June 13, 2020

Randomized Controlled Trials

Source: https://outofpocket.substack.com/p/random-uncontrolled-trialstweets

  • Real-world data / evidence had larger more diverse datasets and can be useful when a Randomized-Controlled-Trials isn’t feasible
  • Single arm trial: When the disease is so rare you only recruit people to get the novel treatment, and compare how they do with others getting their normal treatment (or with past patients)
  • Pragmatic trial: focus is less on showing the intervention works and more on trying to tweak and improve outcomes

Reminds me of how Machine Learning research-is-done-more like an Engineering discipline? https://www.bmj.com/content/bmj/363/bmj.k5094.full.pdf

Updated Sep, 13 2020

#Mental-models #Metascience #Epistemics