Do you regularly read about the economics of mental health? If so, you’ll be well-acquainted with the mind-boggling (and mostly useless) gargantuan numbers that represent the cost to society of mental health problems.
The hugeness of those numbers usually isn’t explained by health care costs. Rather, it’s mostly explained by things going on in the labour market. If people are unable to work, or to contribute to society in other ways, we can see that as a cost borne by society as a whole, or perhaps by employers.
A key consideration for health economists is what we call ‘perspective’; for whom is an intervention cost-effective? The NHS is primarily concerned with health outcomes; the most cost-effective intervention is that which achieves the most health improvement per pound of expenditure. Other perspectives (including a patient’s perspective) are likely to differ.
In the context of workplace interventions, funded by an employer, health outcomes may be secondary. A Dickensian capitalist may want their workforce to produce more widgets at the expense of their health. But even an enlightened employer who cares for their employees needs to consider their bottom line.
So, could people be supported in the workplace in a way that is cost-effective for society and for employers? That’s the subject of a recent review by de Oliveira et al (2020); how might mental health and substance use be addressed in the workplace? In particular, the review is concerned with identifying where it might be possible to get the highest return on investment. Some reviews have been conducted in this context before, but this one is by far the most comprehensive.
Employers need to know which workplace mental health interventions work best, so they get the best bang for the buck from their investment.
Methods
The literature search for this review was built around the following PICOS statement:
- Population: employed adults with a mental health and/or substance use disorder;
- Intervention: employer-funded workplace interventions improving a work-related outcome;
- Comparator: no care or usual care;
- Outcomes: effects on mental health or substance use and work-related outcomes such as productivity or return on investment;
- Study design: economic analysis with measurement of change.
The authors searched a whopping 20 databases, including an impressive coverage of grey literature sources alongside the usual suspects and some smaller collections. They adopted a robust approach to independent screening and data extraction across three reviewers.
Studies were separately evaluated for methodological quality (using QHES) and reporting quality (using CHEERS). Data extraction was only done for studies of high quality and there was some synthesis of evidence according to the disorder.
Results
The researchers’ thorough combing of the literature resulted in them screening 11,018 citations, more than 500 of which were sourced from outside of the academic literature databases.
The screening process resulted in 56 studies being included, with data extraction conducted for the 23 studies that were deemed high-quality.
Depression and smoking got the most coverage and a large proportion of studies were from the US. Of those studies looking at depression, cognitive behavioural therapy (CBT) was the most common intervention. Of those studies concerned with smoking, most looked at pharmacotherapy coverage.
The perspective of the analysis is key. Around a third of studies adopted an employer’s perspective, while the rest variously adopted health care, societal, or employee perspectives.
The included studies tended to use cost-benefit analysis. That’s where both the costs and the benefits of the intervention are valued in terms of money. We don’t tend to use this in the context of health care because we’re usually interested in health as the outcome.
So, what works?
After synthesising the data, the review found at least moderate evidence to support the following interventions:
- For depression
- CBT alone
- CBT in combination with care management
- For multiple mental health disorders
- Active involvement of occupational health professionals in return to work
- For smoking
- Brief counselling
- Varenicline
- Bupropion
Return on investment for CBT with care management was between $0.39 and $3.35 per dollar spent.
Conclusions
- There are workplace interventions available that employers can use to improve outcomes relating to depression or smoking.
- Not only are these likely to be good for employees but they may also be cost-saving for employers.
Employers should consider investing in workplace interventions for depression and smoking.
Strengths and limitations
This study is one of the best reviews of economic evidence I have read. The methodology is thorough and the reporting is clear. It’s difficult to think of ways in which a better review could have been conducted.
There was one thing that bugged me about the review; sleep-related interventions were excluded. Presumably, that’s because of the interests of the funder (the Centre for Addiction and Mental Health Foundation in Canada), but it seems to defy logic. To paraphrase one study by Manocchia et al (2001); good sleep, mental health, and productivity go hand-in-hand.
Another important question that isn’t addressed by this study is what isn’t cost-effective. Employers’ options may be limited by their specific characteristics or context and so it may be just as important for them to understand what the evidence shows that is not cost-effective. This may help facilitate the phasing out of existing ineffective programmes.
But the most fundamental limitations are not the fault of the authors; in some areas there is simply a lack of evidence. And inconsistencies between the studies meant that data couldn’t be pooled and synthesised according to factors other than the relevant disorder. That’s important when we consider the implications of the findings.
For employers, knowing which mental health workplace interventions are NOT cost-effective may be just as useful as knowing what is cost-effective.
Implications for practice
These results should be useful to employers and to those in a position to support employers in making investment decisions. It is reasonably clear from the evidence that workplace CBT should be considered for depression and drug coverage should be considered for smoking.
But there is a big unanswered question, and that is what the necessary conditions are for these interventions to be cost-effective. As the authors of this review explain, the available evidence provides very little information on contextual factors. In particular, we don’t know what kind of organisational change might be necessary to facilitate the cost-effective provision of these interventions.
Before policy-makers rush to hand responsibility to employers, we also need to consider whether NHS provision or workplace provision is better. Sure, CBT in the workplace might be good. But is it as good as CBT in an NHS setting? Is it better? It’s possible to think of reasons why one might be better, or more cost-effective, than the other. This review isn’t able to answer that question.
The study also identifies some gaps in the literature, of which researchers should take note. What about self-harm? What about gambling? What about drug use other than tobacco? This mammoth review didn’t find evidence on the economics of workplace interventions for many important mental health problems.
What kind of organisational change might be necessary to facilitate the cost-effective provision of workplace mental health support?
Statement of interests
I am currently working on an evaluation of a sleep-related technology, which is probably why I mentioned sleep. Other than that, I have no relevant interests in this study.
Links
Primary paper
Oliveira, C. de, Cho, E., Kavelaars, R. et al (2020) Economic analyses of mental health and substance use interventions in the workplace: a systematic literature review and narrative synthesis. The Lancet Psychiatry 2020 7(10) 893–910. 10.1016/s2215-0366(20)30145-0.
Other references
Manocchia, M., Keller, S. and Ware, J.E. (2001) Sleep problems, health-related quality of life, work functioning and health care utilization among the chronically ill. Quality of Life Research 2001 10(4) 331–345. 10.1023/A:1012299519637.
Photo credits
- Photo by Michael Longmire on Unsplash
- Photo by Arlington Research on Unsplash
- Photo by Sharon McCutcheon on Unsplash
- Photo by Yoann Boyer on Unsplash
- Photo by Kelly Sikkema on Unsplash
- Photo by Jametlene Reskp on Unsplash
- Photo by Good Faces on Unsplash
Thanks for the review, Chris. CBT usually comes up as the most cost-effective in economic terms. In studies of quantity of evidence, I wonder if CBT’s dominance becomes self-perpetuating. Here in Britain, CBT is by far the most used by the NHS. This means that there’s more evidence on it, which in turn allows best practice to be better identified and CBT to be made better. Is CBT just as dominant in other countries, or are we British particularly attached to it?
It’s a good point; there are plenty of people who would argue that we are too focused on CBT. But it is surely right that evidence for effective services should be self-perpetuating as implementation and capacity building facilitate the creation of new evidence. Why would we pursue research on services that we didn’t think might be effective? I’m not sure about other countries, but I’d be interested to know.
Thanks, Chris.
An interesting read Chris – thanks for summarising. Was there anything related to the relative merits of pro-active services (i.e. offer these services to all with individuals self-enrolling if they feel they need it) compared to re-active (offered to people who have absence due to mental health)? I imagine this falls into your “But there is a big unanswered question, and that is what the necessary conditions are for these interventions to be cost-effective” comment.