Table 3.
Estimation method | Death from prostate cancer |
Death from all causes |
||
Adjusted HR (95% CI) | P | Adjusted HR (95% CI) | P | |
Observational (unweighted) | 1.59 (1.27 to 2.00) | <.001 | 1.47 (1.35 to 1.59) | <.001 |
PS reweighted (IPTW) | 1.60 (1.40 to 1.83) | <.001 | 1.54 (1.46 to 1.62) | <.001 |
PS reweighted (SMRW) | 1.39 (1.10 to 1.76) | <.001 | 1.46 (1.33 to 1.59) | <.001 |
Instrumental variable† | 0.73 (0.08 to 6.73) | .78 | 1.09 (0.46 to 2.59) | .84 |
Data are from the randomized controlled trial (29) with a follow-up of 12 years that reported the following relative risks (RR): for death from prostate cancer for men aged 65–74 years at diagnosis, RR = 0.87, 95% CI = 0.51 to 1.49, P = .55; and for death from all causes, RR = 1.04, 95% CI = 0.77 to 1.40, P = .81. CI = confidence interval; IPTW = inverse probability of treatment weights; PS = propensity score; SMRW = standardized mortality ratio weights.
Lagged area treatment was the only instrument.