Table 2. Summary of the 30 individual peer-reviewers of the 319 cases.
Reviewer number* | Median score on six point ordinal scale† (IQR) | Number of cases rated as an adverse event‡N (%) | Model-based reviewer operating characteristic | |
Sensitivity | Specificity | |||
1 | 1 (1,2) | 22 (7) | 0.0562 | 0.92 |
2 | 1 (1,2) | 43 (13) | 0.2703 | 0.9824 |
3 | 1 (1,3) | 51 (16) | 0.25 | 0.9181 |
4 | 1 (1,4) | 90 (28) | 0.5542 | 0.9532 |
5 | 2 (1,3) | 55 (17) | 0.3524 | 0.9832 |
6 | 2 (1,3) | 76 (24) | 0.3784 | 0.883 |
7 | 2 (1,4) | 84 (26) | 0.3992 | 0.8542 |
8 | 2 (1,4) | 86 (27) | 0.3129 | 0.7678 |
9 | 2 (1,4) | 96 (30) | 0.4952 | 0.8671 |
10 | 2 (1,5) | 103 (32) | 0.527 | 0.8537 |
11 | 2 (1,5) | 125 (39) | 0.7717 | 0.9366 |
12 | 2 (1,5) | 134 (42) | 0.8347 | 0.9386 |
13 | 2 (1,5) | 137 (43) | 0.8159 | 0.9048 |
14 | 2 (1,6) | 135 (42) | 0.794 | 0.8975 |
15 | 3 (1,4) | 129 (40) | 0.6847 | 0.838 |
15 | 3 (1,5) | 130 (41) | 0.6964 | 0.8424 |
17 | 3 (1,5) | 156 (49) | 0.7858 | 0.7676 |
18 | 3 (1,6) | 155 (49) | 0.8404 | 0.8207 |
19 | 3 (2,4) | 93 (29) | 0.5815 | 0.9593 |
20 | 3 (2,4) | 115 (36) | 0.4909 | 0.7523 |
21 | 3 (2,5) | 154 (48) | 0.7665 | 0.7627 |
22 | 4 (1,5) | 162 (51) | 0.8889 | 0.8217 |
23 | 4 (1,5) | 167 (52) | 0.875 | 0.7804 |
24 | 4 (1,5) | 171 (54) | 0.8485 | 0.7342 |
25 | 4 (2,5) | 160 (50) | 0.8484 | 0.7984 |
26 | 4 (2,5) | 176 (55) | 0.891 | 0.7417 |
27 | 4 (2,5) | 181 (57) | 0.9025 | 0.7224 |
28 | 4 (3,4) | 190 (60) | 0.706 | 0.4999 |
29 | 4 (3,5) | 204 (64) | 0.9314 | 0.613 |
30 | 5 (2,5) | 218 (68) | 0.9104 | 0.513 |
Overall | 3 (2,4) | 3798 (40) | 0.65 | 0.82 |
Physician reviewers were numbered based on their median 6-point scale score ranking.
1– Definitely due to disease, 6– definitely due to medical care.
Ordinal scale dichotomized: 1–3 ‘Due to disease’; 4–6 ‘Due to medical care’.
Area under the receiver operating characteristic curve.