Workplace wellness programmes are an assemblage of wellbeing activities like yoga or cycling clubs, packaged together with diagnostic activities like biometric screenings; their aim is to reduce sickness, increase productivity and cut insurance costs for an organisation’s members. This is big business – in the USA, the market is around $8 billion – with a return-on-investment claim, thanks to a plethora of studies that tout the benefits of these programmes (for example, see this meta-analysis from 2010). But whether staff enter these kind of initiatives in the first place is usually up to them, making it hard to evaluate their effectiveness, as those who choose to participate may differ in key ways from those who do not. To assess the benefits of the programmes accurately therefore requires a randomised-controlled study. This is what the National Bureau of Economic Research published recently, and it leaves these programmes looking sickly.
Damon Jones and his team set up a staff wellness programme at the University of Illinois containing the three core elements of such programmes: a biometric screening to give an indicator of overall health and a health risk assessment to flag up specific risks for the individual and indicate routes to improvement. Finally there were the activities themselves, including programmes for weight management, guidance on smoking cessation and healthy workplace habits, and tai chi sessions. Over 12,000 university employees were invited, of which nearly 5,000 showed an interest and completed surveys covering their health status, health-care use, job satisfaction and productivity.
Around a third of these interested people were then randomly assigned to a control condition and received no further support. The others were invited to be screened, risk assessed, and then to take on the available wellness activities offered over two semesters. Finally, at the end of the academic year in July, one year after they first enrolled in the programme, the participants completed a follow-up survey covering the same topics as at baseline, plus some additional questions.
Sickness takes a toll on national economies – the cost to the US economy where this research was conducted is over $200 billion per year. Based on the new evidence, do wellness programmes show promise in making a dent in it? Unfortunately not.
Compared to the control group, participants in the programme did not take significantly fewer sick days. Nor were they more likely to stay in their job, nor get a promotion or get a pay hike. They didn’t spend any less on medication or hospital visits, undermining the financial rationale in reducing medical insurance claims (or in the UK, burden on the health service). And those in the programme showed no improvement in health behaviours such as using the gym or taking part in a popular annual running event. A few effects proved statistically significant – feeling subjectively happier at work and one measure of job satisfaction – but the remaining 37 measures were non-significant.
It seems that the past studies that did find effects have been riding on a selection effect due to the kind of people who choose to get involved in such programmes.
Looking at the 7,000-plus people who declined to engage in the wellness programme on offer in the current study, we see they were, on average, older and had either more medical spend than average, or were zero-spenders on health (i.e. suggesting either unusually poor or good health prior to the start of the study). Those who signed up and were enrolled in the programme, but dropped out, were also more likely to be either high-spenders or zero-spenders on their health.
These group differences in health were mirrored in income, where those who didn’t engage in the programme tended to be rather wealthy or poor. Those who did engage were already, before the study, much more likely to be active at the gym or go running.
To summarise, those staff most inclined to opt to take part in the wellness programme were active, fairly young, and moderately well-off, with little evidence of chronic health problems. If these patterns generalise to the opt-in patterns at other organisations, it makes it unsurprising that prior wellbeing programme research (that hasn’t used a randomised design) has tended to find good health outcomes for participants – they would probably be doing well regardless.
This stringently-designed study suggests that there is no “return on investment” case for these programmes, at least in the form investigated here. The programmes may have a tangential benefit in attracting healthy workers to the organisation – a sort of targeted perk – but not in raising wellbeing as they proclaim. And the population you would most hope to aid – those who are poorer, currently less active, and with more pressing health issues – are the least likely to engage, although they will still be paying towards the programmes through insurance premium hikes (or taxation in other systems).
This study can’t speak to why certain individuals are deterred from signing up, but perhaps it has to do with their other commitments and dependents, and their perceptions of the programmes as somehow not for them. To increase up-take among these groups will therefore likely require addressing these perceptions and providing additional support to help overcome any obstacles to taking part.
For now, the impression of a positive impact given by wellness programmes looks largely a mirage.