We will release a new Episode Radio Freakonomics later this week on the use of RCTs (randomized controlled trials) in healthcare delivery. It features the work of an MIT economist. Amy Finkelstein and her colleagues in J-PALand includes their analysis of what happened when Oregon expanded its Medicaid coverage.
If you’d like to get started on this topic, check out the new working paper (closed) titled “Does Medicaid for Pregnant Women Affect Prenatal Health Behavior?” The authors Dhawal M. Dave, Robert Kestneras well as George L. Wehby. They did not use RCTs, but worked diligently to answer a complex and important question in healthcare delivery. Their conclusion, highlighted in bold in the annotation below, is completely illogical (and, of course, for some at least, depressing):
Despite plausible mechanisms, few studies have evaluated the potential health behavioral changes resulting from the expansion of Medicaid in the 1980s and 1990s for pregnant women. Accordingly, we provide the first national study on the impact of Medicaid on the health behavior of pregnant women. We use exogenous variability associated with increases in Medicaid income for pregnant women and children between the late 1980s and mid-1990s to examine the impact on some prenatal behaviors and health outcomes using vital statistics data. USA. We found that an increase in Medicaid enrollment was associated with an increase in smoking and a decrease in pregnancy weight gain. A 12 percentage point increase in Medicaid eligibility increased rates of any prenatal smoking and smoking more than five cigarettes per day by 0.7 to 0.8 percentage points. The expansion of the Medicaid program was associated with a decrease in weight gain during pregnancy by about 0.6%. These effects decrease at higher levels of acceptability, consistent with a crowding out from private to public insurance. It is important to note that our data are consistent with the assumed moral hazard, although there is also an income effect. Worsening health behaviors may partly explain why the expansion of Medicaid has not been associated with significant improvements in infant health.
Just to be clear: gaining weight is good when you’re pregnant; And you shouldn’t smoke more than 5 cigarettes a day. Here is how the authors try to explain their conclusion:
One possible explanation for this somewhat paradoxical finding is that health insurance creates incentives for health behavior change along with lower health care costs. Insurance reduces the cost of treating an illness, such as an adverse medical outcome for a mother or child, which can lead to reduced maternal efforts to prevent such events (anticipatory moral hazard).
In addition, Medicaid coverage has an income effect on saved personal expenses and on health insurance premium costs (when private insurance is replaced by public insurance). This income effect can be used to purchase goods that improve infant health as well as goods that may harm infant health (such as cigarettes).
Incentives, incentives, incentives.