The magnitude and cost of alcohol and other drug (AOD) related problems originating in adolescence make it especially vital to monitor these behaviors, identify high-risk groups and evaluate preventive interventions such as policy measures. Adolescent substance use is associated with the four leading causes of morbidity and mortality in this age group and results in more than $25 billion in direct costs due to medical care and lost time at work annually. Denominator-based health surveillance systems provide population-representative structured data to inform policy and guide clinical and research objectives to address this problem. However, survey timeframes and sampling frames do not support real-time monitoring, robust subgroup investigation, flexible follow up or feedback of health alerts and information to subjects. Harnessing youth engagement with online social media to derive timely information about alcohol use behaviors and problems may accelerate public health awareness and drive targeted responses. The purpose of our project is to develop a new approach to public health surveillance of underage drinking that centers on mining social media communications in Facebook. The approach may: fill gaps in knowledge about risky, stigmatizing or sensitive behaviors; describe subgroups of youth who may not have opportunity/willingness to report about risk behaviors to public health authorities or clinicians; accelerate detection of trends to real time; and, enable evaluation of policy interventions to resolve impacts at greater temporal-social-geographic levels.
Our specific aims are:
Aim 1 - Derive metrics of underage alcohol consumption from user-generated Facebook data, testing correspondence of social media and traditional surveillance reports at the state level. The approach is to collect text reports from the public profiles of US Facebook users ages 13-20 years, creating alcohol use metrics from these data by parsing, filtering and classifying them using machine learning techniques. The main hypotheses are that (1.a) patterns of underage drinking derived from social media will correlate with traditional surveillance reports and (1.b) vary in relation to seasons/events.
Aim 2 - Establish the relationship between validated measures of the alcohol policy environment and measures of youth alcohol consumption obtained from social media. The approach is to ascertain associations among extant, validated alcohol policy scores measured at the state level and social media sourced measures of underage drinking and harms (from Aim 1). The main hypothesis is that social media metrics of underage alcohol use will be associated with alcohol policy at the state level. By utilizing the expertise of an experienced multidisciplinary team and a comprehensive rigorous approach, we anticipate findings that will advance public health methods for monitoring underage drinking and harms and generate a paradigm shift for policy evaluation by enabling real-time investigation of impacts at fine temporal- spatial resolution.

Public Health Relevance

The magnitude of alcohol problems originating in adolescence make it one of the greatest and costliest health problems for this age group. In this work, we will derive and validate metrics of alcohol consumption from information posted on Facebook by underage youth from all 50 US states, and establish the relationship between these social media sourced metrics and surrounding state alcohol policies. Findings will enable a new approach to population surveillance of underage drinking that will complement traditional systems.

National Institute of Health (NIH)
National Institute on Alcohol Abuse and Alcoholism (NIAAA)
Exploratory/Developmental Grants (R21)
Project #
Application #
Study Section
Neuroscience Review Subcommittee (AA)
Program Officer
Freeman, Robert
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Children's Hospital Boston
United States
Zip Code
Liu, Jason; Weitzman, Elissa R; Chunara, Rumi (2017) Assessing Behavioral Stages From Social Media Data. CSCW Conf Comput Support Coop Work 2017:1320-1333
Huang, Tom; Elghafari, Anas; Relia, Kunal et al. (2017) High-resolution Temporal Representations of Alcohol and Tobacco Behaviors from Social Media Data. Proc ACM Hum Comput Interact 1:
Chunara, Rumi; Wisk, Lauren E; Weitzman, Elissa R (2017) Denominator Issues for Personally Generated Data in Population Health Monitoring. Am J Prev Med 52:549-553