Abstract
Purpose :
While patient and health system characteristics are known to influence the likelihood of diabetic eye screening, this study aims to identify whether primary care provider characteristics are associated with screening receipt using a state-wide, all-payer claims database covering over 75% of Wisconsinites. We hypothesized that the likelihood of screening receipt would be higher among patients with providers who are female, have greater years in practice, and practice within urban primary care clinics.
Methods :
Using the Wisconsin All-Payer Claims Database, we analyzed adults with type 1 or 2 diabetes (18-75 years old) with claims billed throughout the baseline (1/1/2016- 12/31/2017) and measurement years (1/1/2017- 12/31/2018). Only patients with providers caring for at least 20 diabetes patients were included. We used a mixed-effects, logistic regression to assess provider characteristics associated with diabetic eye screening receipt, adjusting for patient characteristics and using a hierarchical model for the effects of health system and clinic.
Results :
A total of 137,535 adults from 263 Wisconsin health systems cared for by 2,322 primary care providers were included. The overall screening rate was 51.5%. Patients with female providers had a higher likelihood of screening receipt compared to male providers (OR 01.06, 95% CI: 1.03-1.09). Patients with rural providers had a lower likelihood of screening receipt compared to those with providers in urban, socioeconomically-advantaged areas (OR 0.64-0.86, 95% CI: 0.61-0.94). The number of years in practice was not associated with screening receipt. Clinic and health system characteristics had a greater magnitude of effect on the likelihood of screening receipt (variable importance estimation: delta AIC 112.2 and 91.7, respectively) compared to provider rurality, gender, and years (delta AIC 12.7, 8.7, and 1.5, respectively).
Conclusions :
We observed that patients with female primary care providers and those with providers from urban, socioeconomically advantaged areas were more likely to receive diabetic eye screening compared to male and rural providers. Years in practice was not associated with screening receipt. Future studies to investigate how provider gender and rurality affect the likelihood of screening receipt may contribute towards effective interventions to increase diabetic eye screening.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.