Predicted probability of outcome between automated refills and usual care patients, controlling for baseline differences*
| Usual care Mean (95% CI) | Automated refills Mean (95% CI) | Difference (95% CI) | P value | |
| Intention to treat | ||||
| HgbA1C<8, % | 60.5 (51.5 to 69.5) | 60.8 (52.3 to 69.4) | −0.3 (−12.1 to 12.8) | 0.96 |
| HgbA1C, mean | 7.9 (7.6 to 8.1) | 7.8 (7.6 to 8.1) | −0.04 (−0.42 to 0.34) | 0.84 |
| MPR DM | 57.1 (52.5 to 61.6) | 62.2 (56.2 to 68.2) | 5.1 (−2.9 to 13.1) | 0.21 |
| MPR reference | 54.5 (50.2 to 58.7) | 63.9 (59.3 to 68.6) | 9.5 (3.1 to 15.9) | <0.01 |
| Per protocol | ||||
| HgbA1C<8, % | 60.6 (51.4 to 69.8) | 53.6 (40.4 to 66.8) | −7.0 (−23.6 to 9.5) | 0.39 |
| HgbA1C, mean | 7.9 (7.6 to 8.2) | 8.1 (7.7 to 8.6) | 0.3 (−0.3 to 0.8) | 0.33 |
| MPR DM | 57.3 (53.0 to 61.7) | 76.4 (69.1 to 83.8) | 19.1 (10.1 to 28.1) | <0.001 |
| MPR reference | 54.1 (50.0 to 58.3) | 71.4 (62.7 to 80.0) | 17.2 (7.5 to 26.9) | 0.001 |
*Controlling for patient age, race, preintervention primary care visit number, disability rating and primary care clinic affiliation in quarter prior to project. Glycaemic measures also control for baseline A1C in 12 months prior. Medication measures also control for baseline MPR in 6 months prior.
DM, diabetes mellitus; MPR, medication possession ratio.