When choosing a hospital, patients need and deserve meaningful and transparent information about the quality of care an institution will provide.
But that is not what they got when the CMS recently released its new hospital overall star ratings on the Hospital Compare website.
Instead of providing useful information, the new ratings paint a confusing and conflicting picture of the quality of U.S. hospital care because of a deeply flawed methodology that ignores important differences in the patient populations and the complexity of conditions that different types of hospitals treat.
The ratings are particularly misleading about the quality of care provided at the nation's major teaching hospitals—institutions that provide the most advanced care to some of the sickest and poorest patients. These institutions are nationally known for setting the standards of care, teaching the next generation of doctors, and pioneering new treatments for the most complex medical conditions. However, according to the CMS analysis, only one major teaching hospital received five stars; nearly 90% rated three stars or below. How is this possible?
Because of the wide array of complex and common procedures performed at major teaching hospitals, these institutions reported more measures to the CMS. In fact, the agency used more than 60 measures to calculate the ratings for major teaching hospitals compared with as few as nine measures for other hospitals.
The consequences of providing more data? According to an analysis by the Association of American Medical Colleges, the lower the number of measures a hospital reported, the more likely that institution was to receive a higher star rating. Hospitals that reported on only 60% of the metrics or fewer received almost half of the five-star ratings. Of this group, only four received a one-star rating, while 46 received five-star ratings—nearly half of the 102 hospitals that earned the highest rating.
Ratings for two Midwestern hospitals show how misleading the ratings can be. When hospital A, a major teaching hospital, reported on seven quality domains (mortality, readmissions, timeliness of care, efficient use of medical imaging, effectiveness of care, safety of care and patient experience) with “same as” or “above” national average ratings for all, it received a lower rating than hospital B that provides more limited care and reported only three of the seven measures (effectiveness of care, safety of care and patient experience).
The ratings also fail to account for socio-demographic factors such as patients' education, race, economic status and regular access to medical care. All have a tremendous impact on health. As a result, many urban hospitals that provide stellar patient care and pioneer groundbreaking therapies, in addition to caring for large numbers of poor patients, received fewer stars than hospitals in affluent suburbs that treat fewer complex patients.
An analysis of Medicare disproportionate-share hospital payments shows a troubling pattern in the star ratings. In general, hospitals that serve a higher than average number of low-income patients received lower ratings than hospitals that received lower levels of disproportionate-share payments. Of the 334 hospitals reporting the highest amounts of disproportionate-share payments, there are no five-star rated hospitals and only 23 four-star rated facilities.
The reality is that hospitals cannot be rated like movies. The AAMC and the nation's major teaching hospitals strongly support providing patients with transparent and meaningful information. However, the new star ratings offer an overly simplistic picture of hospital quality that could end up driving patients away from the very hospitals where they could receive the best care.
These ratings are bad for patients and bad for the hospitals that care for them. We urge the CMS to re-do the star ratings to address design flaws and account for socio-demographic factors and the complex care many hospitals provide. Only then will these ratings begin to be meaningful for the patients they are intended to help.
