Are women better doctors than men? A new study by Yusuke Tsugawa, Anupam B. Jena, and Jose F. Figueroa, et al., explores this question and finds that patient outcomes do differ depending on whether they were treated by a man or a woman. Specifically, they find that among Medicare patients who had been hospitalized, those treated by female internists had lower mortality and readmissions in the following month than those who saw male internists. The study has attracted an enormous amount of attention — and skepticism from some quarters. I spoke to Dr. Tsugawa about why we might be seeing this difference and what doctors should take away from it. Our conversation has been edited for clarity.

HBR: You found that compared to patients treated by a man, patients treated by a woman had a 4% lower relative risk of dying and a 5% lower relative risk of being admitted to the hospital again in the following month. How significant are these differences? 

Tsugawa: These are statistically significant, and we argue that they’re also clinically significant. [It’s] about the same size in terms of magnitude as the 10-year improvement in the mortality rate in the U.S., which is the aggregated impact of better practice, clinical guidelines, new medications, and new interventions. All these things combined have led to a reduction in mortality in the last 10 years, and the difference between male and Female Doctors is about the same size.

Is this a correlational association or causation?  

If there was a confounder, that would make this a correlation. So if we compared men and women doctors, but they’re either treating different types of patients or female doctors are treating healthier patients—then this is not a causation. But we did account for this and other factors.  Some people are saying this difference might be because of different training, but we account for medical school; some say it’s because females might be treating fewer patients, but we did account for that. So it is possible this is correlational, but the problem is nobody has come up with any plausible reason why this is not causal.

But, even though I mentioned this looks causal, the gender of the physicians cannot cause patients to be more or less likely to die. It is not the gender, it is differential practice patterns that are associated with the gender of physicians which is causing patients to be better or worse off. [Women] listen to their patients, they spend more time with them.

We’re not saying we did a perfect job. We think we did a pretty good job, and we want to know if there’s anything else we can do and adjust for. So far, there’s none. [But] we need more analysis to identify the potential mechanisms.

What other reactions have you gotten so far?

At this point people are trying to understand and digest the findings. It’s quite interesting. A lot of men or men doctors are still trying to find some confounders, some explanations for the difference. I think they’re at the state where they’re still not accepting the data, while women are more supportive of the findings. However, even among men there are a lot of people saying we accounted for everything.

Why hasn’t this been studied before?

There was one other study on this, looking at outpatient care and using a much smaller sample size [21,365 vs. 58,344] and they found no difference—probably due to having a smaller sample size or because they looked at healthy outpatients.

One of the key reasons why this is such an understudied area is there wasn’t good data on doctors and doctor characteristics, like age, gender, and medical school, available. We got this data from the Doximity, which is a social networking service for clinicians in the U.S. They actively collect data from multiple sources, including professional societies, institutions, and medical schools. We linked this with Medicare data on patients and were able to find an association.

Could the fact that the patients are 65 and older possibly affect the finding? Or would you expect to see the same thing in younger populations?

I’d say that maybe the pattern would hold for younger populations as well, because the care clinicians provide with the patients are not so different based on their age. We’d provide the same kind of care for heart attack patients ages 70 and 50. I think that we’d find the same patterns. [But] given that older patients are in general sicker and more likely to die, the magnitude of the treatment effect might be smaller for younger populations.

I thought the differences you found among female physicians were interesting: they tend to be younger, more likely to be DO (osteopathic) than MD, in academic teaching hospitals in the Northeast, and they see fewer patients, more female patients, and slightly higher income patients. Did you account for all of these things?

We adjusted for everything. For the hospital differences, we compared male and female doctors within the same hospital, so it doesn’t matter if it’s teaching or non-teaching.

You did measure some differences in practice between male and female physicians. Like length of stay, use of care, and patient volume — did any of these differ significantly between men and women doctors?

We did find some systematic differences, but they were all pretty small. But we didn’t find that any of those narrowed the gap between male and female doctors.

You pointed to past studies that have shown female physicians are more likely to adhere to clinical guidelines, checklists, and evidence-based practice. Why are women more likely to use best practices?

This is beyond the scope of our study, but this has been [examined] in previous studies. We did look at a lot of research on the gender gap or gender differences across diverse fields outside of health care. These found, for example, that if they’re riding a bicycle, women are more likely to wear a helmet. If they’re investing money, men are more likely to invest in higher-risk things. Clearly there’s some systematic pattern that shows that women are more risk-averse than men. It’s not conclusive, but research suggests they are.

Assuming that female doctors are more risk-averse, it’s understandable that they will go back to their guidelines, they will abide by evidence-based medicine, they will consult more, they want to spend more time with their patients—all these things can happen because they want to make sure they’re doing the right thing. Whereas male doctors, if they’re bigger risk takers, maybe they’re overconfident about their practice patterns and maybe that will make them provide care that’s slightly different from clinical guidelines.

There’s a lot of research suggesting that men tend to be more overconfident than women. Has this been studied among doctors?

We speculated from other fields. As far as we know, there are no studies that look at that in medicine.

Is there any reason to think that the male doctors pursue higher-risk strategies with their patients?

It is possible. But it is also possible that some patients benefit from that kind of risk-taking behavior. It’s not always the case that risk-taking behaviors are harmful to the patient. But on average, if you look at the distribution, more likely than not, these more risk-taking behaviors might be worse than just strictly abiding by the guidelines and the evidence-based practice.

But we don’t know that this is what’s necessarily causing this difference in patient outcomes.

We don’t. The other thing is it’s probably not a single factor that is driving the difference between male and female doctors. As we mention in the paper, there are a lot of things they practice differently, and by combining them together, that’s how we find what’s causing the difference between male and female doctors.

Women have been said to have a more patient-centered communication style. What does this mean?

They spend more time with their patients [and] in terms of patient-physician interactions, I think they are better communicators than men. I heard an anecdote on the radio of someone saying a male doctor always looks at the computer, whereas a female doctor is always looking at you. But there isn’t a one-size-fits-all better communication style.

If differences in practice style are what’s leading to better patient outcomes, should men try to adopt women’s approach?

That’s a reasonable conclusion. We don’t want to say either one is better than the other, that men are worse than women doctors. We want to learn what the determinants of high-quality care are, regardless of who treats the patient. The important point is we shouldn’t stop the conversation here and people should look for female doctors; we should do more studies to find out what the causal mechanisms are. We should be asking: How can we learn from each other? It’s only on average that female doctors have better patient outcomes. Men and women may be doing better on different individual aspects.

You estimated that 32,000 fewer patients would die each year if male physicians had the same outcomes as female physicians.

Assuming this is causal, if the quality of care of the male doctors were equivalent to that of the female doctors, we would have 32,000 fewer deaths in the U.S. every year. The effect size would be even larger if the association holds for younger populations as well.

And yet your coauthor, Dr. Jena Anupam, previously found that female doctors in medical schools earn 8% less — $19, 879, — a year than their male colleagues. Are female physicians paid less in general?

There are a couple studies that look at different contexts, and more likely than not, on average, female doctors are paid less and get promoted slower than their male peers.

And this this been chalked up to women taking more time off for maternity leave or working part-time.

You hear these arguments a lot. The problem is that these imply that the quality of care of female doctors is worse. We found that was not the case, and it was actually the opposite.

Female doctors are doing a terrific job, and they may be treated unfairly compared with male doctors. Now we have the finding that they’re better performers, and we can make a stronger argument that it’s unfair to treat female doctors the way we’re doing right now.