OBJECTIVES: Diabetic retinopathy (DR) affects over one-third of diabetes patients. We investigated the possible synergistic effect of the P4P program and the continuity of care index (COCI) on the occurrence of DR among diabetics in Taiwan.

METHODS: From Taiwan National Insurance claims data of 1997-2013, We selected 15,619 DM patients in the P4P program along with age- and gender-matched non-P4P participants at a ratio of 1:4. Time-weighted average (TWA) of the COCI was calculated and included in the time-dependent Cox proportional hazard models to examine the impact of P4P and COCI on the risk of DR, while controlling for demographics and comorbidities.

RESULTS: During the four-year follow-up of 2010-2013, there were 16,116, 4,390, and 683 participants developed any DR, treatment-needed DR, and surgery-needed DR, respectively. In the time-dependent Cox proportional hazard model, we found the adjusted HRs (aHRs) and 95% confidence intervals (CI) of any DR, treatment-needed DR and surgery-needed DR were 1.57 (1.52-1.63), 1.15 (1.06-1.25), 0.75 (0.59-0.96), respectively, for P4P group in comparison to the non-P4P group. With the low COCI group as the reference for surgery-needed DR, the aHR and 95% CI of surgery-needed DR was 0.69 (0.56-0.86) for the middle COCI group, and that of surgery-needed DR for the high COCI group was 0.37 (0.30-0.47). With the interaction of P4P program and COCI, and subgroup of non-P4P and low COCI as the reference for surgery-needed DR, the aHR and 95% CI was 0.71 (0.56-0.89) for non-P4P and middle COCI, 0.38 (0.29-0.48) for non-P4P and high COCI, 0.81 (0.50-1.29) for P4P and low COCI, 0.50 (0.34-0.75) for P4P and middle COCI, and 0.29 (0.19-0.43) for P4P and high COCI, respectively.

CONCLUSIONS: Enrollment into the P4P program and increasing COCI could increase the early diagnosis and medical treatment of DR, but reduce the risk of surgery-needed DR for DM patients independently and synergistically.