Massachusetts’ coverage expansion, access, and outcomes

by The Incidental Economist on January 14, 2014 · 3 comments

Several recent papers have shed new light on the post-health reform experience in Massachusetts. Since the state is about seven years ahead of the nation in implementing a reform very similar to the Affordable Care Act, it’s always worth paying attention to new work in this area. However, it should be kept in mind that Massachusetts is different from other states culturally, politically, and economically, so it is never clear how much one can generalize from results pertaining to its reform.

First, let’s back up and review some older work. The recent paper by Amelia Bond and Chapin White offers a nice summary of the effect of coverage expansion on access in Massachusetts:

From 2006 to 2010 the percent of the population with a usual source of care increased by 4.7 percentage points, and the percent of the population with a preventative health visit increased by 5.9 percentage points (Long, Stockley, and Nordahl 2012/2013). The percentage of the population with visits to specialists, multiple physicians, and dentists also increased (Long, Stockley, and Dahlen 2012). [A] recent survey also demonstrated decreases in emergency department utilization and in hospital stays. [...]

A study by Miller (2012/2013) used the National Health Interview Survey to look directly at office visits, which suggests an increase of 4% in office visits for those aged 18–64 years in Massachusetts relative to surrounding states. Using the Healthcare Cost and Utilization Project National Inpatient Sample, Kolstad and Kowalski (2012) reported a decreased length of stay and decreased admissions originating from the Emergency Department (ED) in Massachusetts after reform. Their results also suggest that there was little change in the rate of preventable hospitalizations, but that the severity of patients with preventable hospitalizations decreased. Another study by Miller (2012) using administrative data looking directly at ED visits found a 2.4 percentage point decrease in Massachusetts. A number of these results may indicate increased primary care access with patients on average able to access care at an earlier stage.

The one known study looking directly at Medicare beneficiaries uses preventable hospital admissions as a proxy for access to preventative primary care visits and finds no detrimental effect of the insurance expansion on the Medicare population (Joynt et al. 2013). [Links added.]

Bond and White’s contribution is based on Dartmouth Atlas and Medicare utilization data:

Principal Findings. In areas of Massachusetts with the highest uninsurance rates— where insurance expansion had the largest impact—[primary care] visits per beneficiary fell 6.9 percent (p < .001) relative to areas of Massachusetts with the smallest uninsurance rates.

Conclusions. The expansion of coverage for the nonelderly reduced primary care visits, but it did not reduce the percent of beneficiaries with at least one visit. These results could imply restricted access, increased efficiency, or some blend.

In totality, the evidence of these studies doesn’t strongly suggest health care access problems stemming from Massachusetts’ coverage expansion. Bond and White’s results offer a pessimistic spillover effect, and one we might expect. However, if it’s true that Medicare beneficiaries’ access to more than one visit decreased as access for the previously uninsured nonelderly improved, it’s not a forgone conclusion that’s a bad thing in general (though it could be for specific patients). We have underuse (by the uninsured) alongside overuse (by the insured, including Medicare beneficiaries). A redistribution of use from the latter to the former could therefore be efficiency enhancing.

Two other recent papers relied on Behavioral Risk Factor Surveillance System (BRFSS) data to compare outcomes in Massachusetts to those in other states, pre- and post-reform. Courtemanche and Zapata (ungated working paper here) found that the Massachusetts reform, relative to other experiences in other US states, was associated with an increase in the probability of individuals reporting excellent or very good health, improvements in self-reported physical and mental health, functional limitations, joint disorders, and body mass index. Though they estimated far more sophisticated, multivariate models, the thrust of their findings is evident in this descriptive chart:

MA reform 1

You can read more about the Courtemanche and Zapata study in this post by Ezra Klein.

The other recent paper on Massachusetts’ health reform and based on BRFSS data is by Philip Van der Wees, Alan Zaslavsky, and John Ayanian. They examined a complementary set of outcomes, finding that, relative to other New England states, the reform was associated with significant relative increases in measures of access and affordability of care, as well as rates of Pap screening, colonoscopy, and cholesterol testing. Again, though they used sophisticated statistical models, you can see the qualitative nature of their results graphically:

MA reform 2

Of course, none of this was ascertained by randomized trials. The BRFSS is a survey, and a large one, but one must always add a pinch of salt when interpreting observational studies. It’s of some comfort that the results of Courtemanche/Zapata and Van der Wees et al. reinforce each other, are consistent with prior work, and are robust to the many specification variations explored by the investigators. To date, the evidence suggests the Massachusetts reform improved the health of the state’s residents.




Jimbino January 14, 2014 at 4:47 pm

Well, taxing non-gamblers to support casinos would enhance the casino participation rate and the casino experience, no doubt. But what does it do to the non-gamblers, not to mention the public in general?

Crazy analysis.

Dennis Byron January 14, 2014 at 6:15 pm


Right, particularly if you let the gamblers go to the casinos and make the non-gamblers buy his chips. All the research mentioned here about Massachusetts is cherry picked to support the blog author’s politics. I can find research that leans the other way just as easily. Unfortunately almost all of the author’s and the other side’s research is old stuff.

But the last two graphs are interesting. It is from part of a recent interpretation of mostly old data. This Van der Wees, et al research was funded by by the left wing Commonwealth Fund and is based on self-reported data; no real-world research would ever depend on such a source. In the main manuscript of the analysis, the yearly data in the above graphs are apparently averaged… or aggregated… can’t tell which. Taking either approach is a bad way to analyze data (time-series analysis — year to year to year, or week to week to week, or… — is the way research should be best compared).

The data — although it is from the Behavioral Risk Factor Surveillance System as noted — does not report on behavioral factors. You just can’t use data collected for one purpose for another; in this case the BRFSS data collectors specifically say

“The basic philosophy was to collect data on actual behaviors, rather than on attitudes or knowledge.”

Yet the Commonwealth Fund bases its whole strained theory on attitudes (like “did you feel good in the last 28 days?”)

The data is cherry picked (Massachusetts is compared against some states but not others, some years are used but not others, and so forth)

If you actually wanted to compare all of this bad data to prove that insurance made some kind of difference in health, you would compare the trough and peak of insurance rates (not shown on these graphs). That data is in the full report however and when you look it up and compare against this data you find — as would be expected, there is no significant difference over the years. In summary the Van der Wees et al analysis illustrates that in its short now ended existence, RomneyCare insurance made no real difference in people’s health. That would be as expected because only a very small percentage of Massachusetts residents were subscribers to RomneyCare insurance or signed up for the expanded Medicaid that came with the RomneyCare law. In fact what the author of this blog is not telling you is that the conclusion of Van der Wees, et al, was:

“These findings may stem from expanded insurance coverage as well as innovations in health care delivery…”

Which means the result could have been because we here in Massachusetts are younger, richer, less white, more ethnically Irish, eat beans and hot dogs on Saturday, and so forth than in the other states cherry picked for the research.

Comments on this entry are closed.

{ 1 trackback }