Evaluating an algorithm to identify outpatient Lyme disease diagnoses in Iowa
Description
The number of human Lyme disease (LD) cases continues to grow in the United States with risk expanding outward from historically endemic areas into surrounding regions. The Centers for Disease Control and Prevention (CDC) developed a claims-based algorithm to detect LD diagnoses from national outpatient claims data. Use of the LD algorithm could serve as a useful adjunct to surveillance, enabling researchers to study clinical trends across years and regions and evaluate the effectiveness of interventions; however, limited studies have been conducted to evaluate the positive predictive value (PPV) of this algorithm. Algorithm performance likely differs by state incidence. We sought to evaluate the outpatient CDC algorithm in Iowa, a state with emerging LD incidence.
We identified LD diagnoses from University of Iowa Hospitals and Clinics (UIHC) electronic health data using the CDC LD outpatient algorithm during 2017-2024. From 697 identified LD diagnoses, 100 were randomly selected for double chart abstraction and classified according to the 2017 Council for State and Territorial Epidemiologists LD case definition. We then calculated a broad PPV that included confirmed, probable, and suspect LD cases and a narrow PPV that included confirmed and probable cases of LD.
From the random sample (n=100), the broad PPV was 80% (95% CI: 70.8-87.3%) and the narrow PPV was 41% (95% CI: 31.3-51.3%). The results of this study provide valuable information regarding the application of the LD algorithm in Iowa, facilitating its use to study trends in LD diagnoses within the state and in states with similar incidence patterns.
Citation Information
Draus, William; Kesteloot, Joseph; Schwartz, Amy; Carnahan, Ryan; Petersen, Christine; Brown, Grant; Platt, Jonathan; Askelson, Natoshia M.; and Carvour, Martha, "Evaluating an algorithm to identify outpatient Lyme disease diagnoses in Iowa" (2026). Office of Research DMU Research Symposium. 10.
https://digitalcommons.dmu.edu/researchsymposium/2025rs/2025abstracts/10
Evaluating an algorithm to identify outpatient Lyme disease diagnoses in Iowa
The number of human Lyme disease (LD) cases continues to grow in the United States with risk expanding outward from historically endemic areas into surrounding regions. The Centers for Disease Control and Prevention (CDC) developed a claims-based algorithm to detect LD diagnoses from national outpatient claims data. Use of the LD algorithm could serve as a useful adjunct to surveillance, enabling researchers to study clinical trends across years and regions and evaluate the effectiveness of interventions; however, limited studies have been conducted to evaluate the positive predictive value (PPV) of this algorithm. Algorithm performance likely differs by state incidence. We sought to evaluate the outpatient CDC algorithm in Iowa, a state with emerging LD incidence.
We identified LD diagnoses from University of Iowa Hospitals and Clinics (UIHC) electronic health data using the CDC LD outpatient algorithm during 2017-2024. From 697 identified LD diagnoses, 100 were randomly selected for double chart abstraction and classified according to the 2017 Council for State and Territorial Epidemiologists LD case definition. We then calculated a broad PPV that included confirmed, probable, and suspect LD cases and a narrow PPV that included confirmed and probable cases of LD.
From the random sample (n=100), the broad PPV was 80% (95% CI: 70.8-87.3%) and the narrow PPV was 41% (95% CI: 31.3-51.3%). The results of this study provide valuable information regarding the application of the LD algorithm in Iowa, facilitating its use to study trends in LD diagnoses within the state and in states with similar incidence patterns.
