Table 2.
Mean estimated disease prevalence, sensitivity, and specificity (mean asymptotic variance, simulation variance in 10−4 units) from 1000 realizations with different values of α1 and β1 when the sample size is 1000. Models for disease and verification are correctly specified
| (α1, β1) |
||||
| Method | (0.7, 0.7) | (0.5, 0.5) | (1, 0) | (0, 0) |
| Prevalence | ||||
| True value | 0.10 | 0.10 | 0.10 | 0.10 |
| Full† | 0.10 (0.90, 0.90) | 0.10 (0.89, 0.82) | 0.10 (0.90, 0.87) | 0.10 (0.90, 0.90) |
| Naive | 0.21 (4.80, 4.89) | 0.20 (4.61, 4.35) | 0.18 (4.42, 4.59) | 0.12 (3.44, 3.29) |
| BG | 0.10 (1.33, 1.40) | 0.10 (1.51, 1.52) | 0.10 (1.73, 1.84) | 0.10 (2.34, 2.28) |
| MS | 0.10 (1.33, 1.40) | 0.10 (1.52, 1.52) | 0.10 (1.74, 1.85) | 0.10 (2.35, 2.28) |
| IPW | 0.10 (1.44, 1.46) | 0.10 (1.73, 1.63) | 0.10 (2.13, 2.16) | 0.10 (3.21, 3.16) |
| SP | 0.10 (1.39, 1.44) | 0.10 (1.61, 1.60) | 0.10 (1.86, 1.93) | 0.10 (2.56, 2.45) |
| PS | 0.10 (1.47, 1.48) | 0.10 (1.75, 1.67) | 0.10 (2.16, 2.20) | 0.10 (3.21, 3.16) |
| Sensitivity | ||||
| True value | 0.89 | 0.78 | 0.65 | 0.20 |
| Full† | 0.89 (9.8, 9.7) | 0.78 (17.2, 17.4) | 0.65 (22.7, 24.2) | 0.20 (15.9, 16.6) |
| Naive | 0.96 (5.2, 5.5) | 0.91 (11.7, 12.7) | 0.85 (21.5, 22.8) | 0.40 (63.0, 66.7) |
| BG | 0.89 (30.0, 36.0) | 0.79 (46.5, 54.5) | 0.65 (53.8, 63.2) | 0.20 (24.4, 27.3) |
| MS | 0.89 (30.5, 36.1) | 0.79 (46.9, 54.5) | 0.65 (54.2, 63.2) | 0.20 (24.7, 27.3) |
| IPW | 0.90 (34.7, 39.5) | 0.80 (56.6, 65.1) | 0.67 (69.4, 79.4) | 0.21 (31.3, 34.9) |
| SP | 0.89 (33.8, 39.8) | 0.79 (51.4, 60.0) | 0.65 (58.9, 68.3) | 0.20 (25.9, 28.5) |
| PS | 0.89 (34.9, 40.4) | 0.79 (56.0, 65.1) | 0.66 (67.7, 77.7) | 0.20 (29.9, 33.2) |
| Specificity | ||||
| True value | 0.80 | 0.80 | 0.80 | 0.80 |
| Full† | 0.80 (1.77, 1.85) | 0.80 (1.78, 1.68) | 0.80 (1.78, 1.79) | 0.80 (1.78, 1.77) |
| Naive | 0.53 (9.11, 9.66) | 0.53 (9.08, 8.78) | 0.53 (9.06, 8.98) | 0.54 (9.08, 9.64) |
| BG | 0.80 (1.90, 2.01) | 0.80 (1.89, 1.82) | 0.80 (1.89, 1.87) | 0.80 (1.88, 1.85) |
| MS | 0.80 (1.90, 2.01) | 0.80 (1.90, 1.82) | 0.80 (1.90, 1.87) | 0.80 (1.88, 1.85) |
| IPW | 0.80 (1.94, 2.04) | 0.80 (1.93, 1.90) | 0.80 (1.94, 1.93) | 0.80 (1.93, 1.88) |
| SP | 0.80 (1.91, 2.02) | 0.80 (1.91, 1.85) | 0.80 (1.91, 1.88) | 0.80 (1.89, 1.85) |
| PS | 0.80 (1.95, 2.06) | 0.80 (1.95, 1.92) | 0.80 (1.96, 1.95) | 0.80 (1.94, 1.89) |
†
Estimator based on complete data.
BG, Begg and Greenes estimator; MS, Mean Score estimator; IPW, Inverse Probability Weighting estimator; SP, Semi-parametric Efficient estimator; PS, Propensity Score Stratification estimator.