Malpractice In Medical Research

Question 0

Analysis games

Answer 0

Making the analysis up during the study 
Doing lots of stats tests
Too many sub group analyses
Only highlighting outcomes that look good 
Using % relative changes
Test wrong thing, testing changes from baseline
Stopping early 
Weird subgroups
Avoiding intention to treat
Change primary outcome
Ignor baseline difference
Turn a blind eye to rescue therapy

Question 1

Design games

Answer 1

Avoiding active comparators, using obscure comparators, using the wrong dose
Choosing outcomes that inflate effect
Choosing a time assessment that is favourable
Doing lots of placebo studies
Choosing a scale that enhances the effect

Question 2

Write up games

Answer 2

Smoke screen of mechanistic results
Misinterpreting inconclusiveness with equivalence
Forget to mention side effects and patient reported outcomes
Neglect to say how large treatment effect is
Use poor scoring systems
Inflate scales
Positive language
Selective reporting-> outcome bias

Question 3

Publishing games

Answer 3

Don’t or delay publishing negative results
Publish negative studies in obscure journals
Duplicate publish
Salami publish
Citation bias and media attention bias
Data base bias
Grey literature bias