In theory, being able to see which hospitals and physicians best deliver care for various conditions would likely help individuals choose where to receive their health care. But in practice, public reporting of outcomes data can be problematic because it generally doesn’t make the distinction between clinicians who are treating high-risk vs low-risk patients. Thus, a higher or lower rate of adverse events for a procedure may be a reflection of the clinician’s skill or may be mostly the result of the types of patients (higher-risk or lower-risk) that a clinician treats.
A study appearing today in JAMA highlights this conundrum. The researchers wanted to see whether public reporting affects the use of percutaneous coronary intervention (PCI) and death rates in patients admitted with an acute heart attack. To study this, the researchers used data for Medicare patients admitted between 2002 and 2010 with an acute heart attack in 3 states with mandatory public reporting (New York, Pennsylvania, and Massachusetts) and compared it with data for similar Medicare patients treated in several nonreporting states.
They found that overall, PCI use in states with public reporting was 18% lower than PCI use in nonreporting states, while 30-day death rates for both patients who had the procedure and those who did not remained the same. But when they looked specifically at outcomes for patients with ST-segment elevation myocardial infarction (STEMI), a more dangerous form of heart attack, those in public reporting states were 27% less likely to receive PCI and were 35% more likely to die than comparable patients in nonreporting states.
Lead author Karen E. Joynt, MD, MPH, of the Harvard School of Public Health in Boston, discusses her team’s findings:
“We did this study because there is a concern that public reporting might reduce access to PCI. We used acute myocardial infarction patients because they should have an indication for PCI. It’s not 100%, but guidelines suggest they would benefit from PCI. So we compared New York, Pennsylvania, and Massachusetts to states without mandatory reporting.
“There’s plenty of evidence showing that there’s overuse of PCI, but it’s difficult to tease out of our data. The non-STEMI group in our sample has much more of a ‘soft’ indication for PCI; our rates reflected that, and there may have been folks in that group in whom it was overused. My suspicion is that the ‘overuse’ in the cardiogenic shock and cardiac arrest group is mostly futility: cardiologists are doing procedures that they really hope will help, but the patient is so sick that no matter what we do, they probably aren’t going to make it. That is technically overuse, but it’s not overuse in the usual sense of the word.
“We need to figure out how to do transparency right. I believe in it philosophically. To suggest that patients don’t want or need data for making their medical decisions is misguided. We should not discard public reporting but tweak it to make it better as it plays out on the front lines of medical care. We have to make sure we’re doing risk adjustment correctly to take away some of the concerns clinicians have that they’re not getting credit for riskier patients they’re taking on. Everyone gets more risk adverse, and there may be unintended consequences.”