Leveraging an Existing Healthcare Quality Improvement Data Infrastructure for Public Health Surveillance

Introduction: The Wisconsin Obesity Surveillance Partnership is a joint effort of healthcare systems, academic partners, and a voluntary statewide collaborative to use electronic health record data to create public health surveillance metrics related to obesity for the state of Wisconsin. As a voluntary collaborative, the Wisconsin Collaborative for Healthcare Quality (WCHQ) works with providers throughout the state to measure quality and affordability of healthcare services in Wisconsin. The University of Wisconsin Health Innovation Program (HIP) and Wisconsin Obesity Prevention Initiative collaborated with WCHQ to leverage their existing data sharing infrastructure to conduct public health surveillance of obesity at the state and ZIP Code levels.

Methods: HIP and WCHQ built on their existing relationships with healthcare systems to support our vision of using WCHQ’s data sharing infrastructure to conduct public health surveillance of obesity in Wisconsin. Body mass index (BMI) was already a publicly reported metric of healthcare quality among WCHQ healthcare system members, ensuring high-quality and consistent data. To support the use of WCHQ BMI data, HIP uses their HIPAA compliant data infrastructure to extract and consolidate BMI data from the WCHQ data warehouse and partnering healthcare systems for patients who were seen for outpatient care at participating systems during the target year. We cleaned and analyzed these data to create statewide and ZIP Code level estimates of obesity prevalence for ages 2 to 100. We suppress estimates for geographies and age groups that do not have sufficient sample size to protect patient privacy. Approved estimates are shared via the Wisconsin Health Atlas, our web-based platform that includes interactive maps, downloadable estimate tables, and a full statewide report.

Results: We created crude state-level estimates and ZIP Code level estimates for 591 of 774 ZIP Codes in Wisconsin using data from 1.8 million patient visits during 2015-2016. State-level estimates are comparable to estimates based on measured heights and weights from the Survey of the Health of Wisconsin and indicate that the prevalence of obesity in Wisconsin is higher than previously estimated by self-report measures. Rates vary markedly across different communities.

Conclusion: Existing “big data” infrastructures used for healthcare quality improvement can be efficiently leveraged to create public health surveillance estimates to support public health practitioners and others working at the local level to prevent obesity. Future work will expand this project to other obesity-related health outcomes available through the WCHQ data, such as diabetes, to track change over time at sub-state geographies, and to create estimates at Census geographies.