Table 4.
Performance measure score concurrent validity for numerators of RA electronic clinical quality measures.
| Site 1 | Site 2 | Site 3 | Overall | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N K |
Sensitivity | Specificity | N K |
Sensitivity % (95% CI) | Specificity % (95% CI) | N K |
Sensitivity % (95% CI) | Specificity % (95% CI) | N K |
Sensitivity % (95% CI) | Specificity % (95% CI) | |
| Assessment of Disease Activity | 70 1.00 |
100 | 100 | 34 0.17 |
44 (25, 66) | 86 (42, 100) | 117 0.98 |
99 (92, 100) | 100 (92, 100) | 221 .84 |
88 (82, 93) | 99 (94, 100) |
| Functional Status Assessment | 81 1.00 |
100 | 100 | 34 NA |
3 (0.1, 15) | Undef. | 117 0.73 |
96.4 (91, 99) | 100.0 (54, 100) | 232 0.54 |
80.4 (74, 86) | 100 (89, 100) |
| DMARD Therapy | 81 1.00 |
100 | 100 | 34 NA |
94 (80, 99) | Undef. | 58 NA |
98 (91, 100) | Undef | 173 0.85 |
98 (95, 100) | 100 (66, 100) |
| Tuberculosis screening | 62 1.00 |
100 | 100 | 35 0.85 |
93 (76, 99) | 100 (63,100) | 36 0.33 |
69 (50, 84) | 100 (40, 100) | 133 0.73 |
89 (82, 94) | 100 (84, 100) |
K=kappa, a measure of agreement (in these analyses, Kappa=1.0 when the automated EHR data query agrees exactly with data obtained through manual chart extraction, and Kappa=0 when the agreement appears entirely due to chance); RA = rheumatoid arthritis; DMARD = disease modifying anti-rheumatic drug.
Site 1 had perfect agreement on all measures.
2nd site was different for TB screening. 3rd site was different for DMARD therapy and TB screening.
Some Kappas were undefined because there were no true negatives; this is an inherent limitation of the kappa statistic that requires a distribution to produce a result. Similarly, specificity was undefined in instances where the denominator of the specificity calculation (false positive + true negative) was zero.