190953 Using Data Mining and Geo-Spatial Analytics to Predict Member Compliance with Diabetic Retinal Eye Exams

Wednesday, October 29, 2008: 10:35 AM

Patty Howard, RN, BSN , Clinical Improvement, BlueCross BlueShield of Tennessee, Chattanooga, TN
Stephen Jones, MS , Medical Informatics - Accreditation Analytics, BlueCross BlueShield of Tennessee, Chattanooga, TN

In the United States, diabetic retinopathy is the leading cause of new blindness among adults 20-74 years old. Since 1999, the diabetic retinal eye (DRE) screening rate has consistently been lower than the national average and has increased only 1.8 percentage points for the Commercial population of BlueCross and BlueShield of Tennessee.


Adult diabetic members who were non-compliant for three consecutive 12-month periods (n=13,897) were observed for the following four consecutive 12-month periods (observation period) for compliance. Twenty months of prior medical claim history were analyzed to assess predictors of compliance. Univariate analyses and decision tree modeling were conducted on 83 explanatory variables.


Approximately 10% of members became compliant during the observation period. Over-sampling was conducted to explain both compliance and non-compliance. Analyses revealed that increased DRE compliance was related to SIC industry code, diabetes type, recent PCP, endocrinologist, and other specialist visits, copay amount, HbAlc testing, higher home value, and vision benefits.


Results support findings that more frequent health care utilization leads to increased compliance of preventive care. Although DRE screenings are covered under medical benefits, member preconceptions of benefits may affect compliance as members tend to comply when they have full vision benefits and smaller copays. Activities to increase visits to PCP and specialists in addition to education on member benefits for diabetics may prove useful in efforts to increase compliance with DRE.

Learning Objectives:
Describe how geo-spatial and data mining techniques can be used to predict compliance to HEDIS measures.

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: Public Health education included program evaluation and statistical methods. I have been in an analyst role for six years.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.