Identifying the geographic leading edge of Lyme disease in the United States with internet searches: A spatiotemporal analysis of Google Health Trends data

Abstract

Background: The geographic footprint of Lyme disease is expanding in the United States, which calls for novel methods to identify emerging endemic areas. The ubiquity of internet use coupled with the dominance of Google's search engine makes Google user search data a compelling data source for epidemiological research.

Objective: We evaluated the potential of Google Health Trends to track spatiotemporal patterns in Lyme disease and identify the leading edge of disease risk in the United States.

Materials and methods: We analyzed internet search rates for Lyme disease-related queries at the designated market area (DMA) level (n = 206) for the 2011-2019 and 2020-2021 (COVID-19 pandemic) periods. We used maps and other exploratory methods to characterize changes in search behavior. To assess statistical correlation between searches and Lyme disease cases reported to Centers for Disease Control and Prevention (CDC) between 2011 and 2019, we performed a longitudinal ecological analysis with modified Poisson generalized estimating equation regression models.

Results: Mapping DMA-level changes in "Lyme disease" search rates revealed an expanding area of higher rates occurring along the edges of the northeastern focus of Lyme disease. Bivariate maps comparing search rates and CDC-reported incidence rates also showed a stronger than expected signal from Google Health Trends in some high-risk adjacent states such as Michigan, North Carolina, and Ohio, which may be further indication of a geographic leading edge of Lyme disease that is not fully apparent from routine surveillance. Searches for "Lyme disease" were a significant predictor of CDC-reported disease incidence. Each 100-unit increase in the search rate was significantly associated with a 10% increase in incidence rates (RR = 1.10, 95% CI: 1.07, 1.12) after adjusting for environmental covariates of Lyme disease identified in the literature.

Conclusion: Google Health Trends data may help track the expansion of Lyme disease and inform the public and health care providers about emerging risks in their areas.

Fig 1. Designated market area-level percent change in Google Health Trends “Lyme disease” query fractions between 2011 and 2019.

State boundaries from the Census Bureau (public domain) and high-incidence state boundaries (black) are shown for reference. DMA boundaries are the intellectual property of Esri and are used herein with permission. Copyright © 2024 Esri and its licensors. All rights reserved.

Fig 2. Designated market area-level spatial covariation in reported Lyme disease incidence rates and Google Health Trends “Lyme disease” query fractions using 2011–2013, 2014–2016, and 2017–2019 averages.

Breaks for low, moderate, and high incidence are 0–10, 10–50, and 50+; those for query fractions are 0–500, 500–1,000, and 1,000+. State boundaries from the Census Bureau (public domain) are shown for reference. DMA boundaries are the intellectual property of Esri and are used herein with permission. Copyright © 2024 Esri and its licensors. All rights reserved.