All of us have a survivorship bias in regard to how we were trained in medicine


All of us have a survivorship bias in regard to how we were trained in medicine

The change precipitated by covid-19 has shown us that just because it’s “always been done this way”, doesn’t mean it should be, or has to be

During the Second World War, planes would return home with numerous bullet holes in their metalwork. Logically, the US military thought that the way to protect these planes during their next flights was to reinforce the areas that had been hit. But this logic was flawed. Abraham Wald pointed out that these planes had survived enemy fire and made it home. Therefore, these areas of direct harm were survivable. The areas on these planes that hadn’t been hit likely represented where other planes had been hit and not made it home. So those were the areas that should be reinforced. Welcome to the concept of survivorship bias. [1]

All of us have a survivorship bias in regard to how we were trained or are being trained in medicine. We made it through our training (at least, thus far), we believe ourselves to be good doctors, thus my training must have been good to produce me. What we don’t take into account with this bias, is the people who leave training posts. There is no standard practice of exit interviews from training programmes, or debriefing for those are struggling and nearly quit. However, when this has been done, it makes for uncomfortable reading, with evidence of bullying and sexual harassment being reasons for leaving surgical training. [2]

It is clear that our training system is not working for everyone. If it were, then there would be more than 12% female consultants in surgery, we would have more than 25% of medical directors being women, and burnout levels in trainees would be less than they are, with PTSD rates in some specialties, worse than returning military personal. [3-5] If we want to build an inclusive world of training, surely we need to make training fit all shapes and sizes, not just those who have survived it in the past?

Survivorship bias can also be dangerous—particularly when it comes to the “in my day” conversations. You know the type, the ones who wax lyrical about how great “training” was when you worked 100 or more hours a week. How they could do more, sooner and without any kind of need for help or codling. How the doctor-patient relationship was much better, how trainees were just “pretty capable at resisting stress.”

And lets make this clear—increased hours of experience does not equal good surgeons. You can practice something badly, without direction for 100 hours. Your 100 hours experience doesn’t count for much if you are still bad at the end of it. But given direction for 10 hours, with deliberate practice and feedback, those 10 hours are worth far more than the initial 100 hours. If that was the case, with the number of hours I sing, I should be Britney spears by now. I am not. Deliberate, careful and coached practice makes you better, which might mean fewer hours doing the thing, but it creates better surgeons.

“In my dayers” are telling current trainees that they will never be good as those who have gone before them. That they are weak because they are burnout. That it is their fault the system isn’t working, rather than the system that needs examining. That they shouldn’t be in training, because the “standards” previously set, they can’t live up to.

But here’s the problem with that; we know that those standards were not only damaging to trainees, (we’ve all read House of God) but more importantly were damaging to patients. It is one of the reasons NCEPOD happened. [7]

Which is why is it so frustrating when the RCS Bulletin publishes another article by a retired surgeon, who writes among other things “In my youth, the consultant was ‘god’ and as such all elements of the institution tended to acquiesce to their request.” [6] There are good reasons why doctors are no longer seen as gods—I refer you to Harold Shipman and Ian Patterson. God cannot be questioned and their behaviours are usually accepted, because God knows best and is unquestionably in charge. God is also, in many cultures, white and male. These attitudes have no place in modern medicine, where we should be striving to compassionate leadership and collaborative team working, to achieve the best for our patients, and weeding out all forms of “god like personality” for the destructive behaviour that it allows. [8,9]

But what this RCS piece really shows us is a classic survivorship bias. Yes, the training worked for those who have come before, but what about the ones who didn’t make it through, who it wasn’t designed for? Who it harmed. The world has changed and we cannot keep allowing those with a survivorship bias, to tell us how to run the modern world of medicine. Is it perfect? No. Could we do better? Of course. Do we have our own biases to address? Definitely. But by continually allowing those who have come before to tell those currently in training they are not good enough is not helping anyone.

During the past year, we have seen the healthcare system change, in some ways overnight. Virtual clinical care, interviews, and mandatory expectations for end of year sign off all adapted. [11] Not all these changes have been welcomed, but they show that the status quo can be changed, and just because it’s “always been done this way”, doesn’t mean it should be.

Out of the darkness of this year, there is an opportunity to rebuild and rethink not only how we deliver healthcare, but how we deliver training. If done well, there can be a phoenix of innovation that rises from the nightmare of 2020 and shines the way to a more inclusive, patient focused future. But please. Leave the in my day conversations where they belong—in the past.

Greta McLachlan, General surgical trainee, Cleveland Clinic London Leadership Fellow, and co-founder of Women Speakers in Healthcare (WISH). Twitter: @geemclachlan.

Simon Fleming is a trainee orthopaedic surgeon and PhD Candidate at Barts and the London School of Medicine and Dentistry. Twitter: @orthopodreg

Competing interests: None declared


  5. Am J Surg 2017 Dec;214(6):1118-1124. doi: 10.1016/j.amjsurg.2017.08.037. Epub 2017 Sep 28