In response to concerns about incomplete and misleading data, 60 senators have sent a letter to the Centers for Medicare and Medicaid Services (CMS) requesting that the agency delay releasing overall star ratings for hospital quality on April 21.
In their letter to acting CMS administrator Andy Slavitt, the senators said they were concerned that the current star ratings system may not accurately take into account hospitals that treat patients with low socioeconomic status or multiple complex chronic conditions.
“We are concerned that the star ratings system may be misleading to consumers due to flaws in the measures that underpin the ratings,” the senators wrote. “Many prominent hospitals that are in the top echelon of other quality rating reports, and handle the most complex procedures and patients, may receive 1 or 2 stars (out of a possible 5), indicating that they have the poorest quality in comparison to all other hospitals. As the Medicare Payment Advisory Commission (MedPAC) and other researchers have noted, these measures are not appropriately adjusted for socioeconomic status and patient complexity.”
In addition, the senators pointed out, the CMS has revealed insufficient details regarding the methods used to determine its star ratings and has not provided hospitals with the data used to derive the ratings. As a result, the senators heard from hospitals that had been unable to replicate or evaluate the CMS’s work to ensure that its methods were accurate or fair. For that reason, the senators requested that the CMS delay the release of the star ratings to provide the necessary time to more closely examine the methods involved, to analyze their effect on different types of hospitals, and to provide more-transparent information regarding the calculation of the ratings to determine their accuracy.
The 2016 star ratings and the methods used to create them have long been a source of controversy; in a January report, the CMS offered a complex explanation of how it arrives at its rating scores. The report states:
“The measures [of quality] are first selected based on their relevance and importance as determined through stakeholder and expert feedback. Next, the measures are standardized to be consistent in terms of direction and magnitude. These standardized measures are then used to construct a single hospital summary score in a two-step process. In the first step, the measures are organized into seven groups by measure type. For each group, CMS constructs a latent variable statistical model that reflects a dimension of quality represented by the measures within a given group. Each of the seven statistical models generates a hospital-specific latent variable, or group score, as a prediction of that hospital’s performance based on the available individual measures scores within a given group. CMS uses these seven group scores to organize hospitals into one of three group performance categories (above, same as, and below the national average), providing additional detail for patients and consumers for comparing across the seven groups. In the second step, a weight is applied to each group score, and all available groups are averaged to calculate the hospital summary score. Finally, to assign a star rating, the hospital summary score is organized into one of five categories using a clustering algorithm.”