Study to Evaluate the Effectiveness of a Program Developed to Improve Eye Care for Veterans With Diabetes

Diabetes is a common cause of blindness and much of this blindness is preventable by early detection and treatment. Although VA and HEDIS quality criteria now allow some individuals with diabetes to have biannual examinations, current diabetes eye care policies continue to emphasize routine, annual eye exams for most diabetes patients. Strong evidence suggests that the current �one-size fits all� method wastes resources while ignoring an opportunity to substantially improve outcomes for high-risk patients.

Recent research shows that patients referred for screening examinations (i.e., those without retinopathy) require different follow-up then those referred for surveillance examinations (i.e., those with retinopathy). Using a more targeted, risk-based criteria for scheduling eye examinations together with system level interventions designed to assure their application may lead to improved healthcare outcomes. Although, the efficacy of such approaches have been demonstrated in other systems and organizational research, it has not yet been demonstrated for diabetes eye care.

Therefore, we propose to conduct and evaluate a prototype translational research project examining the impact of the Proactive Diabetes Eye Care Program, a coordinated and targeted system-level intervention, on: 1) the optimal timing of photocoagulation; 2) the optimal timing of eye care visits; 3) patient and provider satisfaction; 4) health care resource use; and 5) the overall cost-effectiveness of a targeted eye care program.

The primary intervention will involve the use of an innovative �Progressive Reminder and Scheduling System� in which intensity of the reminders is based on the patient�s degree of risk for developing proliferative diabetic retinopathy or macular edema. At the intervention sites, there will be separate clinics for screening (those whose last examination was normal) and surveillance (those with known retinopathy). This two-year prototype translational project will have a quasi-experimental design. Six facilities will be recruited: three will receive the intervention without the system design components. The control and intervention sites will be matched for comparability to baseline screening rates and similar patient populations. The intervention will be evaluated using historical controls (pre-post analyses) and by comparison to control sites.

Data will be collected from three sources. We will use the VISTA database to determine resource use, patients demographics, co-morbidities and medications. Trained medical personnel will conduct chart reviews on a random sample of patients undergoing photocoagulation to determine whether it was sub-optimally timed (i.e., the patient already had a major retinal hemorrhage or advanced macular edema at the time of the procedure). A random sample of patients will be surveyed, at baseline and after 12 months, about non-VA eye care services they received and their attitudes and satisfaction toward eye care. We will also survey health care providers regarding diabetic eye care services.

If successful, this program will serve as a model for disseminating diabetes eye care best practices throughout the VA system and could provide further information about the best approaches to managing other diseases in which patients may benefit from risk stratification rather than being treated according to a single standard.

Inclusion criteria

The population for the project will be all diabetic patients treated at the three VAMCs selected as intervention sites. Therefore, patient recruitment is not required for participation in the initial implementation activities nor in assessing compliance with recommendation for eye examination and treatment. Although the scheduling system will be redesigned and the clinical guidelines will be used to inform eye care clinical decision-making, the final decision about whom should receive diabetes eye care and at what interval will continue to be left to the best clinical judgment of the health care professionals and their patients at each study site. Nothing in this study will preclude a patient and their physician from seeking eye care more frequently or less frequently than recommended by the guidelines. In one regard, this study could be thought of as a study of usual care under two different management and organizational systems without any direct patient-level intervention.

For data analysis purposes, there will be two sets of inclusion criteria ? EPRP visit criteria and diabetes identification criteria. To be included in the primary data analyses a patient must meet the criteria of the Office of Quality & Performances quality monitoring program. This criteria includes at least one outpatient visit during the current fiscal year and at least one outpatient visit in the fiscal year prior to the current year (see http://vaww.oqp.med.va.gov/oqp_services/performance_measurement/tech_man.asp). The previous and current qualifying visits must be in any one of 8 clinics (Primary Care (301), General Medicine (323), Cardiology (303), Endocrinology/Metabolism (305), Diabetes (306), Hypertension (309), Pulmonary/Chest (312), or Women's Health (322)) per the West Virginia's Peer Review Organization's External Peer Review Program (EPRP) selection criteria. Eligible patients would be excluded if they had a life expectancy of 6 months or less (e.g., diagnosis of pancreatic cancer or in hospice). For epidemiological comparability, separate cohorts will be constructed for evaluating eye care at the beginning of the study period (for the pre-intervention analyses) and at end of the study period (for the post-intervention analyses). We will identify diabetics (for analytic purposes) by requiring them to meet one of the following criteria in the previous year: 1) one or more prescriptions filled for hypoglycemic medications or self-monitoring blood glucose supplies (as determined by VISTA records), 2) having a diagnosis of diabetes (ICD-9 250.x) recorded in at least 2 separate outpatient encounters, or 3) having a diagnosis of diabetes recorded for one or more inpatient stays.

Exclusion criteria

None if patient meets inclusion criteria