Table 4.
Single-disease and overall composites evaluated at primary care physician (PCP) group level.
| Composite Measure | Total Number of Measure Observations | Measure Adherence | Measures per PCP Group Mean (range) | Casemix- Adjusted PCP Group ICC* (SE) | Panel Needed for 80% Reliability† | PCP Groups with Adequate Panel Size‡ |
|---|---|---|---|---|---|---|
| Diabetes | 85,709 | 71.6% | 5,042 (1,493-10,420) | 0.2 (0.2) | -- | -- |
| CHF | 4,967 | 82.9% | 292 (73-665) | 0 (0) | -- | -- |
| Ischemic Vascular Disease | 15,949 | 76.2% | 938 (185-2,069) | 0.2 (0.2) | -- | -- |
| Prevention | 72,326 | 72.4% | 4,255 (1,389-9,473) | 1.0 (0.5)§ | 400 | 100% |
| Overall | 164,765 | 73.8% | 9,692 (2,877 – 20,802) | 3.2 (1.6)§ | 121 | 100% |
The “Case-mix-adjusted PCP-Group ICC” (primary care physician-group intra-class correlation coefficient) is the amount of total variance in the composite score that is due to differences between PCPs grouped by site after controlling for patient factors, measure factors, and random measurement error.
The panel size needed to obtain 80% reliability in a PCP Group's score for a single-disease or overall composite using the case-mix-adjusted ICC and the Spearman-Brown Prophecy Formula. Only reported for statistically significant case-mix-adjusted ICCs.
Percent of PCP Groups with an adequate panel size to achieve 80% reliability for single-disease or overall composite measures.
Denotes statistically significant ICCs