Table 2

Estimated proportions of publications using inappropriate/desirable algorithms in multivariate analysis stratified by whether medical statistics experts were included as author or not.

OutcomesProportion95% CI
Lower (%)Upper (%)
1. Using only significant variables in univariate analysis6.44.88.5
Subgroup analysisMedical statistics experts are included as
First authorCoauthor
NoNo12.28.716.8
NoYes3.52.06.1
YesEither1.10.33.5
First author or coauthor2.11.33.6
2. Using too many covariates for few events17.410.228.0
Subgroup analysisMedical statistics experts are included as
First authorCoauthor
NoNo22.113.533.9
NoYes11.53.333.1
YesEither19.03.858.5
First author or coauthor13.65.131.5
3. Fitting several models for the same outcome and selected factors14.411.118.3
Subgroup analysisMedical statistics experts are included as
First authorCoauthor
NoNo7.34.611.4
NoYes19.011.529.7
YesEither30.723.039.7
First author or coauthor26.220.532.9