Could your doctor’s age affect how you fare during treatment? Perhaps — though the impact is fairly small, according to a new study.

A team of researchers led by Dr. Anupam Jena of Harvard Medical School looked at a random sample of Medicare data for more than 700,000 hospital admissions from 2011 to 2014. The patients received treatment from whichever physicians happened to be on duty at the time of their admission.

When the team looked up the ages of the nearly 19,000 doctors who handled those Medicare patients, a trend emerged: The older the doctor, the higher the patient mortality rate.


For physicians under the age of 40, patients’ mortality rate was 10.8 percent. That edged up to just over 11 percent for patients treated by doctors in their 40s.

The patient mortality rate was 11.3 percent for physicians aged 50-59, and rose above 12 percent for physicians over 60.


These findings appeared Tuesday in the BMJ.

Why might older doctors have worse outcomes? “There’s a fear that as doctors get further away from residency, they might be out of touch with new technologies and treatments,” said Jena.

The study drew praise from Rebecca Mary Myerson, a health economist at the University of Southern California, who was not involved with the research. She said the analysis appropriately controlled for a number of variables — including the possibility that the sickest patients were assigned to the most experienced physicians on call at the hospital on any given day.

One interesting twist in the results: Doctors who saw a large number of patients did not seem to lose any skills as they aged. Indeed, patient mortality rates were fairly consistent among high-volume physicians of all ages.

Jena suggested that seeing a lot of patients might have the effect of keeping doctors on top of advances in treatment.

“It could be that high-volume doctors are experiencing no decrease in their skills or expertise,” he said. “Maybe low-to-medium-volume doctors just don’t see enough patients to have to keep up. Or maybe those doctors are less knowledgeable, so they see fewer patients. It’s not clear what comes first.”

Dr. Jack Sobel, a Detroit physician not affiliated with the study, said he is not surprised by the findings.

“It’s not that clinical skills deteriorate,” he said. “People over the age of 65 are just not as familiar with the new methods. That’s what gives younger doctors the edge. It’s access to newer technology, and knowing the newer drugs.”

Sobel, who is 74, said he tries to keep his skills current by reading five medical journals a day, in between treating patients and teaching students at Wayne State University, where he is dean of the medical school.

“I happen to be addicted to keeping up to date,” he said. “But I’m not the norm.”

  • As with most such population-based survey studies I suspect there are a lots of gaps in how the information was collected. I also agree with one commentator who stated that older physicians have elderly patients that have stayed with them for many years thus skewing the curve. I’m sure as with most things in life some elderly physicians are less equipped and others are more equipped based on experience. The other thing that I have observed is that older physicians are a bit more thorough in their clinical skills as the newer generation of physicians and providers seems to be more niche and narrow focused. I suspect many things are missed due to the new paradigm of a 10 minute encounter including probably more than 50% spent on the computer. I guess I’m just an old misanthrope.

  • So the study was about patient volumes, not physician age. And also about the addiction of many physicians in the big academic institutions to writing sensationalist articles that are picked up by the lay press, thereby boosting their visibility and careers.

  • Why don’t you do a parallel study in which you evaluate the amount of testing done by the young doctors versus the older doctor. You can then dovetail that with the article recently suggested that says more testing result in higher mortality. I believe this would offset The small percentage difference between young and old doctors and the outcomes would be a that young doctors cause more mortality Try it.

  • As a senior ophthalmic surgeon I am routinely called to another OR where a fairly skilled younger surgeon has gotten themselves into a situation they can’t handle. What would the patient pay to have me there, a 71 year old with tons of surgical experience who has seen every complication, lots of them and now knows and uses the latest techniques? The combination of experience, skill and new knowledge outpaces any young surgeon. They are great, however with complication rates now lower than 1/100 they never get any experience dealing with these rare problems. When I was a young surgeon, the most skilled surgeon would see these now “rare” complications on a weekly basis. And the old saying that “good judgment” comes from experiencing the outcomes of previous misadventures, is still with us!

  • I would generally agree – But I had a neurologist at Mt Sinai in NYC who is now approaching 90….yes .
    He no longer practices, but that is more a function of his sight. He still does research and is phenomenal.

    His name is Bernard Cohen – he is a neurologist who specializes in movement disorders and vestibular problems. He may be one of the smartest and kindest docs I have ever seen. I now travel back to NYC to see his protege Catherine Cho, MD who is young…in her 40s.

    But I can tell you wonderful stories about Dr. Cohen – I have seen over 15 docs about this amorphous vestibular disorder. And I was so fortunate to have gotten his name. He did numerous clinical trials for NASA and treated all the astronauts (which helped me also!).
    I owe my current stability to Dr Cohen. He was willing to try any new medications or therapy.

  • Do this article just suggest there was an interesting ” twist” in a “control” in this study.

  • To every reader of this article. Please read and then re-read Dr. Bowman’s brilliant critique of this article. It deserves to be considered in virtually every health related article that finds its way into the main stream press. Particularly those that are chosen by the Stat editors.

  • So the headline should read, “Doctors who see more patients maintain lower average mortality rates.” Dedication matters. Who knew.

  • And by the way, this study did not find gender differences as in the previous study. So we again have contradictions and the researchers examine data in a different way. The fact of the matter is that convenience data has little to do with patients, more to do with billion, and least relationship to health outcomes that are shaped predominantly by non-clinical factors.

  • One should ask whether convenience big data studies prevent real progress. Why do studies not consider time too limited with patients, patient ratios too high for hospitalists or nurses, discharges too fast, community resources, poor communication skills, or poor interactions with nursing teams, family, primary care practices, or community resources?

    Until studies do hypothesize and test for areas of real concern, they will distract from real improvements. Health outcomes are predominantly shaped by patient, local, community, and non-clinical factors that act before, during, and after hospital encounters. To actually demonstrate differences by age or gender or race or ethnicity would require a randomized study of thousands over years.

    Hospitalists are atypical as this is the most recent workforce added explosively to 50,000. The differences between physicians choosing hospitalist careers by different age groups are not studied.

    The care of a hospitalized patient is provided by different teams that are mostly not a single physician as assigned in these studies.

    The readmissions relationship has already been demonstrated to be flawed and should not be used as an outcome. Readmission is most consistently linked to the patient factors.

    BMJ should pay attention to the critique of experts such as Dr. Saurabh Jha regarding lazy generalizations (great for dramatic headlines, rarely relevant), the tyranny of the mean (outcomes only relevant within a narrow range), and study flaws introduced with variations within a comparison group as great or greater than between groups. We supposedly dealt with these 100 years ago in race vs ethnicity differences from intelligence testing, but obviously we still try to compare by race, ethnicity, gender, and age. Speculation about the reasons for differences allow the worse kind of assumptions and media sharing.

Comments are closed.

A roundup of STAT’s top stories of the day in science and medicine

Privacy Policy