The Ugly Truth About Patient Satisfaction

Years ago, the Disease Management Care Blog crossed swords with a clinic administrator who was unsatisfied with its Department's patient satisfaction metrics. The DMCB argued sometimes having to say "no" to patients' demands for easy cures, specialist referrals, sophisticated imaging studies or off-formulary drugs could result in low satisfaction rates. It was about then that the DMCB was informed that the patient satisfaction surveys were going to be tied to portion of its income.

Administrator: 1, DMCB: 0.

That's why years later, the DMCB feels partially redeemed by an interesting peer-reviewed publication that examines the lack of any relationship between patient satisfaction and patient centered care.  In case you think they're both the same, think again.

Writing in the July 11 issue of JAMA, authors Joel Kupfer and Edward Bond point out that the former is a consumerist concept that compares service or product delivery to customers' expectations, while the latter is a medical concept that uses patient values to guide medical decision making.

It's been long-known that exceeding expectations leads to customer loyalty which, in turn, leads to greater profits. That's one of the reasons why business-minded hospital administrators typically rely on satisfaction surveys as a key metric of success. Despite their wide use in health care settings, however, Drs. Kupfer and Bond find that "satisfaction" is a poorly researched concept that has little correlation with the important domains of quality, safety, effectiveness, efficiency or equitability. For example, it's possible for over-testing and over-prescribing to give patients the false impression of action and progress. In addition, because patients don't usually bear the full cost of medical care, their opinions don't have to reconcile price and product, like they would for a car or a bottle of shampoo.

What did the DMCB learn, other than, once again, it was right all along?

1. Building a patient centered medical care practice won't necessarily generate high patient satisfaction. Conversely, pursuing high satisfaction rates is not the same as initiating patient centered medical care.

2. "Patient satisfaction rates" are a Ver. 1.0 primitive window into one slice of patient experience that, by itself, ultimately offers little insight about overall quality.  Patients and their doctors deserve better.

Does Lots of Tests for Patients Generate Lots of Satisfaction? Maybe Not.

There is nothing like a blood test or an x-ray to build patient confidence in a diagnosis. Being able to point to that laboratory report or CAT scan as proof positive that a disease was indeed present (Mrs. Smith, the pregnancy test confirms it!) or absent (Mr. Jones, the scan shows there is no brain tumor!) remains a time-honored part of the doctor-patient journey from uncertainty to reassurance. Never mind Mrs. Smith’s gravidity would be obvious even to a professional wrestling referee or that the likelihood that Mr. Jones’ 20 year history of dizziness could be ascribed to any tumor anywhere is remote, it’s **The Test** that seals the deal.

Is it really that simple though? This study (provocatively titled "Is patient satisfaction influenced by the intensity of medical resource use by their physicians?") by Bryan Down, John Kralewski, Amer Kaissi and Sally Irrgang from the University of Minnesota, Texas’ Trinity University and Blue Cross Blue Shield of Minnesota in the American Journal of Managed Care says otherwise. Using a validated 2002 patient survey that assessed satisfaction with a) overall primary health care, b) adequacy of time spent with the physician and c) likelihood of recommending the clinic to others, the authors sought to correlate the satisfaction with 2001 insurance network and claims data for 62 practices with 2800 patients.

While the actual calculation of claims expense was complicated (‘dollar-weighted resource use’), the bottom line was that there was no statistical correlation between more intensive medical resource use and a) overall patient satisfaction or b) the likelihood of recommending the clinic to others. What’s more, higher levels of use were negatively correlated with satisfaction over time spent with the physician. In fact, the degree correlation looked striking at -0.92. Interestingly, high resource use was not associated with being a busy clinic, prompting the authors to suggest that busy clinics are not trying to substitute tests for meaningful hands-on patient care – something else was going on.

The Disease Management Care Blog isn’t sure what else is going on and the authors didn’t spend a lot of effort discussing the implications of their findings, other than stating that more research is needed. The DMCB wonders if some physicians try to substitute tests for ‘quality time’ independent of how busy they are. The reverse may also be true: patients who are dissatisfied with any amount of time they’re getting with their doctor may end up demanding additional tests. There were two weaknesses in the research design too: 1) in general, these Minnesotans were highly satisfied with their care and Minnesota is the land of low variation to begin with, which may have obscured the ability of the researchers to detect a correlation and, 2) if multiple correlations are performed, one may randomly show up as spuriously ‘significant.’

Whatever the reason, public and private insurers will view these data as one more reason to do something about unsustainable upward trend in medical testing, including high cost radiology procedures. While physicians may claim that their patients ‘demand’ the testing, it didn’t turn up in this research. When policy makers get around to the notion of impeding physician's ability to order tests ad lib, they may want to use this paper to point out that patient satisfaction levels may remain unchanged or - assuming there's a causal relationship - even go up.