Prospective Randomized Controlled Trial Comparing the Outcomes and Costs of Two Eyecare Adherence Interventions in Diabetes Patients

Abstract

Objective

Diabetic retinopathy is one of the leading causes of vision impairment among adults in the USA. While it is suggested that diabetics receive annual dilated fundus examinations (DFE), many patients do not adhere to these recommendations. This paper investigates the outcomes and costs of an educational and telephone intervention on DFE follow-up adherence in patients with diabetes.

Methods

In a prospective trial, 356 diabetic patients due for a DFE at an urban eye clinic were randomly assigned to usual care (UC; reference case), mailed intervention (MI), or telephone intervention (TI). UC patients (n = 119) received a standard form letter. MI patients (n = 117) received a personalized letter encouraging scheduling of an eye examination with an educational brochure about diabetic eye disease. TI patients (n = 120) received personal calls (up to three attempts) to schedule a follow-up with standard form letter. The primary outcome was obtaining a DFE within 90 days of suggested return. Costs (US$ 2013) included time costs for staff, phone charges, supplies, and postage. Since TI involved greater cost components compared to MI, univariate sensitivity analysis examined the impact of reducing phone costs.

Results

Patients were mostly female (66 %) and African American (70 %) with a mean age of 61 years. TI patients were more likely to schedule DFE [65 vs. 42 %; relative risk (RR) 1.54; CI 1.20–1.96; P < 0.001] versus UC patients. Obtaining a DFE within 90 days of suggested return was also significantly higher among TI patients compared to UC patients (51 vs. 36 %, RR 1.41; CI 1.05–1.89; P = 0.024). MI patients were slightly less likely to schedule DFE versus UC patients (38 vs. 42 %, RR 0.90; CI 0.66–1.22; P = NSS) and obtain a DFE (32 vs. 36 %; RR 0.90; CI 0.63–1.28; P = NSS). The total cost of TI was US$798.28 or US$6.65/patient and the cost/follow-up DFE was US$26.05. Sensitivity analyses revealed that the cost/follow-up can be greatly reduced but remains greater compared to UC (US$2.76 if US$0.25/call, US$11.13 if $1/call; US$22.29 if US$2/call).

Conclusions

Personal phone assistance in scheduling DFE follow-up is more effective but also more costly. Follow-up research has been initiated to determine whether automated phone reminders can achieve similar effectiveness at a lower cost.

Statistical Appendix: Poisson Regression Models

The results presented in these tables are from Poisson regression models where the dependent variable is keeping timely appointment. Models were fit in SAS PROC GENMOD using generalized estimating equation (GEE) methods with empirical standard errors. Score tests were used to determine the adjusted association of each variable with keeping timely appointment. When a categorical variable had more than two categories, pairwise category comparisons were performed using Wald P values. Table 7 reports results for the main effects model. Parameter estimates are estimates of the log relative risk. Equivalent results on the relative risk scale for Table 7 are reported in the main text Table 3.

Table 7 Poisson regression model for keeping timely appointment: main effects