Rolling the Dice on Health Insurance

The Affordable Care Act will extend health coverage to millions of low-income Americans beginning in 2014. This expansion of care has engendered fierce debate in Congress, the media, and the Supreme Court, as opponents argue that government-provided health insurance will not benefit the poor enough to justify the cost and may even make them worse if the quality of care is low. Quantifying the effect of gaining health insurance on this low-income population is critical to forecasting the impact of health care reform. Isolating the effect of health insurance has proved difficult. The uninsured differ from the insured population in many ways that could be correlated with health outcomes. Random assignment of health insurance to individuals by researchers would raise serious medical issues and be extremely costly. Thus, despite numerous studies, until now this debate suffered from a lack of reliable, randomized data on the effect of health care, leaving the national argument mired in ideological contention.

In new research from The Oregon Health Care Experiment: Evidence From the First Year, authors Finklestein, Taubman, and Wright et. Al, take advantage of a rare random sample to examine the effect of accessing health insurance on low-income individuals’ health outcomes, financial stability, and use of medical care. In 2008, Oregon realized it had enough funding to reopen enrollment for up 10,000 adults into its state health insurance program for those earning up to the Federal Poverty Line, Oregon Healthcare Plus (OHP). Foreseeing that demand would vastly outstrip supply, Oregon sought permission from the Federal Government to offer a lottery to apply for health insurance. Ninety-thousand individuals entered the lottery, of which 29,600 were selected. About 30 percent of the lottery winners successfully enrolled in OHP, as 40 percent of winners failed to return applications and about half of those that did return them failed to prove eligibility.

The researchers studied health outcomes and financial stress across the treatment group of lottery winners and the control group of lottery losers for an average period of 14 months, wrapping up their observations in September 2009. They used hospital discharge records to measure health care utilization, a survey mailed to lottery winners to gauge self-reported health outcomes, and TransUnion’s Consumer Credit Database to examine the effect of health insurance on financial strain. Credit reports were matched to 68.5 percent of 29,600 lottery winners and 45,800 of the control population of lottery losers studied, while researchers obtained survey results from approximately 50 percent of lottery participants.

The researchers studied a largely low-income and unhealthy sample. Seventy percent of respondents reported incomes below the program’s poverty line, and almost 70 percent had only a high school diploma or less. Eighteen percent of survey respondents reported being diagnosed with diabetes, 28 percent with asthma, and 56 percent had screened positive for depression. Oregon lottery applicants were not representative of an average national population, as 56 percent were female, and only four percent were African American.

Researchers found that gaining access to health insurance had a substantial positive impact on this population. Using a two-stage estimation method combining the effects of being selected in the lottery with gaining insurance through the lottery, the researchers isolated the effect of attaining insurance on health use, outcomes, and financial strain. Two-stage estimation also ensured that the effects of winning the lottery on health operated through take up of health insurance by those who would not otherwise have gained health insurance over the year of study. Researchers also checked to see if winning the lottery might also affect health through uptake of other public services such as food stamps, but only found very small effects.

Gaining insurance was associated with dramatic increases in health care utilization according to the hospital discharge and survey data. The probability of being admitted to the hospital increased by 2.1 percent and the probability of having a procedure performed increased by 45 percent. Responses to the survey implied that gaining insurance through the lottery led participants to make 55 percent more outpatient visits and take 15 percent more prescription drugs. New access to insurance especially influenced compliance with preventative care. For example, the probability that lottery winners had their cholesterol checked increased by 20 percent. Using a rough estimate of costs, the researchers estimated that higher use of medical services increased medical spending by $778 per new enrollee, an increase of 25 percent relative to the control population. The authors note, however, that spending on preventative care might lower future health costs over a horizon longer than the 16 months studied.

This substantial increase in health care utilization led to significantly higher self-reported health outcomes, although the limited time-scale of the study prevented researchers from making claims about improvements in actual physical well-being. Lottery winners reported good or better health at a 25 percent higher rate than the control group average, a ten percent lower incidence of depression, and a 32 percent increase in self-reported happiness. The authors suggest that some of this improved sense of well-being may have stemmed from improved financial security, as collections for unpaid medical bills fells by 20 percent relative to the control mean and the probability of having to borrow money or skip bills to pay for medical expenses fell by 40 percent compared to the control group.

The striking effects of health insurance on health utilization and outcomes for low-income individuals shown by this rare experiment has major policy implications, as the Affordable Care Act requires states to extend Medicaid eligibility to all adults earning up to 133 percent of the poverty line. Access to health insurance increased health care use and improved self-reported health, suggesting that extending Medicaid could offer dramatic benefits for low-income and previously uninsured populations.

 Feature Photo: cc/(canonsnapper)