Using Physical Tracking to Predict Sunburn

Using Retrospective and Real-Time Physical Activity Tracking to Predict Risk of Sunburn in Outdoor Exercisers on Strava

Recreational UV exposure is associated with every form of skin cancer and individuals who engage in more physical activity have a higher prevalence of sunburn, a proximal biomarker of melanoma risk, perhaps explaining why melanoma is the only cancer with which physical activity is positively correlated. Mobile technology for tracking physical activity has become increasingly prevalent and Strava, an activity tracking app and social networking site for athletes, is one of the most popular of these technologies. This research will test the feasibility of delivering location-based, ecologically-valid sun safety advice to Strava users at times when they are predicted to be engaged in outdoor physical activity, by utilizing Strava's public open-source Applications Programming Interface.

Skin cancer is the most common cancer in the United States. Over 5.4 million cases of keratinocyte cancers (basal and squamous cell) are diagnosed annually and incidence of melanoma has increased 3% each year for the past 30 years. In the 2014 Call to Action to Prevent Skin Cancer, the U. S. Surgeon General included strategies for coordinating messages on sun safety and physical activity, recognizing the need for sun safety among populations that engage in physical activity outdoors. Recreational UV exposure is associated with every form of skin cancer; individuals who engage in more physical activity have a higher prevalence of sunburn. Melanoma is the only cancer with which physical activity is positively correlated. Strava, a tracking app and social networking site for athletes, is one of the most popular of mobile technologies for logging physical activity and providing social feedback on activities by other Strava users. The goal of this R21 research is to test the feasibility of interfacing with the Strava website and its mobile app to develop an algorithm that a) predicts when individuals are likely to be engaged in physical activity outdoors in the sun when ultraviolet radiation is high and b) delivers sun safety advice tailored to the time, location, and personal risk (e.g., skin sun sensitivity) of the Strava users. Concept focus groups with Strava users (n=16) and non-users (n=16) will provide input on the feasibility, content, and functions in the Strava Sun (SS) intervention. The algorithm that predicts future outdoor exercise will be developed using machine learning modeling on activity data provided by Strava users (n=1000). The SS intervention will be programmed to deliver ecologically-valid sun safety advice through Strava's open-source Applications Programming Interface (API) via email and comments in the Strava interface. The sun safety advice will be tailored to location, time, season and user's personal risk, employing advice algorithms we developed for a sun safety mobile app, sunZapp, and a social media message library in our Go Sun Smart intervention for outdoor recreation resorts. SS will be tested for usability (n=30 Strava users) prior to conducting a pilot field trial to establish feasibility of SS in practice (n=226 Strava users). The project is significant and innovative. A sun protection interface for the Strava platform will provide individuals who engage in regular physical activity personalized, ecologically-valid advice to help them practice sun safety during outdoor activities. It will contain a novel machine learning algorithm that predicts future high-risk behavior, i.e., outdoor physical activity, to provide precision prevention advice. The intervention will be delivered over an established and popular technology platform with over 42 million registered users.

Participants will be enrolled on a rolling basis and participate in SS for 4 months in the intervention condition. After pretesting, participants will receive sun protection messages from SS through emails/comments based on algorithm results from their profile and activity data. All participants will complete posttest survey 4 months from randomization.

A small control group will be included to see if a no-treatment control condition is acceptable to users and estimate follow-up rates for planning a randomized trial. Participants will be enrolled on a rolling basis and will complete a pretest at randomization. All participants will complete a posttest survey 4 months from randomization.

Test the developed algorithm that delivers location-based, tailored sun safety advice to Strava users through Strava comments for acceptability, ways of interacting with it, and barriers/willingness to allowing access.

What proportion of 8 sun protection practices did participants do on reported days of exercise. Results were calculated by summing sun protection practices done and dividing by the number of days they reported exercise times the total number of potential practices assessed. Range = 0-100%

Participants were asked if their skin was sunburned (described as being red and/or painful from exposure to the sun) in the past three months, and if so, how many times.

Participants were asked if their skin was sunburned (described as being red and/or painful from exposure to the sun) in the past four weeks, and if so, how many times.

Duration in minutes of exercise multiplied by number of times per week, in the past 7 days. Total minutes per week calculated; mean per treatment group reported. Range = 36-2580 minutes.

Duration in minutes of exercise multiplied by number of times per week, in the past 7 days. Total minutes per week calculated; mean per treatment group reported. Range = 0-2220 minutes.

Participants in intervention only were asked about how many messages they read. Scale ranges from 0 - 3. 0 = none of the messages; 1 = some of the messages; 2 = most of the messages; 3 = all of the messages. Higher score indicates higher message recall.

Participants were asked: how confident are you that you can practice sun safety next time that you exercise. 1 = not confident; 2 = a little confident; 3 = somewhat confident; 4 = very confident. Higher scores indicates more self-efficacy for sun safety.

Inclusion Criteria: 18 or older; reside in the United States; read/write in English; upload outdoor activity to Strava ≥3 times per week Exclusion Criteria: Younger than 18; reside outside of the United States; read/write in a language other than English; uploads Strava activity less than 3 times per week OR does not use Strava; having a cognitive or visual impairment that would interfere with participating; having another family member participating in the research; having participated in the focus group and usability testing