Adjusted Differences in Practice System Scores and Diabetes Care Measures
| Variable | Mean, % | Difference Between Groupsa | P Value |
|---|---|---|---|
| Practice systems score | |||
| Overall | 77.7 | +5.3 | .002 |
| Access | 74.8 | +4.5 | .04 |
| Registry | 92.0 | +1.5 | .38 |
| Coordination of care | 72.8 | +7.7 | .001 |
| Care plans | 72.4 | +9.5 | .002 |
| Quality improvement | 92.6 | –1.3 | .50 |
| Diabetes care measure | |||
| Composite optimal | 45.8 | +5.1 | <.001 |
| diabetes care | |||
| Aspirin use | 99.6 | +0.1 | .048 |
| Hypertension control | 84.1 | +1.0 | .11 |
| Statin use | 87.4 | +4.4 | <.001 |
| Hemoglobin A1c control | 70.1 | +1.5 | .006 |
| Nonsmoking status | 84.3 | +1.7 | <.001 |
HCH = health care home.
Note: After regression adjustments. For practice system scores, marginal effects of difference in certification status computed from clinic-level linear regression analyses controlling for practice characteristics (system size, Federally Qualified Health Center status, and rurality). For diabetes care outcomes, marginal effects of difference in HCH certification status computed from both patient-level logit regression analyses controlling for patient characteristics (age, sex, presence of ischemic disease or depression, insurance type, and census-based estimates of wealth, education, and race in patient neighborhood) and practice characteristics (system size, Federally Qualified Health Center status, and location). Diabetes care outcome regression analyses include practice random effects.
↵a Difference in mean score or mean probability of meeting the diabetes care measure between HCH-certified and -uncertified practices.