Crowdsourced Photographs as an Effective Method for Large-Scale Passive Tick Surveillance

As tick vector ranges expand and the number of tickborne disease cases rise, physicians, veterinarians, and the public are faced with diagnostic, treatment, and prevention challenges. Traditional methods of active surveillance (e.g., flagging) can be time-consuming, spatially limited, and costly, while passive surveillance can broadly monitor tick distributions and infection rates. However, laboratory testing can require service fees in addition to mailing and processing time, which can put a tick-bite victim outside the window of potential prophylactic options or under unnecessary antibiotic administration. We performed a retrospective analysis of a national photograph-based crowdsourced tick surveillance system to determine the accuracy of identifying ticks by photograph when compared to those same ticks identified by microscopy and molecular methods at a tick testing laboratory. Ticks identified by photograph were correct to species with an overall accuracy of 96.7% (CI: 0.9522, 0.9781; P < 0.001), while identification accuracy for Ixodes scapularis Say (Ixodida: Ixodidae), Amblyomma americanum Linnaeus (Ixodida: Ixodidae), and Dermacentor variabilis Say (Ixodida: Ixodidae), three ticks of medical importance, was 98.2% (Cohen's kappa [κ] = 0.9575; 95% CI: 0.9698, 0.9897), 98.8% (κ = 0.9466, 95% CI: 0.9776, 0.9941), and 98.8% (κ = 0.9515, 95% CI: 0.9776, 0.9941), respectively. Fitted generalized linear models revealed that tick species and stage were the most significant predictive factors that contributed to correct photograph-based tick identifications. Neither engorgement, season, nor location of submission affected identification ability. These results provide strong support for the utility of photograph-based tick surveillance as a tool for risk assessment and monitoring among commonly encountered ticks of medical concern.