As part of its push to pay providers based more on their outcomes and less on just delivering a service, the Center for Medicare Services created a series of star ratings for providers based on various supposed measures of quality. I say supposed because there isn’t always good research to support a connection between the measure and actual improved health outcomes for patients. In any event, one such Medicare rating system is used for hospitals and is supposed to help patients select high-quality facilities. The measures also end up being a factor in the hospital’s reimbursement through the value-based purchasing system and through a penalty if quality data is not submitted. A study carried by Health Affairs, however, suggests that there are flaws in the way CMS ranks all hospitals together. (HA Article) Part of the problem is that large urban facilities are being compared with small rural ones and another part is that not all hospitals report on all measures, in fact hospitals may not have enough data to be able to report on some; they simply lacked the volume to make a meaningful calculation. For 2017 there were 57 measures composing the star system, grouped into seven categories–mortality, safety of care, readmissions, patient experience, effectiveness of care, timeliness of care and efficient use of medical imaging.
There were 3692 hospitals that got a star rating in 2017. The hospitals fell into roughly three clusters, basically divided by number of measures they reported on. The cluster that reported on the most measures accounted for 36% of one-star hospitals, 27% of two-star hospitals, 18% of three and four-star hospitals and 33% of five-star facilities. The cluster that reported on the fewest measures, had only 10% of one-star hospitals, 20% of two-star institutions, 33% of three-star facilities, 36% of four-star ones and 22% of five-star facilities. So you can see that there is a clear difference in where you might rank depending on which cluster you are in. The cluster one hospitals tended to be large and busy and were in metropolitan areas. The third cluster was comprised of more smaller and rural facilities. If different hospitals report on different measures, it pretty obviously seems inaccurate, and unfair, to compare their performance. And it isn’t providing patients with truly useful information. As the authors suggest, it would seem more useful to put different types of hospitals, or hospitals reporting on different numbers of measures, into separate comparison groups.