Table 3.
Average effects of the interventions on hip fractures per 100.000 person-years for the entire sample and various age and sex-specific groups
| Sample | Effect measure | Controls | |||
|---|---|---|---|---|---|
| Age group | Sex | DD | CE | IRR | n |
| 65+ years | Both | −14.3 (−98.8, 70.3) |
−21.6 (− 149.3, 106.1) |
0.99 (0.92, 1.06) |
4 |
| 80+ years | Both | − 274.7 (− 595.2, 45.8) |
− 135.8 (− 294.2, 22.7) |
0.90 (0.80, 1.02) |
1 |
| 65+ years | Men | − 131.0 (− 314.0, 52.1) |
−90.7 (− 217.4, 36.0) |
0.85 (0.66, 1.09) |
1 |
| 80+ years | Men | − 222.8 (− 705.8, 260.2) |
−39.7 (− 125.6, 46.3) |
0.85 (0.66, 1.09) |
3 |
| 65+ years | Women | −1.3 (− 138.3, 140.8) |
−1.1 (− 118.1, 115.9) |
1.0 (0.91, 1.09) |
5 |
| 80+ years | Women | − 126.1 (− 422.2, 170) |
−39.6 (− 132.5, 53.4) |
0.96 (0.87, 1.06) |
13 |
Notes: *Significant at the 5%-level. DD = difference-in-differences estimate (in rate per 100.000 person-years), CE = cumulative effect (in number of hip fracture patients), IRR = incidence rate ratio effect. The number of controls to include in each analysis is determined via cross-validation (see text for details)