A mandatory cognitive testing program for clinicians aged 70 and older at Yale New Haven Hospital found that 12.7% had “cognitive deficits that were likely to impair their ability to practice medicine independently.”

In a report published in this week’s Journal of the American Medical Association, the study by Leo Cooney, a geriatrician at Yale School of Medicine, and Thomas Balcezak, chief medical officer at the hospital, reported that none of the clinicians had previously been reported for performance problems.

All of the doctors who performed poorly either voluntarily quit their practices or agreed to work in a “closely proctored environment.”

Physicians typically do not have a required retirement age. According to an editorial that accompanied the study, 15% of practicing doctors were over age 65 in 2017. Some cognitive functions tend to decline with age and the risk of dementia rises.

Monitoring the cognitive abilities of older physicians has been controversial. Some older physicians contend that it is age discrimination to focus only on older practitioners when younger doctors can also have thinking problems. Older doctors also disproportionately provide care in rural areas, which already are struggling with clinician shortages.

But health systems worry about their responsibility to patients if doctors in the early stages of dementia make mistakes. Several, including Scripps Health Care, Stanford Hospitals and Clinics, and Penn Medicine, have age-based screening programs, the editorial said.

Many Americans now remain healthy, both physically and mentally, long past the traditional retirement age, and there is a trend toward working longer.

Yale began requiring clinicians aged 70 and older to participate in cognitive screening in 2016 as part of its re-credentialing process. The study analyzed results for 141 clinicians with an average age of 74. Of those, 88.7% were physicians. The rest were nurses, dentists, psychologists, and other medical professionals.

Yale is keeping secret the precise battery of 16 short tests so that clinicians cannot prepare in advance. The tests focus on multiple cognitive domains that help people “think, monitor, and adapt,” the study said.

The process itself was confidential and results were reviewed by a special committee. The group looked at problems likely to affect a physician’s job, so problems with visualizing spatial relationships were more important for doctors who perform surgery and other procedures.

Fifty-seven percent of the candidates did fine on the test and proceeded with credentialing with retesting in two years. About one-quarter went on with the process but had to be rescreened in one year because of “minor” abnormalities.

The rest either underwent more comprehensive testing or were counseled about next steps.

In the report, Cooney and Balcezak said the most important steps in the process were the selection of the neuropsychologist who administered the tests and of the members of the review committee.