Estimated proportions of publications using inappropriate/desirable algorithms in multivariate analysis stratified by the number of events, impact factor and whether medical statistics experts were included as author or not
| Subgroup | Using only significant variables in univariate analysis | Fitting several models for the same outcome and selected factors | |||||
| 95% CI | 95% CI | ||||||
| Proportion (%) | Lower (%) | Upper (%) | Proportion (%) | Lower (%) | Upper (%) | ||
| Medical statistics experts included as first author or coauthor | The number of events* | ||||||
| No | <51 | 20.2 | 12.5 | 31.1 | 2.1 | 0.7 | 5.9 |
| 51–100 | 9.4 | 3.2 | 24.7 | 3.2 | 1.1 | 8.6 | |
| >100 | 8.6 | 5.1 | 14.2 | 10.7 | 6.3 | 17.7 | |
| Yes | <51 | 7.7 | 2.9 | 18.9 | 12.6 | 5.0 | 28.2 |
| 51–100 | 4.0 | 1.2 | 13.0 | 30.1 | 16.5 | 48.6 | |
| >100 | 1.6 | 0.8 | 3.2 | 27.0 | 20.6 | 34.6 | |
| Medical statistics experts included as first author or coauthor | Impact factor | ||||||
| No | Under 2 | 30.6 | 17.1 | 48.4 | 4.0 | 1.1 | 13.7 |
| 2–<4 | 6.5 | 2.4 | 16.3 | 3.4 | 0.8 | 13.1 | |
| 4–<6 | 10.8 | 5.8 | 19.2 | 11.7 | 6.1 | 21.5 | |
| 6 or over | 12.9 | 7.5 | 21.1 | 9.0 | 4.2 | 18.4 | |
| Yes | Under 2 | 6.0 | 1.9 | 17.2 | 16.2 | 5.4 | 39.6 |
| 2–<4 | 3.1 | 1.1 | 8.6 | 22.8 | 10.5 | 42.6 | |
| 4–<6 | 0.2 | 0.0 | 1.1 | 23.7 | 16.1 | 33.5 | |
| 6 or over | 3.5 | 1.7 | 6.9 | 35.5 | 25.9 | 46.4 | |
*The category of ‘<21’ has been integrated with the category ‘21–50’ because of insufficient numbers.