Session: Using Big Data to Support Prevention Strategies (Society for Prevention Research 27th Annual Meeting)

Prevention scientists have long employed a plethora of methodologies to collect and take advantage of large datasets, most frequently in the form of randomized clinical trials. However, the types of data encompassed by the notion of “big data” have been less frequently used. The three studies and discussion that comprise this symposium will illustrate how big data and methods can support preventive interventions. The first study by Ramirez and colleagues used longitudinal within-patient healthcare records to test the comparative efficacy of mental health and psychiatric services for Latinxs who are or are not English speakers. They are currently merging the healthcare records with geospatial and census data to further understand the role of social determinants of health in their outcomes and will present these results as part of the symposium. These data will inform the investigators’ strategies for improving prevention programming specifically for the Hispanic community. The second study by Ridenour and colleagues was conducted to evaluate the feasibility and potential value of pediatrician screening, brief intervention and referral treatment (SBIRT) to prevent or curb young adolescents’ risky behaviors. North Carolina healthcare costs during fiscal year 2012 that stemmed from adolescent risky behaviors demonstrated that prevention of adolescent risky behaviors cost third-party payers and healthcare providers hundreds of millions of dollars annually. These results support third-party payer investment in preventive intervention. The third study by Tueller and colleagues presents a methodology and software for conducting N=1 analyses of intensive longitudinal data collected from hundreds of chronic migraineurs using a smartphone app. Results provided each migraine patient and her/his neurologist with the individual’s profile of migraine triggers that contribute to the severity of her/his headaches. This information and analytic service is especially important for treating chronic migraine headaches because of the tremendous heterogeneity among patients that has long interfered with translation of population-based studies into clinical services. Results of analyses in turn are being used to inform patients’ life style changes and treatment strategies to manage the severity of their migraine headaches. Software created to conduct these analyses requires several automated functions and high-throughput processes to meet the requirements of the study’s more than 55,000 analyses. Discussion of these studies by Dr. Morgan-Lopez will consider sources of big data that prevention scientists could take advantage of to support their research and programs, highlighting novel analytic techniques to process big data.