Table 2.
Characteristics of studies conducting sensitivity analyses, n (%)
| Characteristics | Number of studies (N = 152) |
|---|---|
| Number of sensitivity analyses (median [IQR]) | 3 (2.6) |
| Only one sensitivity analysis | 35 (23.0) |
| 2 to 5 sensitivity analyses | 72 (47.4) |
| 6 to 10 sensitivity analyses | 33 (21.7) |
| More than 10 sensitivity analyses | 12 (7.9) |
| Type of sensitivity analysis | |
| Alternative study definitions | 99 (65.1) |
| Exposure/control | 66 (43.4) |
| Primary outcome | 49 (32.2) |
| Covariates | 28 (18.4) |
| Exposure/control and primary outcome | 1 (0.7) |
| Alternative study design | 74 (48.7) |
| Population under study | 74 (48.7) |
| Alternative modeling | 85 (55.9) |
| Statistical strategy/model | 74 (48.7) |
| Approach to handling of missing data | 12 (7.9) |
| E-value | 7 (4.6) |
| Multiple alternatives | 24 (15.8) |
| Reporting results of sensitivity analyses | |
| Reporting least one result of sensitivity analyses | 131 (86.2) |
| Reporting no results of sensitivity analyses | 21 (13.8) |
IQR Interquartile range