Data Analysis Finds Hospitals in Low-Income Areas Scored High in MI Care, Even if Patients Had Poor Outcomes

A study presented Monday at the 64^th Annual Scientific Session of the American College of Cardiology (ACC) found that the quality of the care for myocardial infarction (MI) was just as good if not better in hospitals serving low-income areas than it was elsewhere in the United States, even if patient outcomes did not always reflect those high standards.

The results could have policy implications, as CMS moves to tie Medicare reimbursement to readmission rates and patient outcomes. The study attaches data to a complaint raised by hospitals that serve the poor: strict attention to outcomes without accounting for how long-term poverty affects health will cripple those institutions that most need federal dollars.

“Despite getting good care, some patients may be prone to poor outcomes,” said Jacob A. Udell, MD, MPH, of the Women’s College Hospital and Peter Munk Cardiac Centre, University of Toronto. He presented the study, “Socioeconomic Disparities and Quality of Hospital Care after Myocardial Infarction in a National Cardiovascular Data Registry.”

Many institutions affected by CMS’ reimbursement policies are academic medical centers that train physicians; if current policies remain, Medicare penalties will affect medical education programs. Changes are deeply felt in cardiovascular care, because CDC statistics show poverty is associated with higher rates of diabetes, obesity, and tobacco use, which all affect the ability to survive a cardiovascular (CV) event. These are the patients that many academic centers treat, in what Udell called the “risk-treatment paradox.”

“Patients who are going to be the worst off get the best care in teaching hospitals, because they go a great job with acute management,” he said.

Udell and his co-authors took a unique approach, one made possible with the rise of quality metrics reporting: it paired data from the ACC Foundation’s National Cardiovascular Data Registry and overlaid the MI quality scores for 390,692 cases with socioeconomic measures based on US Census data collected by ZIP code. (Use of ZIP code data as a proxy for individual wealth measures is a longstanding demographic tool.)

Specifically, researchers used data collected between July 2008 and December 2013 in the ACTION Registry-GWTG, 1 of 6 hospital registries that are part of ACC NCDR. Researchers looked at 6 quality measures during the acute phase of MI care and 6 different quality measures at discharge. For the analysis, researchers divided the hospitals into quintiles based on the socioeconomic data of their patient populations. Notably, there was greater separation from the lowest to highest quintile by average housing value--$58,450 to $156,900—than there was by income, where the averages in the lowest and highest quintiles were $27,086 and $57,317, respectively.

Data revealed that hospitals in the lowest-income quintile neighborhoods served patients who were least likely to have private insurance coverage and most likely to have diabetes and smoke. These results reflect CDC statistics that show 29.2% of adults who are below the poverty level smoke, compared with 16.2% of adults who are at or above the poverty level. Among adults under age 65, 30% of Medicaid enrollees smoke, compared with 15% of adults with private insurance.

Overall quality scores—which track how well hospitals performed procedures--were very close across all quintiles in the acute care phase. The lowest-income quintile scored a 91.0%, slightly better than the 90.2% scored by the highest-income group. Scores for procedures at discharge were less favorable among all groups, and this area has been a target of CMS. In this area, overall scores for the lowest-income group lagged 70.0% to 76.7% for the middle-income group, which scored highest.

Even though hospitals in the lowest-income areas were adhering to quality standards, in some cases better than counterparts in wealthier areas, overall mortality rates were higher than in most other quintiles (although lower than in the highest-income hospitals). Hospitals in the poorest areas had particularly high mortality rates due to heart failure or bleeding, compared with deaths rates from MI itself.

Low scores on discharge performance by hospitals in the poorer areas will attract CMS’ attention. The movement toward accountable care organizations, which take responsibility for patient health on an ongoing basis, would ask academic centers what they can do with community groups to promote continuity of care for their populations. However, not all lower-income Americans have gained access to coverage under the Affordable Care Act; the study showed the highest number of institutions serving the poor was in the South, where several states have not expanded Medicaid.

A limitation of the study, which drew attention from commentators at the ACC session, was the 17% of patients who could not be analyzed because of missing or incomplete data for the socioeconomic analysis. However, Udell noted that the distribution was random; as a practical matter, this is still the type of data CMS uses to make policy decisions on Medicare.

The good news, the study concluded, is that quality initiatives appear to be working. Hospitals taking part “appear to provide equitable in-patient care to individuals living in the most disadvantaged neighborhoods compared with the rest of the nation,” researchers concluded.