I write about incentives to change behavior quite a bit. Sometimes, it's to acknowledge that it might work. Other times, it's to question if it might not. But there's no question that incentives do have the potential to help people achieve goals. A new study in JAMA tries to tease out to whom we should be directing those carrots.
David Asch and colleagues recognize that financial incentives both to patients and physicians are being used more and more often. But research often is lacking on how those function in practice. They wanted to see how incentives could be used to help patients with a high level of cardiovascular risk lower their low-density lipoprotein (ie bad) cholesterol.
They set up a four-group randomized controlled trial in a number of primary care practices. They enrolled both primary care physicians as well as patients. Patients were eligible if they had a 10-year Framingham risk score of 20% or higher with an LDL level of at least 120 mg/dL or a risk score of 10%-20% with an LDL of at least of 140 mg/dL. They were able to blind those gathering data for the study, but not the participants, obviously.
Doctors were eligible for $256 for each patient who met a quarterly goal (up to $1024 per patient per year). This money was kept separate from their other salary, so that it could be easily identified as part of this initiative.
Patients were incentivized differently. Each day, they had a 10% chance to win $10, and a 1% chance to win $100, but only if they had taken their medication the day before. If they were completely adherent, they could potentially win (on average) about $1022 per year. Information about compliance was uploaded automatically from electronic pill bottles. The daily drawings also reinforced the idea that medication needs to be taken every day.
One group was physician incentive only, as described above, and one group was patient incentive only. A third group had both physician and patient incentives, but at half the value of the individual incentive groups. A fourth group received no incentives and served as the control.
The outcome measure of interest was the change of LDL level one year out. Simple as that. The results, however, were interesting.
Patients in the control group wound up lowering their LDL levels by just over 25 mg/dL, which shows how just being in a study can have an impressive effect on health. People in the patient-incentive-only arm saw a reduction of the exact same amount. They didn't seem to work at all. Patients in the physician-incentive-only arm saw a reduction by just under 28 mg/dL which wasn't significantly different from control either. In the joint incentive arm, however, the average reduction was 33.6 mg/dL, which was significant.
In other words, incentivizing either group alone didn't appear to do much at all, with respect to lowering LDL values. But incentivizing both, even though the money spent was no greater, had a significant effect.
I've been down on physicians incentives at times because I don't think they've been shown to have the effects people sometimes attribute to them. But this study is interesting. It shows that by focusing either on patients or providers alone, we might not do much good. Focusing on both, however, seems to work. It has the added benefit of also aligning with many of the models of ideal care, where doctors and patients work together to achieve better health. Something to think about.