Diagnostic Testing Characteristics of the Various Manifestations of Lyme Disease in the Emergency Department.

Abstract

Background: Previous studies have shown that clinician's abilities to diagnose Lyme disease, particularly in the emergency department (ED), from symptoms alone are limited. The aim of this study is to determine the diagnostic testing characteristics of demographic and clinical characteristics in ED patients being evaluated for Lyme disease in a hyper-endemic region. Materials and Methods: This is a multicenter retrospective chart review between 2016 and 2017. Eligible cases were identified by searching the electronic health record ED database. Patients were excluded if they were miscoded or had missing clinical information. We calculated the sensitivity, specificity, positive predictive value, negative predictive value, LR+, and LR- with 95% confidence intervals (CIs) for 39 predictor variables using a gold standard of Lyme disease diagnosis, defined as a positive standard two-tier test or clinician-directed ICD-10 code A69.2. All analysis was performed using MedCalc online statistical software. Results: Of the 1527 eligible patients, 577 patients were included in the data analysis. Of these 577, 72 (12.5%) were diagnosed with Lyme disease. Of the predictor variables analyzed, the following were statistically significant: rash (LR+ = 1.73 [95% CI: 1.07-2.78]), joint pain (LR+ = 1.55 [95% CI: 1.17-2.07]), rural residence (LR+ = 1.29 [95% CI: 1.04-1.61]), winter season (LR+ = 0.18 [95% CI: 0.05-0.72]), summer season (LR+ = 1.34 [95% CI: 1.06-1.70]), age less than 16 years old (LR+ = 1.87 [95% CI: 1.21-2.89]), male sex (LR+ = 1.48 [95% CI: 1.24-1.77]), female sex (LR+ = 0.54 [95% CI: 0.36-0.81]), recent tick bite (LR+ = 1.94 [95% CI: 1.02-3.69]), and recent travel (LR+ = 2.24 [1.34-3.74]). Conclusions: No single demographic or clinical characteristic is a strong independent predictor for Lyme disease in ED patients being evaluated for Lyme disease in hyper-endemic regions.