Background. Crowd-sourcing methods have been used in surveillance for influenza-like illness (ILI) to detect epidemic influenza earlier than the Centers for Disease Control and Prevention's (CDC) ILINet. Mining of search engine query data (passive surveillance) and email-based approaches (active surveillance) have been used, but both primarily focus on detecting epidemic flu at the national level.
Methods. At University of Wisconsin-Madison (UW), during the 2013-14 academic year, we began passive surveillance for ILI using a smartphone app called OutSmart Flu (OSF). Users reported their ILI symptoms or feeling well. As feedback, users received near-real time, past 7-day rates of ILI among fellow OSF users and one-week lagged ILI rates estimated by UW University Health Services (UHS), a CDC ILINet sentinel site. We promoted OSF using social marketing, social media, student volunteers, email, and raffles. Nine months after deployment of iOS and Android versions of the app, we examined the composition of OSF users, app engagement, and compared ILI trends among OSF users with trends among UHS attendees.
Results. Overall, 776 users (35% men, 65% women) installed OSF, 583 (75.1%) completed a baseline research survey, 557 (71.8%) completed a profile, 556 (71.2%) submitted at least one symptom report during the academic year. The median (IQR) time elapsed between joining OSF and last use of the app (i.e., length of engagement) was 102 (31, 209) days. Although women were more likely to join OSF, their relative hazard (RH) of using the app was greater than men's use for the first month (RH = 1.41; 95% CI: 1.03, 1.95; p = 0.035), but not thereafter (RH = 0.79; 95% CI: 0.64, 0.99; p = 0.038). Individuals who reported not receiving the previous year's seasonal flu vaccine (2012-13) were significantly more likely to use OSF (RH = 1.64; 95% CI: 1.39, 1.95; p < 0.001). ILI rates mostly paralleled UHS ILI rates, but neither detected the ILI peak identified by national surveillance systems as this fell during UW Winter Recess when both OSF app use and UHS clinic attendance was lower.
Conclusion. College campuses are unique environments for influenza transmission. In its first year, uptake and use of a smartphone-based, surveillance system for ILI among UW students was moderately successful and provide us reason to expand OSF to more users and more campuses during the 2014-15 flu season.
Disclosures. All authors: No reported disclosures.