There have been many studies looking at factors that can be used to predict whether a person will return to work after developing pain. Many, quite naturally, look at factors relevant to acute and subacute pain problems, aiming to reduce the number of people eventually progressing to longterm disability. Some people will slip through these interventions – for many reasons – and the news for returning to work is not good for those who do have extended periods of time away from the workforce.
The factors that might precipitate stopping work in the acute or subacute stages may not be the same as those involved in maintaining longterm disability. For example, it’s well known now that pain-related fear and avoidance can influence people’s initial disability (see Prkachin, Schultz,& Hughes, 2007), but studies looking at whether interventions to reduce avoidance behaviour through psychosocial methods haven’t been entirely positive (e.g.Frost, Haahr,& Andersen, 2007).
Marois and Durand describe a descriptive correlational study of over 200 people with work disability who participated in the Canadian ‘Previcap’ work rehabilitation programme.
This programme is a roughly 12 week interdisciplinary programme based on the ‘Sherbrooke’ model ( go here for a nice powerpoint by Patrick Loisel on this programme). The programme has three components -
Work disability diagnostic interview, which is intended to ‘rule out a serious medical diagnosis and to formulate a diagnosis based on one or more factors explaining long term work disability of musculoskeletal origin.’
Preparation for Therapeutic Return to Work – clinic-based interdisciplinary rehabilitation
Therapeutic Return to Work – graded exposure to a real work environment
Participants in this study had been enrolled in these programmes and included in this study if they had persistent musculoskeletal pain, an existing employment relationship (i.e. had a job to return to), and had been completely off work for a minimum of 12 weeks.
Data was collected twice during the participants progress through the Previcap Programme – once at the initial assessment, and again at completion of the programme. I’m not going to go through the method or measures, suffice to say that 83 different factors were measured, and the authors state that ‘these variables … include most of the factors identified in the literature as being associated with long-term work disability during the sub-acute phase’. Several additional factors were added by a committee of clinicians.
Work status was measured at the end of the programme in terms of ‘back at work’ full time or part-time, or ‘able to go back to work’ but had no job available or had non-musculoskeletal problem reasons for not returning. The final group were ‘unable to return to work’ related to their musculoskeletal problem.
I don’t want to go into any detail with regard to the analysis undertaken – suffice to say that it involved a lot of logistic regression both multivariate and univariate! The full details are described in the article itself. What is interesting are the characteristics of the people in the programme, and the findings themselves. What is even more interesting are the implications.
Cutting to the chase: the participants were around 40 years old, had an average time off work of 31 weeks, particpated in the Previcap Programme for about 15 weeks or so, and mainly, returned to work – more women than men did so, however.
The factors that ended up predicting return to work were different for men and for women, and three different models of return to work were developed. Out of the original 65 factors measured, only 17 were retained – and these were mainly work related and psychosocial.
So, what were they?
The key ones from this study for the ‘general’ model were:
fear of aggravating symptoms
no light duties
awkward postures needed at work
static work postures needed
longer time on the Previcap Programme
For the men’s model:
absence of musculoskeletal history requiring time away from work
longer time away from work
no light duties
longer time on the Previcap Programme
For the women’s model:
higher psychological distress
fear of aggravating symptoms
presence of recent personal events
presence of work equipment that is unsuitable
Of lower odds ratio, a common factor for all models was the person’s perception that treatment had failed.
These models were able to predict between 84 – 96% of those who returned to work, but were less successful at identifying those who didn’t do so (37 – 48%).
So, where does that leave us?
Now first of all, some interesting things to ponder.
At least in New Zealand, people who still have an employment relationship are more likely to return to paid employment than those who don’t. Three months isn’t a long time in the rehabilitation game, but it is to an employer who needs to have someone on deck to keep the wheels of industry turning. In another roughly three months, (six months total time away from work) some estimates suggest that the chances of returning to work are around 50% (see Waddell, 2004 for example), while by 12 months, the chances have reduced to around 10% or less. So my best guess is these people were slightly different from those in New Zealand – after roughly 30 weeks off work, I’m not sure that as many NZers would still have jobs to return to.
Another point to consider is the way ‘able to return to work’ is measured. I can readily understand those who do return to their original job (either in suitably selected or reduced hours), but when someone is judged ‘able’ to return to work in the absence of actually testing it in the real workplace, I’m suspicious. Read here for a great discussion of ‘ designated guessers ‘ by Dr Jennifer Christian.
There are several factors that seem to need to line up for someone to return to work: they need to believe they can be successful, they need to believe they won’t harm themselves – and that they have what it takes to cope. There also need to be suitable work duties and employer support, and the support from a health provider (preferably someone familiar with the workplace).
The factors that were associated with unsuccessful attempts to return to work were quite interesting. I’m not at all surprised by the finding that men with a shorter period of time in employment at that place of work were less likely to re-engage. For women, attempting to return to work but failing (notably because of increased pain), and ‘wanting’ to return to work were both associated with poorer outcome.
The authors suggest that ‘wanting’ to return to work might be influenced by social desirability in women – I’d suggest there could be a lot of reasons for women to say they want to return to work, but not be able to, including the influence of others (who benefits when a woman doesn’t return to work?) and logistic problems around re-engaging in the work role alongside other occupational demands for women (transport to and from work, integration of grocery shopping, childcare, household management and so on).
Another interesting finding is the relationship between health services factors and return to work – the person’s perception that health care has failed (maybe a mismatch between expectations and results?), and, in men at least, duration of a work rehabilitation programme. Perhaps this means more time to address the obstacles that are identified by the person, maybe it’s partly the mediating role that health providers can play to liaise between the person and the employer (acting as a buffer to set limits perhaps).
More on return to work and similar things tomorrow – sorry for the length of this post, but returning to work is a complex business!
If you’ve enjoyed this post, and want to continue reading – subscribe using the RSS feed link above, or you can bookmark and return at your leisure. I post most days during the week, and love comments and discussion.
Marois, E., & Durand, M. (2009). Does participation in interdisciplinary work rehabilitation programme influence return to work obstacles and predictive factors? Disability & Rehabilitation, 31 (12), 994-1007 DOI: 10.1080/09638280802428374 Frost, P., Haahr, J. P., & Andersen, J. H.(2007). Reduction of Pain-Related Disability in Working Populations: A Randomized Intervention Study of the Effects of an Educational Booklet Addressing Psychosocial Risk Factors and Screening Workplaces for Physical Health Hazards. Spine 32(18),1949-54. Prkachin, K. M., Schultz, I. Z., & Hughes, E. (2007). Pain Behavior and the Development of Pain-related Disability: The Importance of Guarding. Clinical Journal of Pain Vol 23(3) Mar-Apr 2007, 270-277.