Epidemiology of Lyme disease in Pennsylvania 2006–2014 using electronic health records

Katherine A. Moon, Jonathan Pollak, Annemarie G. Hirsch, John Aucott, Cara Nordberg, Christopher David Heaney, Brian S Schwartz

Research output: Contribution to journalArticle

Abstract

Lyme disease is the most common vector-borne disease in the United States. Electronic health record (EHR)-based research on Lyme disease is limited. We used Geisinger EHR data from 479,344 primary care patients in 38 Pennsylvania counties in 2006–2014 to compare EHR-based Lyme disease incidence rates to surveillance incidence rates, evaluate individual and community risk factors for incident Lyme disease, and to characterize the proportion of cases with diagnoses consistent with post-treatment Lyme disease syndrome in the EHR (PTLDSEHR). We primarily identified Lyme disease cases using diagnosis codes, serologic testing order codes, and medication orders but also completed subgroup analyses among those with positive serology and those with both diagnosis code and antibiotic treatment. We compared annual incidence rates from the EHR to surveillance by age, sex, and county. In case-control analyses, we compared cases to randomly selected controls (5:1) frequency-matched on year, age, and sex. We identified 9657 cases of Lyme disease, including 1791 cases with positive serology and 4992 cases with both diagnosis code and antibiotic treatment. Annual incidence rates in the EHR were 4.25–7.43 times higher than surveillance. In adjusted analyses, white non-Hispanic race/ethnicity (vs. black, Hispanic, or other) was associated with higher odds of Lyme disease (odds ratio [OR]: 2.06, 95% confidence interval [CI]: 1.73–2.44). Medical Assistance insurance use (always vs. never; OR: 0.77, 95% CI: 0.68–0.88), and higher community-level socioeconomic deprivation (quartile 4 vs. 1 OR: 0.50 (95% CI: 0.42–0.59) were associated with lower odds of Lyme disease. Within 4–52 weeks after Lyme disease diagnosis, 20.8% (n = 735) of cases with a diagnosis code and treatment had a diagnosis of malaise or fatigue, pain, or cognitive difficulties not present in the past 26 weeks. These results highlight the utility of EHR data for epidemiologic research on Lyme disease for case-finding, surveillance, risk factor evaluation, and characterization of PTLDS using EHR data.

Original languageEnglish (US)
Pages (from-to)241-250
Number of pages10
JournalTicks and Tick-borne Diseases
Volume10
Issue number2
DOIs
StatePublished - Feb 1 2019

    Fingerprint

Keywords

  • Electronic health record
  • Epidemiology
  • Lyme borreliosis
  • Lyme disease
  • Post-treatment Lyme disease syndrome
  • Surveillance

ASJC Scopus subject areas

  • Parasitology
  • Microbiology
  • Insect Science
  • Infectious Diseases

Cite this