Health Insurance and Mortality, Part 2

Please read Part 1 of this post before continuing.

Not surprisingly, there have been several criticisms of the Sommers, Long and Baicker study. Here are the main ones I’ve been able to identify so far, and my comments on each.
  • The results contradict those of the Oregon Medicaid study. A 2013 study by Baicker, et al., examined the effects of Medicaid expansion in Oregon using a randomized control group design. (The opportunity to sign up for Medicaid was determined by lottery.) Although this study found economic and psychological benefits of Medicaid, its effects on objective measures of health were disappointing.  Blood pressure, cholesterol and blood sugar levels were all lower in the Medicaid group, but the differences were not statistically significant. However, the Oregon study’s sample size was too small to detect medically important health benefits. The Massachusetts study has many more particpants. If the same percentage decline in mortality observed in the Massachusetts study had been seen in Oregon, it too would not have been statistically significant. The two studies are not inconsistent. An important strength of the Massachusetts study is its larger sample size. Furthermore, it measured the effects of the entire health reform package, not just Medicaid expansion.
  • It costs too much. Using the Sommers, et al., data, Cannon calculated that it cost Massachusetts $4 million per life saved. He argues that this is not cost effective. However, this assumes that the only benefit people received from Romneycare was when it saved their lives. It ignores the many health and quality of life benefits people receive from medical treatments for non-life-threatening illnesses and injuries. What is the value of a knee or hip replacement that allows a person to walk free of pain for 20 years? How do you measure the benefits to a family of avoiding bankruptcy and the loss of their home?
  • Massachusetts is different from other states. Massachusetts is “whiter and more affluent” than most other states, but the matched comparison groups control for race and income. It could be argued that Massachusetts has a more effective health care system (more doctors, better-equipped hospitals) than other states. However, a convincing alternative explanation must explain not just lower mortality in Massachusetts, but all the results. Why did the mortality rate change from 2001-2005 to 2007-2010? Wouldn’t a better health care system be expected to help people over 65 as well? As Sommers, et al.state in their conclusion:
Although we cannot rule out unmeasured confounders, it is challenging to identify factors other than health care reform that might have produced this pattern of results: a declining mortality rate in Massachusetts since 2007 not present in similar counties elsewhere in the country, primarily for health care-amenable causes of death in adults aged 20 to 64 years (but not elderly adults), concentrated among poor and uninsured areas and not explained by changes in poverty or unemployment rates.
Of course, health insurance is useless if there are no doctors or hospitals in your area. The quality of the health system may have interacted with health care reform to produce a better result in Massachusetts than would be expected in other states. This does not explain away the results, but it may limit our ability to generalize from them.
  • More research is needed. This cliché is, of course, trivially true. However, it is unrealistic to expect definitive studies of effects of the Affordable Care Act (ACA), since it is being implemented simultaneously in all 50 states. There are no experimental and control groups, only before and after comparisons, the results of which can easily be dismissed as caused by other changes taking place in society at the same time. About the best we can expect will be comparisons between states that do or do not expand Medicaid. (Pennsylvanians will be happy to know that, thanks to Governor Tom Corbett and our legislature, future researchers will be counting the number of excess deaths in our state.) However, states that are not expanding Medicaid are already known to differ both economically and politically from other states. And while Medicaid expansion is an important provision of the ACA, it is only a part of it.

In short, the Sommers, et al., study may be the best that is available for the foreseeable future. Even conservative critics of health care reform are granting it grudging respect.  Megan McArdle stated, “(A)fter yesterday’s report, I’ve revised the probability of ‘huge benefits’ [from health care reform] upward, and you should do the same.”
You may also be interested in reading: