It is rare and hard to see true randomized trials in the health insurance or health service delivery sector. One regarding use of primary care is presented in a new study published at the National Bureau of Economic Research. (NBER Paper) The theory driving investment in primary care is that it can keep people in better health and coordinate care in a manner that reduces utilization of more expensive sites of care, like the ER or inpatient hospitalization. The researchers attempted to test this theory by inventing a group of people to use primary care and see what the effect was on overall utilization and spending. The researchers had previously shown that paying people $25 or $50 led to a greater likelihood of an initial primary care visit. In this trial the authors looked at whether the incentives encouraged an ongoing relationship with the PCP and at use and cost within 12 months after enrollment in the study. The basic setting for the trial was a community-based care program established by Virginia Commonwealth Medical Center, which provided free or low-cost care to low-income residents without other insurance coverage. Participants were randomized to receiving an incentive to have a primary care visit or not getting one. There were a little over 400 patients in each of the no incentive, $25 incentive or $50 incentive groups. To get the incentive, a PCP visit had to occur in the first six months after enrollment in the study. A variety of health, demographic and socio-economic data was collected on participants.
In both the initial six months after enrollment and the second six months, patients receiving the $25 and $50 incentive during that first six months had more PCP visits. Generally, however, ER use in the first six months was not statistically different among the groups, and that lack of difference persisted in the second six months. Most ER visits were for conditions deemed to be true emergencies. Inpatient utilization was also not significantly different. There were more outpatient visits among the incentivized group, probably because once they saw a PCP, there may have been referrals to treat specific issues. Although spending was higher for the incentivized groups, the difference did not reach statistical significance for most comparisons. The effect of the incentives are shown most clearly in the speed with which patients had a visit with a PCP. The incentive led to people having a visit one to two weeks more quickly than when there was no incentive. It also appeared that the strongest effects were on patients who were more likely to have health issues and needs, which is good. This is a relatively short follow-up period, and it would not be unreasonable to expect that over longer periods of time there might be a larger effect, particularly for those patients who continued to see their primary care doctor regularly.
As the researchers note, they did not look at the quality effects of encouraging people to establish a primary care relationship. It is almost certainly a good thing for everyone to have an annual review of their health and health issues and to have a relationship with a clinician who knows them and their health status and needs. It is problematic to me to pay people to do what they should do anyway. It is unfair to those who do the right thing in the absence of an incentive. Finding some sticks in addition to the carrot would help eliminate that tendency among some people to behave irresponsibly and then also avoid the consequences of that behavior. A lot of people who get new insurance coverage, usually Medicaid, tend to continue to use the ER or use it even more, just because they find it convenient. Refusing to pay for such inappropriate use is the only way to stop that behavior.