Be wary of doctor-rating sites: Column

The public can rate almost everything on the Internet today: books, hotels, restaurants — and even doctors. But while your chances of getting a great meal at a 5-star restaurant are pretty high, receiving excellent care from a 5-star doctor is less certain.

Doctor ratings generally focus on more subjective issues, such as patient wait times, time spent with the doctor, and physician courtesy. Those are obviously important issues, but they paint an incomplete picture. Doctors with stellar interpersonal skills may not be the best at controlling patients' blood pressures or managing their diabetes.

Moreover, a doctor's quest for ratings perfection can influence medical decisions, since patient satisfaction increasingly affects a doctor's salary. According to the management consulting firm Hay Group, two-thirds of physician pay incentives are based on patient satisfaction scores. And Medicare withholds as much as $964 million in payments to hospitals who fail to meet various quality metrics, with patient satisfaction being a significant component.

But doing what's best for patients won't necessarily make them happy. Denying antibiotics for viral infections or saying no to routine MRIs for patients with back pain are both sound medical decisions, but can anger patients. Those same patients can vent their frustration by giving their doctor a poor rating. It's no wonder that many physicians give in to patient requests. In a survey by Emergency Physicians Monthly , 59% of emergency physicians said patient satisfaction surveys increased the amount of tests they ordered.

In another survey by the South Carolina Medical Association , nearly half of physicians said that pressure to improve patient satisfaction led them to inappropriately prescribe antibiotics or narcotics. In fact, Sens. Dianne Feinstein, D-Calif., and Charles Grassley, R-Iowa, wrote a letter to Marilyn Tavenner , administrator of the Centers for Medicaid and Medicare Services, saying that "there is growing anecdotal evidence that these (patient satisfaction) surveys may be having the unintended effect of encouraging practitioners to prescribe opioid pain relievers (OPRs) unnecessarily and improperly, which can ultimately harm patients and further contribute to the United States' prescription OPR epidemic."

These extra tests and treatments are expensive and can hurt patients. A landmark study from JAMA Internal Medicine analyzed more than 50,000 patient satisfaction surveys, and while the data pre-dates online ratings, it found that patients who were more satisfied with their doctors had higher health care costs, were hospitalized more frequently, and had higher death rates compared to less satisfied patients. Patients who receive more drugs and tests are exposed to their harmful side effects and complications.

Now, I'm not saying physicians shouldn't be graded by patients. But these online ratings need to be complemented with objective measures of medical care, like a surgeon's operation complication rate, for instance. The major physician rating sites such as Healthgrades, Vitals, Yelp and RateMDs do not include individual physician metrics.And Medicare's Physician Compare does not include subjective patient reviews. Physician ratings that do not include a more holistic representation of doctors must not be used to determine how physicians are paid.

And patients may find that doctors who have mixed reviews may actually provide better care because these physicians occasionally say no to patients.

Kevin Pho, a primary care physician in Nashua, N.H., blogs at MedPage Today's KevinMD.com and is a member of USA TODAY's Board of Contributors.

In addition to its own editorials, USA TODAY publishes diverse opinions from outside writers, including our Board of Contributors .To read more columns like this, go to the opinion front pageor follow us on twitter @USATopinionor Facebook.