Table 1.
Comparison of sensitivity analyses reported in medical and health economics journals in January 2012
| Variable | Medical journals | Health economics journals |
|---|---|---|
| Number with statistical analysis |
64$ |
71 |
| Number with sensitivity analysis (%) |
13& (20.3) |
22 (30.9) |
| Type of sensitivity analysis |
|
|
| • Methods of analysis |
5 |
12 |
| • Outcome definitions |
4 |
1 |
| • Distributional assumptions |
1 |
0 |
| • Key assumptions* |
2 |
4 |
| • Missing data |
1 |
4 |
| • Baseline imbalances | 0 | 1 |
$Eighteen (18) of these were randomized controlled trials, of which only 3 reported sensitivity analyses.
&Of which 3 were randomized controlled trials and 10 were observational studies.
*Assumptions related to the participants, interventions or outcomes that can affect the results of the trial. For example, considering a second episode of cancer as a relapse instead of a continuation of the first; in a cost-effectiveness analysis, modifying the anticipated frequency of the intervention.