Politics & Government

A mixed record

Many studies tout the quality of the Department of Veterans Affairs health system based on what researchers call the "process" of health care; for example, how regularly important tests are performed. The VA does well on those measures.

There are fewer studies documenting the outcomes of the agency's health care. That is: Do patients in the VA system do better or live longer? There, the record is mixed.

-A 2004 study, covering patient care from the same period as the RAND Corp. study the VA often cites to prove its quality, examined survival rates among patients treated in either VA or Medicare-funded hospitals after heart attacks. It found that patients in VA hospitals had a higher death rate than those in the Medicare system, even after accounting for the severity of each veteran's illness. Mary Beth Landrum, an associate professor at Harvard Medical School who led the study, said it had prompted the VA to change its cardiac care procedures.

Using slightly different methods, the VA said its own analysis of heart attack mortality found no difference between the VA and Medicare systems. That analysis hasn't been published in a peer-reviewed journal, although the agency said it planned to submit it to a journal soon.

Landrum said the different results showed how difficult it was to research outcomes in health-care systems that record information in different ways.

-A separate study, from 2004, compared diabetes care among VA patients with those in private managed-care programs. It determined that the VA was superior in process issues, but the record was mixed on outcomes issues. For example, patients in the VA system had better control over their cholesterol and one type of hemoglobin than those in managed-care systems, but the VA and managed care were the same at controlling blood pressure.

-A 2005 study compared veterans who got coronary artery-bypass surgery at VA hospitals with those in private hospitals; the death rates were about the same.

-A 2006 study compared older patients in the VA system with those in the Medicare managed-care system. It started with two groups - one from the VA, one from the Medicare system - and determined how many of them died from 1998 to 2004.

After adjusting for health status, the VA patients had a lower death rate overall than those in Medicare. Donald Miller, a VA researcher and Boston University professor who co-authored the study, said the differences were big enough to be meaningful for the average veteran using the system.

But there was a caveat in the study: Once patients were pegged as either VA or Medicare users, those labels stuck; researchers didn't track where the patients received their care. Another study, out in March, found substantial overlap between the systems: Forty-six percent of veterans in the study who were older than 65 used the VA and Medicare for outpatient care, and only 18 percent used just the VA.

-In a statement to McClatchy Newspapers, the VA made the case that a measuring system from the Joint Commission, which accredits health care organizations, demonstrates the VA's superior care.

For about two dozen measures - mostly process measures, but some outcomes measures - the VA compared its performance with the average performance of other accredited hospitals. In its statement, the VA said its hospitals equaled or beat other hospitals 91 percent of the time.

What the agency didn't factor in was all the times there were too few cases to make a comparison. Of 2,738 comparisons, the VA hospital was better than the average accredited hospital about 21 percent of the time, it was the same about 36 percent of the time and it was worse about 6 percent of the time. In the final 38 percent, there were too few patients to make a comparison.

On the specific issue of heart-attack care, VA hospitals and other hospitals were essentially even. On the crucial measure of death after a heart attack, the results found VA hospitals better 0 percent of the time, the same 59 percent of the time and worse 1 percent of the time. In the remaining 40 percent, a comparison couldn't be made.

-Miller, the Boston researcher, participated in another study that detailed a possible problem with any attempt to assess the care at VA facilities.

That study, published in 2004 in the American Journal of Managed Care, looked at more than 200,000 veterans with diabetes who were enrolled in the VA and Medicare.

Researchers found that the systems' medical records differed. Looking only at VA records, for example, indicated that 2,300 veterans had had amputations. But combining VA with Medicare records showed that there were 3,900 veterans with amputations.

The bottom line: VA records gave only a partial picture of a veteran's total health care. Given such record-keeping gaps, the study concluded that "comparisons among national systems may not be possible until there is a single electronic medical record."