Pay for performance program reduces treatment needed diabetic retinopathy - a nationwide matched cohort study in Taiwan

Table 2 Incidence of Treatment Needed Diabetic Retinopathy (TNDR) and sensitivity analysis, data presented as patient number (incidence rate/1000 person-year)

	P4P group	Non-P4P group	Unadjusted HR (95% CI)	Adjusted HR^* (95% CI)
TNDR	N = 2157 Event = 43 (4.53/1000 person-year)	N = 7154 Event = 119 (3.84/1000 person-year)	1.17 (0.82–1.68)	0.99 (0.93–1.05)^a
Sensitivity analysis
Prior MPR > 80%	N = 1622 Event = 26 (3.67/1000 person-year)	N = 3871 Event = 60 (3.67/1000 person-year)	0.98 (0.58–1.63)	0.95 (0.87–1.05)^b
Propensity score matching	N = 1790 Event = 39 (4.88/1000 person-year)	N = 1790 Event = 67 (8.64/1000 person-year)	–	0.63 (0.39–1.00)
Restrict eye exam within one year in Non-P4P group	N = 2157	N = 734	0.44 (0.28–0.71)	0.78 (0.64–0.94)^c
Restrict eye exam within one year in Non-P4P group	Event = 43 (4.53/1000 person-year)	Event = 31 (10.29/1000 person-year)	0.44 (0.28–0.71)	0.78 (0.64–0.94)^c

*P4P as time-dependent variable
^aAdjusted variables: age, insulin, sulfonylurea, metformin, liver disease, CIC score and DCSI score
^bAdjusted variables: age, corticosteroid, insulin, sulfonylurea, metformin, alpha-blockers, CIC score, DCSI score, history of admission for stroke and heart failure, DM nephropathy
^cAdjusted variables: age, insulin, sulfonylurea, metformin, beta-blockers, CIC score, DCSI score, liver disease, renal disease, admission for heart failure

ISSN: 1472-6963