Therefore, our certainty in the effect estimates vary between very low and moderate. Because the present evidence on work related rehabilitation is weak, there is a need for further high-quality randomized trials that involve a sufficient number of participants, and with satisfactory reporting of both methods and results. Such studies will improve our certainty in the findings and that the true effects of the interventions are reported. Our findings are similar to several other systematic reviews, which also found no effects of work-related interventions on RTW. Further, our results correspond with a systematic review from 2005. The researchers reported high heterogeneity among the studies and underlined the need for identifying which intervention components had an effect. The results from our review and previously published reviews show that the research so far has not documented that work-related rehabilitation is a more effective follow up of persons on sick leave than usual care.
Conclusion
Work-related interventions does not appear to increase RTW more than other active interventions or usual care. We have very low to moderate certainty in the results, and future research might change this conclusion. None of the included interventions had any negative effect on RTW. The effect of the interventions on symptom reduction, physical, social and cognitive function is reported as better or the same as other active interventions or usual care.
This systematic review did not uncover any new evidence compared to previous reviews, and our conclusion reflects the findings of previous systematic reviews.
