Comparison of Tick-Related Emergency Department Visits and Lyme Disease — Pennsylvania, 2006–2011

BACKGROUND: Lyme disease is endemic in Pennsylvania and prevention and surveillance remains a challenge. To supplement current passive surveillance efforts, an existing statewide hospital emergency department (ED) syndromic surveillance system could be leveraged to monitor for increases in tick-related ED visits (TRVs). This study was conducted to characterize TRVs in Pennsylvania and to determine if they are consistent with the epidemiology of Lyme disease in Pennsylvania.

METHODS: Syndromic surveillance data were queried for hospital ED visits with a chief complaint containing the word “TICK” with visit dates from 2006–2011. Data were also obtained on confirmed cases of Lyme disease cases with erythema migrans (LD_EM) reported to the Pennsylvania Department of Health’s reportable disease surveillance system from 2006-2011. LD_EM and TRV population rates were calculated using 2009 county-level census estimates. Basic epidemiologic data were summarized. Characteristics of TRVs and LD_EMwere compared using Pearson correlation, Wilcoxon rank sum, and binomial proportions tests.

RESULTS: Analysis included 26,246 TRVs and 16,453 LD_EM from 2006-2011. The age distribution was bimodal for LD_EM and TRVs; however, the age distribution of TRVs and LD_EM was statistically different (z=30.7; p<0.001). The proportion of patients 10 years of age or younger was higher among TRVs compared with LD_EM (28.6% vs. 16.3%; p<0.001). Additionally, males accounted for a lower proportion of LD_EM compared with TRVs (55.5% vs. 62.4%; p<0.001). LD_EM and TRV population rates by county were not highly correlated (r=0.51). LD_EMonset dates peaked in early summer, at a median week of 26. In contrast, the distribution of visit week for TRVs was bimodal, with equally large spring (median = week 18) and fall (median = week 44) peaks.  

CONCLUSIONS: In Pennsylvania, the epidemiology of TRVs and LD_EM appear to be different. The difference in age and sex distribution may reflect differences in care-seeking patterns. County-level incidence of LD_EM was not necessarily associated with its TRV rate. This may be due to geographic variations in care-seeking behavior, differences in how hospitals document chief complaints, and other factors. In the absence of a fall peak in reported LD_EM, the observed seasonal pattern of TRVs might reflect adult deer tick activity, rather than nymphal questing activity. Despite these findings, syndromic tick-bite data may be useful in monitoring entomologic activity and tickborne disease risk. Additional description of TRVs is needed to determine if this large burden of ED visits could be prevented through patient education.