Abstract:
Diabetic retinopathy (DR), the leading cause of visual impairment, is preventable
by early detection and appropriate laser treatment. The purpose of this study is to assess
the cost-effectiveness of various screening intervals using indirect ophthalmoscopy
performed by ophthalmologists for detecting DR among type 2 diabetic patients in
hospital perspective.
The structure of the DR model consisted of six health states: NDR (no DR), BDR
(background DR), PDR (proliferative DR), ME (clinically significant macular edema),
Blindness, and Death. Transition probabilities were derived from published literature
and expert opinions. Cost data were obtained from Ramathibodi hospital.
In base-case analysis, a Markov model was used for simulating a cohort of
10,000 newly diagnosed type 2 diabetic patients, who were followed from 40 years of
age until the age of 75 years or death, whichever occurred first. The incremental costeffectiveness
ratio (ICER) comparing the group being screened every 4 years with the
unscreened group found that it cost about 85,976.89 Baht to prevent blindness per eye.
The ICER of increased screening frequency from every 4 years to every 3 years was
62,806.34 Baht to prevent blindness per eye. The ICER of increased screening
frequency from every 3 years to biannual was 70,553.97 Baht to prevent blindness per
eye. Finally, the ICER of increased screening frequency from biannual to annual was
95,865.04 Baht to prevent blindness per eye.
For sensitivity analysis, if the cost of eye screening, cost of laser treatment,
probability of medical treatment seeking among unscreened, probability of screened
patients being treated with vitrectomy, and annual mortality rate were increased, the
ICER would be increased. If the progression of disease, effectiveness of treatment, the
BDR risk at diagnosis of DM, discount rate, probability of unscreened patients being
treated with vitrectomy, sensitivity of screening, and specificity of screening were
increased, the ICER would be decreased. In addition, if the level of glycemic control
among screened patients was incorporated in the model, the cost-effectiveness of
screening would be increased dramatically. Additional analysis in societal perspective
demonstrated that all screening intervals resulted cost-savings.