In other words, a patient in a large, high-volume hospital that is highly rated under Lieberthal's model might dislike the noise and bad food but will survive a life-threatening heart attack. "Based on this study the hospitals that have the best survival outcomes are not doing the best job of satisfying patients," Lieberthal says.
Lieberthal believes his method could be used by the federal government as a way to correlate the different quality measures that they collect and put them into a single quality score. "For example, right now Medicare has a model that they use to do mortality risk adjustment. The hospitals that tend to see sicker patients get an adjustment for that and the mortality scores that are reported in Hospital Compare," he says. "We would definitely see a value in Medicare applying this model to all of the data they generate, not just the data they put in Hospital Compare but in their much larger set of claims and other data that they generate as a large health insurer."
The study was funded by the Society of Actuaries.
"I was commissioned by them to develop a way to predict the quality of hospitals so that insurance companies and hospitals could plan their reimbursement rates," Lieberthal says.
"We see an implication of this study using the overall scores that we developed to pay more for the hospitals that were better or include the hospitals that were better in preferred provider networks. For the hospitals that didn't do as well, and some of that was because of these satisfaction measures, insurance companies might want to consider not including them in a preferred provider network."