Paying for value, rewarding high-value care, pay-for-performance—all are examples of terminology used to describe aligning financial incentives with clinical goals and processes. Essentially, these policies and programs seek to link quality to payment and their influence is growing, extending even to Medicare. While these concepts have been discussed repeatedly by many in healthcare, including the Institute of Medicine, most of the focus has been on incentives for providers in the areas of chronic care management or preventive care. Other studies have discussed the impacts of different data collection mechanisms (EHR vs. chart review) on these financial incentives.
Notably absent from the mass of studies of financial incentives are studies about the use of financial incentives in mental health care. Although the number has grown substantially since the last time my colleagues and I looked at the issue, there are still relatively few programs with results published in the peer-reviewed literature.
In a recently published Medical Care article, Puyat and colleagues (from the University of British
Columbia and Simon Fraser University) provide evidence on the use of payment incentives to promote improved depression care in a Canadian cohort. The authors use administrative data from 2005-2012 to evaluate the impact of financial incentives on treatment for patients with newly diagnosed major depression. A few indicators were used to measure mental health care for depression: receipt of any counseling/psychotherapy, receipt of any antidepressant therapy, minimally adequate counseling/psychotherapy (4 or more sessions), minimally adequate antidepressant therapy (84 or more days), as well as care continuity (whether care was provided by the same general practitioner (GP) within a 12 month period).
The government of British Columbia introduced a set of financial incentives [PDF] largely attached to the existing fee-for-service system—the mental health portion of which that took effect in 2008. The incentives introduced payments for:
- Assessment and coordination of mental health care by general practitioners,
- Telephone and email management/coordination between clinicians, and
- Increased numbers of mental health visits.
After using segmented regression [PDF] to assess the impact of the policy change on the mental health outcomes outlined above, Puyat and colleagues found that the incentives only modestly changed provider behavior (at best). Downward trends in continuity of care were disrupted after the policy change. Slight increases were observed for any counseling/psychotherapy and minimally adequate counseling/psychotherapy. On the other hand, slight decreases were observed for receipt of any antidepressant therapy and minimally adequate antidepressant therapy. Given the contemporaneous publications of studies about antidepressants’ lack of effectiveness, perhaps this is not so surprising.
The failure of incentives in mental health care to have large, lasting impacts on mental health care has been seen in other studies. Depression in Primary Care: Linking Clinical Systems and Strategies [PDF], was a $12 million dollar program funded by the Robert Wood Johnson Foundation from 2000 to 2007 which had the goals of (1) encouraging the use of a chronic care models to recognize and treat depression and (2) developing financial (and organizational) incentives to encourage physicians to use the chronic care model. Among the many results of the study, success in maintaining new payment methods was limited at a majority of sites.
Among the many potential reasons given by the authors for the mixed effects, additional barriers to receiving treatment for antidepressants or to receiving mental health care in general are strong contenders. In particular, barriers for patients may limit the impact of financial incentives for providers. As has been seen in some other contexts, financial incentives may change the relative delivery of care, towards the actions that create the greatest rewards. Additional research in other samples with other incentive structures would be useful for the design of these programs going forward.