1 Comment
User's avatar
⭠ Return to thread
Michael Campbell's avatar

Dear John

Shocking example!

If you are collecting articles by authors on what I call 'quasi-experimental studies' , i.e. non-randomised prospective studies , who can't be bothered to do a proper analysis , here is another example:

Pan H, Zhou X, Shen L, Li Y, Dong W, Wang S, et al. Efficacy of apatinib +radiotherapy vs radiotherapy alone in patients with advanced multiline therapy failure for non small cell lung cancer with brain metastasis. Br J Radiol (2023) 10.1259/bjr.20220550

Propensity score matching is not a universal panacea, but can be useful. I wrote a discussion on propensity scores here:

247. Campbell MJ (2017) What is propensity score modelling? Emergency Medical Journal. 10.1136/emermed-2016-206542

Expand full comment