the state of the medical literature Flashcards
in 2005 how much of the published medical literature were true
In 2005 John Ioannidis
estimated that only
about 15% of
published research
findings are true
why most published findings are wrong three things
1-Smaller studies are
more likely to give
false results
2-Smaller effects are
more likely to be
false
3-The more things
that are tested, the
more likely some
are to be false
positive or negative findings are more likely to be published
Positive findings are
twice as likely to be
published
how much of the researched have the fabrication data
13% of researchers have witnessed data fabrication
how we can increase medical literature reliability
- Increase replicability
- Reduce questionable
research practices, bias,
& conflicts of interest - Educate professionals
4.Encourage the next
generation to be
leaders in the use of
good evidence
do all of the findings follow the design flow
A paper discussing the
finding that half of clinical
trials contain easily
avoided design flaw
what are the pilot studies
Pilot studies are small, underpowered trials used to determine whether a
proposed study is feasible
what is the question that is asked before the pilot study
Will the study design and methods work, or do they need to be improved before a full-scale trial is conducted?
what does the data show in the pilot studies
- The data should be used to draw conclusions about feasibility
should pilot studies use to draw the conclusion for efficiency
they should not be used to draw conclusions about efficacy
what is scientific reward system looking for
The scientific reward system selects for
poor methodology & false positive results