In the "Baddest Bias" Session #SER2019 I said: "I'd like to add the 'MD Bias' to this list because while most Epis will be thoughtful (if imperfect) about bias, much of the problematic research is coming non-Epis - so how do we change this?" A (long) Thread. 1/n

#EpiTwitter https://twitter.com/EpiEllie/status/1142891789508128769

Once upon a time I REALLY wanted to be a medical doctor. Tho I (obv) changed fields, I spent a semester shadowing a pediatrician, took physiology, patholophys, med school Repro Med & Immunology courses, have done rounds w mult physicians, and became a licensed EMT. /2

EMT class taught me how to deliver a baby. I attended every PNC visit with a friend, attended multiple births and even cut an umbilical cord. I am quite fluent in basic obstetrical care. So I have a decent level of formal training and informal experience under supervision. /3

AND YET… if I decided to walk into a hospital and start delivering babies I'd be arrested. And rightfully so because I would be putting people's lives in danger! /4

But this level of training is similar to - if not MORE - than many physicians receive in Epi and biostats and yet they are permitted to "practice" health research without the involvement of experts in that research.

This is equally dangerous w *millions* of lives at risk! /5

Poor quality research both yields incorrect (sometimes completely incorrect) information which is then used in clinical practice, AND undermines faith in all medical research when it is later corrected.

Make no mistake, this loss of faith is driving the #MeaslesEpidemic. /6

On the flip side, while epidemiologists understand we are not clinicians ( #NotASkinDr) we often do research without clinical input. I argue that this can also be dangerous as understanding clinical medicine is needed to design good Epi studies & properly interpret our data. /7

Never mind that if we wish to conduct #ConsequentialEpi we need to understand the questions clinicians and patients need to have answered.

Attempting to understand the health of populations is complex and requires multiple domains of expertise. #ResearchIsATeamSport /8

#EpiTwitter: I encourage you to reach out to clinicans working in your field to bring their insight into your work. I have formed valuable partnerships this way which have made my research better.

#MedTwitter: Unless you are MD/PhD in Epi/Biostats, don't do research alone. /9

#MedTwitter Cont: Please understand that the design & conduct of research is as complex as medical care. Even if you have taken courses (maybe have an MPH) you really need good Epi & Stats ppl on your team. Treat them like the experts they are & you all can do great things! /10

#MedEd / #PublichHealthEd: Please be clear with students w clinical backgrounds when the goal of a course is to understand research lit vs contribute to research lit. We sometimes make ppl feel qualified to do research when they are not. Emphasize the need for research TEAMS /11

#MedEd / #PublicHealthEd: Emphasize the complexity of epi. Simplification is helpful for learning but not for understanding limitations. Just bc you can program a regression doesn't mean you are qualified to lead research w/o expert supervision. THIS STUFF IS HARD. Be clear. /12

#MedJournals: Encourage collaboration between research experts & clinicians. Require epi review if an all clinical team & take seriously. @bmj_latest asks if patients were involved in planning. Do something similar & ask for contributions of clinicians, statisticians, & epis. /13

#AnthroTwitter, #SociologyTwitter, patient groups, community groups, and others are other underutilized partners that can helps us ask the right questions and get to the best answers. Engaging with these groups and disciplines can only make us better. /14

Appropriate understandings of our limitations for EVERYONE involved in public health/medical research is necessary. We need to build a culture of collaberation and building off each others strengths.

It is ONLY together that we can make things better. /FIN