An example of guidelines for interpreting secondary or subgroup analyses using the mortality analysis from the Multi-Center Study of Hydroxyurea (MSH) trial (adapted from Oxman, et al).5
| 1. Is the magnitude of the difference clinically important? | Yes, 40% reduction in mortality. |
| 2. Was the difference statistically significant? | No, P values were not less than .01. |
| 3. Did the hypothesis precede rather than follow the analysis? | No, mortality was not mentioned in the primary manuscript. |
| 4. Was the subgroup analysis one of a small number of hypotheses tested? | No, several were tested (for example, episodes of acute chest syndrome, time to pain episodes, transfusion requirements). |
| 5. Was the difference suggested by comparisons within rather than between studies? | Yes, comparisons were within the study. |
| 6. Was the difference consistent across studies? | There are no other studies to compare. |
| 7. Is there indirect evidence that supports the hypothesized difference? | Yes, it makes biologic sense that hydroxyurea could improve survival. |
| 1. Is the magnitude of the difference clinically important? | Yes, 40% reduction in mortality. |
| 2. Was the difference statistically significant? | No, P values were not less than .01. |
| 3. Did the hypothesis precede rather than follow the analysis? | No, mortality was not mentioned in the primary manuscript. |
| 4. Was the subgroup analysis one of a small number of hypotheses tested? | No, several were tested (for example, episodes of acute chest syndrome, time to pain episodes, transfusion requirements). |
| 5. Was the difference suggested by comparisons within rather than between studies? | Yes, comparisons were within the study. |
| 6. Was the difference consistent across studies? | There are no other studies to compare. |
| 7. Is there indirect evidence that supports the hypothesized difference? | Yes, it makes biologic sense that hydroxyurea could improve survival. |