Table 4.
Characteristics predicting the relative likelihood (odds ratio) that an expert changed the CT ordinal ratings by ≥1 category and ≥2 categories based on logistic regression models
Probability | Intensity | Frequency | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
% of jobsa | ≥1 category change (n = 293) | ≥2 category change (n = 215) | % of jobs | ≥1 category change (n = 310) | ≥2 category change (n = 68) | % of jobs | ≥1 category change (n = 663) | ≥2 category change (n = 328) | |||||||
OR | CI | OR | CI | OR | CI | OR | CI | OR | CI | OR | CI | ||||
CT agreement rate | |||||||||||||||
>0.9 | 32 | Ref. | Ref. | 18 | Ref. | Ref. | 11 | Ref. | Ref. | ||||||
0.8 to ≤0.9 | 12 | 0.23 | 0.06–0.8 | 0.30 | 0.08–1.2 | 19 | 0.7 | 0.14–3.0 | 0.10 | 0.02–0.46 | 47 | 5.0 | 1.3–19 | 0.41 | 0.07–2.23 |
0.7 to ≤0.8 | 7 | 1.7 | 0.62–4.8 | 1.9 | 0.68–5.6 | 15 | 6.1 | 1.3–28 | 0.38 | 0.05–2.6 | 28 | 5.9 | 1.0–8.0 | 2.0 | 0.52–8.0 |
≤0.7 | 50 | 2.6 | 0.81–8.5 | 3.1 | 0.93–10 | 48 | 7.2 | 1.9–27 | 0.69 | 0.10–4.6 | 14 | 27 | 7.7–97 | 16 | 3.9–69 |
CT ratingb | |||||||||||||||
0 | 16 | Ref. | Ref. | 42 | Ref. | Ref. | 25 | Ref. | Ref. | ||||||
1 | 9 | 38 | 13–114 | 9.1 | 3.1–27 | 23 | 2.8 | 0.98–8.0 | 1.4 | 0.39–5.3 | 58 | 0.23 | 0.07–0.82 | 0.19 | 0.05–0.74 |
2 | 10 | 2.6 | 0.93–7.2 | 0.53 | 0.17–1.6 | 16 | 2.4 | 0.81–7.3 | 33.4 | 10–109 | 15 | 5.8 | 1.8–18 | 2.9 | 0.73–12 |
3 | 64 | 0.15 | 0.06–0.37 | 0.12 | 0.05–0.31 | 18 | 0.51 | 0.09–2.8 | — | 1 | 0.25 | 0.07–0.90 | 0.20 | 0.04–1.0 | |
Diesel module | |||||||||||||||
No | 45 | Ref. | Ref. | 45 | Ref. | Ref. | 45 | Ref. | Ref. | ||||||
Yes | 55 | 2.0 | 0.85–4.7 | 2.3 | 1.0–5.5 | 55 | 1.7 | 0.78–3.7 | 0.90 | 0.22–3.6 | 55 | 2.1 | 1.1–4.2 | 2.1 | 0.91–5.0 |
CI, confidence interval; OR, odds ratio; Ref., reference group.
aJobs in the reliability assessment or in the subsequent expert review (n = 1442).
bCT rating for the metric being assessed.