The story starts 12 years ago when I was a
#whyCMR fellow
@BIDMChealth building a homegrown reporting database in need of robust normative data… 3 years prior I had developed a Bayesian methodology to meta-analyze normative data for the
@ASE360 RV guidelines… Should be easy enough to apply to CMR, right? (spoiler alert - no) Oh wait, there are 1000s of studies to screen, I’ve never been one to shy away from scut work… Uh oh, lots of studies didn’t specify how they traced so need to email the authors, lots of them… I finish my fellowship and move from 🇺🇸 to 🇨🇦
@mcgillu where I set up shop… Original statistician retires, new statistician suggests some tweaks… First look at meta-analysis results in 2015, present proof-of-concept abstract
@scmr 2016… Something is not 100% right, why isn’t the contribution of pap muscles consistent across all parameters?? Must be because of variability in individual studies… Need to smooth this out with informative prior knowledge about the expected difference between tracing techniques… Where can I find patient-level data to derive these priors? Thankfully
@mgwfriedrich is now in Montreal and he has the goods… More time goes by, new studies have been published and need to refresh (actually redo) the lit search… Recruited up-and-coming cardio resident
@YangZhanMD to motor through this (he ended up being absolutely vital to the success of the project)… Combing through all these studies is actually a massive undertaking and another star resident (now star staff)
@ChetritMichael joins the team… Statistician #2 goes MIA and luckily we enlist top-notch
@NandiniDenduku1 who tightens everything up and runs the (near) final analyses… Catch a bug in the code, fix it, open source it… Paper is done, was supposed to be a society guideline but leadership change requires de novo vetting, too late in the game for that… Submit to
@CircImaging and the rest is history…