When antibiotics equals ratings

A new study published in JAMA last week (summarized by NPR) is another great example of the obvious negative externalities of prioritizing patient satisfaction scores (i.e. the Yelpification of medicine). It analyzed a large number of telemedicine visits for URI:

Seventy-two percent of patients gave 5-star ratings after visits with no resulting prescriptions, 86 percent gave 5 stars when they got a prescription for something other than an antibiotic, and 90 percent gave 5 stars when they received an antibiotic prescription.

In fact, no other factor was as strongly associated with patient satisfaction as whether they received a prescription for an antibiotic.

The outsized and misplaced importance of patient satisfaction scores is a perfect embodiment of Goodhart’s law, well-paraphrased as “when a measure becomes a target, it ceases to be a good measure.”

If you make patient satisfaction scores a critical target—and they are—you will see consequent mismanagement. This is so blatantly apparent when it comes to urgent care and pain management that, if anything, high satisfaction scores are likely a more meaningful signal of poor care (like in this study when patient satisfaction scores positively correlated with patient mortality).

I used to know a bunch of residents who would moonlight at a doc-in-the-box for-profit standalone urgent care. They were, apparently, told to make the patients happy and provide antibiotics for most URI visits.

Even outside of quality metrics, you need patients to make money, and the “customer” is always right.