Whether health insurance improves health became a controversial topic when a large quasi-experimental study found mixed results regarding the effects of Medicaid on health. Before we examine that study, it is worth discussing the difficulties in answering this question.
People who gain insurance generally do so because they have become sick, or so poor that they are eligible for public insurance or subsidies to purchase insurance. As such, these people are not comparable to the population that already had insurance. They are likely to either make larger than average gains (if insurance does indeed improve health) or smaller than average gains (if insurance does not improve health). And if there are heterogeneous effects, then the effects could cancel – for example, insurance may work for people with injuries, but not for those with chronic illnesses.
The real world offers few experimentally pristine conditions. One option is for researchers to conduct their own experiment. The RAND Health Insurance Experiment, conducted from 1973 to 1982, assigned families to one of 14 different plans, each with a different cost-sharing structure. The study found little evidence that higher cost-sharing levels impacted health, except for one group – the poorest group who began the experiment with high blood pressure. For this group, only those randomized to a plan with no cost-sharing experienced a reduction in blood pressure, suggesting that the generosity of insurance may be clinically meaningful.
Medicare provides useful “quasi-experimental” conditions, because those who have just become eligible for Medicare are not much different than those who been eligible for just a short while. No one chooses to turn 65 – it just happens. Exploiting this design, researchers found that Medicare improved access to breast cancer screening, especially among black and Hispanic women; and that it reduced deaths among emergency department patients by 20%.
Finally, Oregon provided researchers an opportunity to study the effects of insurance when it decided to expand its Medicaid program. In 2008, Oregon offered the chance to apply to Medicaid to 35,000 people who were drawn randomly by lottery from a list of 90,000. Because not all applied, and others were ineligible, only 30% enrolled.
In the first year of the expansion, Medicaid increased the probability that recipients reported their overall well-being as “excellent” or “good” (compared to “fair” or “poor”) by 25%; reduced depression by 30%; and reduced financial hardships, such as skipping paying bills because of medical expenses, and having unpaid medical bills sent to a collections agency. On the other hand, the authors did not find evidence that Medicaid reduced blood pressure, HDL or total cholesterol, glycated hemoglobin, or overall cardiovascular risk. Medicaid effects on health persisted after 2 years of the program.
These studies do not answer the question about how much coverage is warranted, or how benefits should be structured, but they do point to the worthiness of health insurance coverage. Insurance improves access to care, mental well-being, and financial stability. For some populations, it may not improve physical health, but for others, it saves lives.