19th ANNUAL EARLY HEARING DETECTION & INTERVENTION MEETING
March 8-10, 2020 • Kansas City, MO
3/09/2020 | 2:30 PM - 3:00 PM | Improving Record Matching for the Integration of EHDI Data with Other Data Sources | Chicago B
Improving Record Matching for the Integration of EHDI Data with Other Data Sources
Information systems for Earlier Hearing Detection and Intervention can be and often are integrated with other information systems to improve follow-up, inform health-care providers, and compute population-wide screening statistics. Such data integration requires the matching of person records across various information systems with a high degree of accuracy. However, cross-database person matching for infant records is particularly challenging, due to the lack of common personal identifiers and missing or incorrect values for critical types of information, such as first names, parent information, whether the child is part of a multiple birth, and birth weight. One of the best ways to test and improve a cross-database person-record matcher is to compare its decisions to second matcher’s decisions for the same source records and then examine the differences. This presentation, along with a supporting white paper, describes a study where the author examined the matchings of child records for two different matchers and two data sources. The two matchers come from Utah’s Child Health Advanced Record Management System (CHARM) and the Department of Health Master Person Index (DOHMPI). The two data sources are Utah’s information system for Early Hearing Detection and Intervention, called HiTrack, and its birth registry, called Uintah. This presentation describes the study’s purpose, methodology, results, and conclusions, which include recommendations for the improving both matchers, as well as the quality of the data in HiTrack and Uintah. Although the study was performed on Utah Department of Health databases, its methodology can be adapted to other contexts and systems. Also, kinds of matching and data problems it uncovered may be indicative of problems common to other EHDI-data integration efforts.
- Attendees will become familiar with a reusable study methodology for improving person matchers
- Attendees will become familiar challenges associated with data integration
- Attendees will become familiar ideas for improving person matching and data quality
Presentation:
21060_12549StephenClyde.pdf
Handouts:
Handout is not Available
Transcripts:
CART transcripts are NOT YET available, but will be posted shortly after the conference
Presenters/Authors
Stephen Clyde
(), Utah State University, Stephen.Clyde@usu.edu;
Dr. Stephen Clyde is an associate professor at Utah State University in the Computer Science Department, specializing in software engineering; distributed systems; and integrated health-care information systems. Since joining USU in 1993, Stephen has been the principle investigator on dozens of research and development projects, including the Utah's Child Advanced Record Management (CHARM) project and a system that support automated data exchange between Earlier Intervention Part-C and Part-B. Prior to joining USU, Stephen worked in the software-development industry for 18 years in a variety of positions, including as Chief Technical Officer, Chief Scientist, Project Lead, Senior Systems Analyst, and Software Architect. He has co-founded several successful software-development businesses. Dr. Clyde earned his Ph.D. in Computer Science, with a specialization in Software Engineering, from Brigham Young University in Provo, Utah.
ASHA DISCLOSURE:
Financial -
• Receives Ownership interest for Ownership from Multimedia Data Services Corp..
Nonfinancial -
No relevant nonfinancial relationship exist.