Table 4.
Results from the random effects meta-analysis in which studies were excluded stepwise with specific exposure assessment descriptors.
| Incremental exclusion | No. of studies included | I2 (%) | Meta-α (95% CI) | Meta-KL × 100 (95% CI) | AIC | Studies includeda | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| None (all 19 studies) | 19 | 64.1 | 1.47 (1.14, 1.81) | 0.13 (0.04, 0.22) | 28.2 | 1–19 | ||||||
| Studies with insufficient documentation | 11 | 67.4 | 1.29 (0.87, 1.71) | 0.18 (0.04, 0.33) | 30.6 | 1, 4, 5, 8, 9, 12, 13, 15, 17–19 | ||||||
| Studies with external conversion factors | 9 | 68.5 | 1.38 (0.89, 1.86) | 0.19 (0.03, 0.35) | 30.6 | 1, 4, 5, 8, 9, 12, 13, 18, 19 | ||||||
| Studies with insufficient job histories | 5 | 73.7 | 1.44 (0.78, 2.10) | 0.36 (0.10, 0.61) | 26.4 | 4, 8, 9, 18, 19 | ||||||
| Studies with CE ratio ≤ 50 | 3 | 84.4 | 1.32 (0.50, 2.14) | 0.56 (0.12, 1.00) | 25.0 | 4, 9, 19 | ||||||
| Studies with coverage ≤ 30% | 2 | 88.4 | 1.42 (0.40, 2.44) | 0.55 (0.11, 0.99) | 25.3 | 4, 9 | ||||||
| aStudy numbers are given in Table 1. | ||||||||||||