You are less likely to die if your doctor is female

Jena, Anupam Bapu Jena at HCP.MED.HARVARD.EDU
Wed Dec 21 18:12:26 UTC 2016

Hi Helene, great question and hard to say. I’d say two buckets at play here: (1) M/F differences in acquiring information, which may be a function of time spent, patient comfort with the physician, attention to detail in acquiring information, and (2) differences in how M/F ‘think’, which is a loose way to ask whether there are cognitive biases that are more common in men than women, on average. These could include anchoring on a prior (ED) diagnosis, premature closure, i.e., failing to keep a broad differential and narrowing in.

I’d emphasize that I think that the difference within male and female physicians is much larger than just comparing male and female docs on average. But these are interesting questions that are raised by our analysis!

Anupam B. Jena, MD, PhD
Ruth L. Newhouse Associate Professor
Harvard Medical School

From: HM Epstein [mailto:hmepstein at]
Sent: Tuesday, December 20, 2016 6:52 PM
To: Society to Improve Diagnosis in Medicine; Jena, Anupam Bapu
Subject: Re: [IMPROVEDX] You are less likely to die if your doctor is female

So, Dr. Jena, where do you think gender plays a role in diagnostic accuracy? Patience in listening to patients?  I don't want to display gender bias but there are many studies that demonstrate that while our brains are very similar, stereotypical female brains process data differently than stereotypical male brains do. [See cartoon below ;-) ]

[Inline image 1]

Or perhaps we can identify it within Buster Benson's Cognitive Bias Cheat Sheet <> under Problem #3: Need to Act Fast such as "Illusory Superiority" or "Egocentric Bias"?

 "In order to act, we need to be confident in our ability to make an impact and to feel like what we do is important. In reality, most of this confidence can be classified as overconfidence, but without it we might not act at all."
See: Overconfidence effect, Egocentric bias, Optimism bias, Social desirability bias, Third-person effect, Forer effect, Barnum effect, Illusion of control, False consensus effect, Dunning-Kruger effect, Hard-easy effect, Illusory superiority, Lake Wobegone effect, Self-serving bias, Actor-observer bias, Fundamental attribution error, Defensive attribution hypothesis, Trait ascription bias, Effort justification, Risk compensation, Peltzman effect

Diagnostic Errors on Facebook<>

On Tue, Dec 20, 2016 at 9:39 AM, Jena, Anupam Bapu <Jena at<mailto:Jena at>> wrote:
Hi Mark, wonderful to hear from you! I’ll share my thoughts and Ashish should chime in.

First off, terrific question, but hard to say.  The first issue we have/had to address is whether our findings are driven by something unmeasured about the patients male and female docs see. We don’t think that’s likely b/c we focus on hospitalist physicians who within the same hospital are plausibly ‘quasi-randomized’ to patients, i.e., patients don’t choose M vs F docs in that setting. We also observe that the characteristics of patients are similar, which is consistent with the randomization hope.

That begs the question, why? It’s not about spending intensity, LOS, or where patients are discharged. We look at those. So, we are left w/ two things, are there slightly more misdiagnosis, incorrect treatments, or a hybrid problem of prognosis (i.e., may not matching the best treatment to a given patient).  If I had to guess, misdiagnosis would have to be an element b/c we are looking at mortality here. I think it’s more likely that a misdiagnosis may be made than the correct diagnosis made but the incorrect treatment offered. I say that b/c I think (w/o data) that doctors would be better at identifying the correct trx if they have the right diagnosis, but the right diagnosis may be elusive.

There are reasons to think that misdiagnosis may play a role. There may be differences in the ability to ‘extract’ important information, due to time spent, patient comfort w/ the doctor, etc. All this is plausible but hard to know if it could actually improve mortality. I would lay my money on differences in clinical decisionmaking, e.g., anchoring, premature closure of diagnosis; these are features of decisionmaking that are different on average between M and F in other settings (e.g., finance is where this has been shown I believe).

Those are my 2 cents!

Hope all is well, Bapu

Anupam B. Jena, MD, PhD
Ruth L. Newhouse Associate Professor
Harvard Medical School

From: Mark Graber [mailto:graber.mark at<mailto:graber.mark at>]
Sent: Monday, December 19, 2016 12:51 PM
To: Listserv ImproveDx
Cc: Jha, Ashish; Jena, Anupam Bapu
Subject: You are less likely to die if your doctor is female

Patients of female hospitalists had lower 30 day mortality rates and readmissions - see article attached, or here:<>

Two of the authors were on the IOM panel that drafted “Improving Diagnosis in Health Care”, copied on this message.  Congratulations guys, and maybe you can help us:  Does this difference relate to doing a better job with diagnosis, treatment, or both?


Mark L Graber MD FACP
Senior Fellow, RTI International
Professor Emeritus, SUNY Stony Brook
President Society to Improve Diagnosis in Medicine (SIDM)

[cid:image003.png at 01D25B8B.DEA65470]




Visit the searchable archives or adjust your subscription at:<>
Moderator:David Meyers, Board Member, Society for Improving Diagnosis in Medicine

To learn more about SIDM visit:<>

To unsubscribe from the IMPROVEDX:
or click the following link: IMPROVEDX at LIST.IMPROVEDIAGNOSIS.ORG


For additional information and subscription commands, visit:

http://LIST.IMPROVEDIAGNOSIS.ORG/ (with your password)

Visit the searchable archives or adjust your subscription at:

Moderator: David Meyers, Board Member, Society to Improve Diagnosis in Medicine

To unsubscribe from the IMPROVEDX list, click the following link:<br>
<a href="" target="_blank"></a>

HTML Version:
URL: <../attachments/20161221/e7b29e44/attachment.html> ATTACHMENT:
Name: image002.png Type: image/png Size: 212873 bytes Desc: image002.png URL: <../attachments/20161221/e7b29e44/attachment.png> ATTACHMENT:
Name: image003.png Type: image/png Size: 61111 bytes Desc: image003.png URL: <../attachments/20161221/e7b29e44/attachment-0001.png>

More information about the Test mailing list