In a New Yorker article this week, Atul Gawande revisits McAllen, Texas. In 2009, Gawande pointed out that Medicare spent about $7,000 more per person annually than the national average in McAllen, sparking a flurry of activity to figure out what was behind the high spending and ultimately leading to efforts that dramatically reduced it.
The findings from subsequent research on spending trends throughout Texas, for private health insurance as well as Medicare (including several Commonwealth Fund–supported studies), demonstrate a key lesson: to be effective, efforts to contain health care costs need to be grounded in local conditions.
An in-depth look at cost data for McAllen identified home health and other postacute care services as major drivers of high spending. Shortly after Gawande’s article was published, a research team from the University of Texas School of Public Health (UTSPH) pointed out that Medicare home health use in 2007 was 4.63 times higher in McAllen than in El Paso, a demographically similar area of Texas, and 7.14 times the national average. Moreover, McAllen Medicare beneficiaries were far more likely to be admitted to the hospital and to die there than they were in El Paso. Medicare beneficiaries in McAllen were also more likely to be seen by more than 10 physicians near the end of their lives—a strong indicator of uncoordinated care.
More recent UTSPH research shows McAllen looking much more like the rest of the state: whereas in 2008, per Medicare enrollee spending in McAllen was more than 40 percent above the state mean, in 2012 per enrollee spending was about 16 percent higher. This reduction in spending has been driven largely by reductions in postacute care spending. In 2008, McAllen’s spending on such care was a staggering 97 percent higher than the state average. By 2012, this ratio had dropped to 35 percent.
It’s likely that both local activities and federal health care reform contributed to this change. In 2014, a New York Times editorial by Bob Kocher and Farzad Mostashari highlighted the changes in McAllen, pointing to the proliferation of accountable care organizations (ACOs) created under the Affordable Care Act as a major factor in curbing costs. For example, two ACOs in the McAllen area, Rio Grande Valley ACO Health Providers, LLC, and Rio Grande Valley Health Alliance, LLC, have been highly successful at reducing spending for their Medicare patients (reducing Medicare spending over $26 million below expectations in 2013 combined), in part by focusing on preventive care and better coordinating care, including postacute care. The UTSPH researchers also pointed to new policies governing postacute care at both the local and national levels that may have helped to rein in fraud that was believed to have contributed to the high spending.
But how does the Medicare population in McAllen compare to those with employer-sponsored coverage? In a 2010 Health Affairs article on spending in McAllen, the UTSPH researchers compared Medicare spending for an elderly population to trends for a nonelderly population covered by Blue Cross Blue Shield of Texas (BCBSTX), using El Paso as a comparison group. A key finding was that per-person Medicare spending was 86 percent higher in McAllen than in El Paso, but 7 percent lower for the population insured by BCBSTX, the state’s largest private insurer. This finding made clear that Medicare should not be used as a proxy for the entire health care system.
In more recent analyses, UTSPH researchers looked at Medicare and BCBSTX spending across the entire state of Texas and similarly found that there was not a strong correlation between BCBSTX and Medicare in terms of spending, prices, and utilization. There are areas in the state that are relatively high Medicare spenders, but relatively low BCBSTX spenders (e.g., McAllen and Houston) and vice versa (e.g., El Paso and Austin). Moreover, spending trends look different in BCBSTX and Medicare from 2008 through 2012, with Medicare growing at a much lower rate (1.5% annually) than BCBSTX (4.9%).
These patterns at the local and state levels suggest that private and public spending may be driven by different causes. As McAllen demonstrates and as evidenced by a recent Institute of Medicine report analyzing national health care spending trends, postacute care services are a major driver of spending for the predominantly elderly population covered by Medicare. For the predominantly nonelderly privately insured population, these services are less of an issue since younger people are less likely to need them. Instead, spending is driven more by negotiated prices for inpatient hospital services, because, unlike Medicare, private insurance plans cannot set standardized prices across geographic areas.
Media attention on McAllen helped create an impetus for change—and encouraged researchers to dig deeper into the story. The evidence they uncovered helped identify opportunities to reduce costs and improve care. However, their analysis also shows that what worked for McAllen may not work for other areas, unless there are similar local forces at play.
Texas’ story suggests that policymakers should not use one segment of the population or geographic area to make decisions about how to control spending growth across the health care system. Rather, they should consider a variety of strategies, depending upon local conditions and population needs.