BACKGROUND
Minimize errors
Epidemiologists from NYC’s Bureau of Epidemiology Services study the patterns, causes, and effects of health and disease conditions in New York City neighborhoods to help keep residents healthy. Difficulty in accurately matching patient identity across multiple data sets from different sources and programs limits the effectiveness of this analysis.
NYC decided to create a master patient index (MPI) that would contain unique identifiers for individuals from multiple, disparate data sources. The MPI was also envisioned to contain a consolidated demographic record assembled from the contributing data sources. The project was expected to improve the accuracy of patient matches, reduce the errors in consolidated data, and improve the quality of the resulting analyses. Furthermore, it is designed to serve as an ongoing, reusable matching solution that should eliminate the need to repeatedly conduct one-off matching efforts between these data sources every time that a new study is performed.
APPROACH
Maximize leverage
HLN created an MPI by integrating an open source probabilistic matching engine (Choicemaker) with a set of RESTful MPI services and a new MPI database. Three years of data consisting of more than 21 million patient records were then loaded, matched, and merged from a variety of different sources (births, deaths, Medicaid claims, and hospital discharge records for emergency department visits, inpatient, ambulatory surgery and outpatient visits) into an MPI database that now contains 12 million distinct patient records.
The MPI services were originally developed by HLN for a different part of the agency to support the integration of the Citywide Immunization Registry (CIR) with LeadQuest, the blood lead level testing tracking system of the Healthy Homes program. By reusing key components and modifying others to meet the specific requirements of this project, NYC was able to leverage and capitalize upon prior investments and experience.
RESOURCES
- Journal Article: Is There a National Strategy Emerging for Patient Matching in the US?
- Foundational Work: Connections Community of Practice, Unique Records Portfolio
- Choicemaker Open Source Matching Engine
Want to see if
HLN can help
your agency?