Want to know what's happening next in healthcare?

MedCityNews is the leading online news source for the business of innovation in healthcare.

“I like the diversity of topics that MedCity News covers – ranging from devices to policy to start-up companies. Always something interesting to find.”

Robert J. Szczerba, Ph.D., CEO, X Tech Ventures

Sign up for our daily newsletter

Study finds little correlation between patient satisfaction scores, quality of care in hospitals

8:40 am by | 16 Comments

St. Vincent Indianapolis Hospital, photo by Denise Cox

When it comes to hospitals, patients apparently don't know what's good for them.

After crunching data from 4,655 hospitals, a Thomas Jefferson University health economist concluded that the best hospitals were not the ones with the highest satisfaction scores. Instead, the best hospitals, which tended to be bigger and busier, got more lukewarm ratings from patients.

Robert Lieberthal doesn't know why, but said he's not the first to find little correlation between satisfaction scores and the important stuff, like whether you'll come out alive and stay that way.

Medicare publishes satisfaction ratings for individual hospitals and uses them to some degree when it sets payment rates, said Lieberthal, who works in the Jefferson School of Population Health. Asked whether that's a good idea, he said, "I think the primary purpose is to get people well . . . and I would like to see Medicare focus on refining or improving those measures and paying for hospitals that prove that they have the best clinical quality."


His work was funded by the Society of Actuaries, who are interested in quality measures because insurance company payments increasingly are tied to quality. They wanted a relatively simple, stable rating system for hospitals.

Lieberthal tested PRIDIT, a statistical model originally developed to flag car-insurance fraud. It can reduce a lot of complex data to a single number.

Using data from Medicare's Hospital Compare program and the American Hospital Association, Lieberthal said he found that PRIDIT could give the actuaries what they wanted. Key measures were how well patients with heart failure, heart attacks, and pneumonia fared.

Typically, the hospitals with the highest quality were the largest, hospitals like, say, Thomas Jefferson University Hospital. Lieberthal said he didn't break down the numbers for individual hospitals, so he doesn't know how Jefferson would rate.

Hospitals on the low end were small and not part of a bigger health system. They were not teaching hospitals, not accredited by the Joint Commission, and lacked emergency departments.

One of the things that Lieberthal learned while trying to figure out what correlated with quality was that patients' answers to standard questions about their experience were not helpful. Patients are asked whether nurses and doctors communicated well, whether their pain was well controlled, whether the facility was clean, whether their room was quiet. The answers could be always, usually, sometimes or never.

Lieberthal found that the highest-quality hospitals got more "usually" answers. High rates of "always" answers were negatively correlated with medical quality.

"The hospitals that are doing the best job of satisfying patients are not the ones with the highest quality," he said.

He did not have data on two things that patients often fuss about: parking and food.

While Lieberthal thinks insurance companies will find this information useful, he's not so sure what patients can do with it. They can look at the Hospital Compare charts about medical quality measures themselves, but even Lieberthal finds those to be a "very challenging resource."

He suggested choosing a hospital the old-fashioned way: find a doctor you trust and ask for a recommendation.

(c)2013 The Philadelphia Inquirer

Copyright 2014 MedCity News. All rights reserved. This material may not be published, broadcast, rewritten or redistributed.


Visit website | More posts by Author


@hhask I'm less interested in the results of these studies & more interested in the wording of the questions. Did they ask what matters?


This is veryinteresting, what could be a good framework of measuring the quality of care with respect to 

( in extreme cases) one staying alive post treatment and continuing to stay that way in future (at least in the context of the treated problem). Who could clinically vet the quality of treatment results because it depends on so many variable factors: Treatment given during admission, treatment course followed post discharge, keeping regular follow ups, hospitals ability to keep costs lower ensuring superior clinical quality. Looking at these parameters does the need to have multiple quality parameters arise ? For example one parameter which is more focused on immediate results/ discharge time outcome based and one for a longer term.


Are we looking at the right data & where are we headed? Concierge care or outcomes driven quality - is there a perfect amalgam? @Dermdoc

Hear the latest news first

Get our daily newsletter or follow us.