Counterpoint

We know some of the people involved in the Dartmouth Atlas Project, and we think their analysis is important. Even so, using 2005 Medicare data to inform comprehensive payment reform is inadequate.

As such, we are surprised and dismayed at how policymakers are using the findings as the map for healthcare reform in Washington, D.C. We are also frankly appalled at how The New Yorker article by Dr. Atul Gawande has seemingly become the guidepost of reform for policymakers. The reason is that the conclusions that The White House and much of Congress have drawn from The New Yorker article are, at best, suspect and, at worst, completely wrong. Reengineering 20% of the economy is a large task, in our view, and getting the facts straight is important.

So, what have we done? Instead of using an “Atlas” to analyze McAllen and El Paso, we suggest using a “GPS” to triangulate the position that hospitals played in overall excess cost and utilization. Doing so provides some critical facts that The New Yorker failed to report.

At first blush, McAllen and El Paso are quite similar:

• 2008 populations are within 1% (752,020 for McAllen vs. 759,868 for El Paso).

• Median age of the population is similar, at 28.2 years for McAllen compared to 30.6 years for El Paso.

• Per capita income for each market is depressingly low, with $12,276 for McAllen and $16,838 for El Paso (making El Paso 37% wealthier, as suggested by the physicians in McAllen).

• Medicare hospital utilization rates are similar, with 28% Medicare utilization in McAllen and 30% Medicare utilization in El Paso.

• Total hospital utilization (i.e., all-payer data) when compared to the population were similar in calendar year 2007 (the most current year that all payer data is available), with 12% hospital utilization in McAllen versus 10% hospital utilization in El Paso.

• Each market has 2% workers’ compensation hospital utilization.

• Per capita hospital utilization is similar, with a rate of .48 patient days per capita in El Paso compared to .53 patient days per capita for McAllen.

• McAllen cost per case is 5.4% lower than El Paso, and McAllen’s average length of stay is 9.6% lower than El Paso.

Based on these similarities, McAllen is in many ways a more desirable option for hospital care.

So, what about the real differences between McAllen and El Paso?

Overall, and not just for the Medicare and Medicaid population data (which were central to the Atlas and The New Yorkerperspective), McAllen’s average cost per case is $315.00 less than in El Paso, representing in total $23.6 million in incremental costs that could be saved if all of the El Paso cases had been treated in McAllen hospitals. For policymakers who are concerned about the price paid by the uninsured, the average charge per case is $7,841 more in El Paso than in McAllen.

Importantly, the “excessive” costs attributed to McAllen do not occur in McAllen, or even in Hidalgo County. A full 6% of McAllen residents left McAllen for care to other markets such as Brownsville, Houston, San Antonio, Corpus Christi and Dallas! A total of $283 million in charges migrated away from McAllen, yet those costs are attributed to the population and demographics of the beneficiaries living there. As a result, the Dartmouth Atlas analysis overestimates the costs attributed to McAllen. As a comparison, $63 million of charges out-migrated from El Paso to other Texas hospitals during the same period (the all-payer analysis does not reveal out-migration to any other states; El Paso is closer to Phoenix than Dallas).

What about the important things, like quality? The March 2009 release of the Hospital Value Index™ reports McAllen’s average index score at 42.76 with El Paso’s being 43.83, just over one basis point difference. This indicates that the markets are nominally different on quality, core process measures, mortality, patient safety and patient satisfaction and experience. Shorter lengths of stay, lower costs, and lower mark-ups for charges on patient bills make for a more desirable profile of McAllen hospitals than El Paso.

In summary, the most current all-payer data (2007) simply does not support The New Yorker piece, which was partially based on 2005 Medicare data from The Dartmouth Atlas. For both McAllen and El Paso, the cost per beneficiary would decrease if the beneficiaries did not leave the market.

These markets have a great deal in common, but critical differences not discussed in The New Yorker. We are reminded how important it is to “follow the money”, yet without the anecdotes about what is going on in McAllen, the empirical data report that the hospitals in McAllen aren’t the problem.

We think that there are several important questions that arise:

• Could an entire industry be led astray by the miscalculations of Medicare spending delivered by a half dozen hospitals in McAllen and El Paso?

• Should policymakers draft legislation to reform the provision and coverage of healthcare based solely on (old) Medicaredata?

• Is the nation going to allow a handful of well-meaning, but uninformed, policy-makers to reform healthcare based on the view of an article in The New Yorker?

Heaven help us if we do…

(Source: Texas Health Care Information Collection, TX Public Use Data File, State Hospital Data, Calendar Year 2007)