OBJECTIVES: Multiple methods to identify post-marketing adverse events related to medications exist, yet identifying and reporting adverse drug events (ADEs) remains problematic. The advent of social media platforms provides a robust source to mine pharmacovigilance data. The purpose of this study was to automatically identify associations of ADEs and adalimumab from Twitter accounts using natural language processing techniques and compare ADE tweet rates to known ADE sources. METHODS: Data were collected from Twitter Public API using keywords Humira, adalimumab, and common misspellings. The Twitter API makes available a sample of all posted tweets. Collected tweets were processed by the information extraction system ADRMine, designed to extract potential ADE mentions. Extracted ADRs were mapped to the standard Unified Medical Language System (UMLS) concepts automatically, using a custom-built lexicon. UMLS concept names were categorized by frequency. Disproportionality analyses were conducted to determine the relation of ADR signals to tweets. Rates of UMLS concept names were compared to ADEs reported in the drug compendia Clinical Pharmacology, Micromedex, and Lexicomp. RESULTS: A total of 10,188 tweets mentioned adalimumab; 1382 contained mention of an ADE. Many tweets mentioned medication names but not ADEs. Of ADE tweets, 192 unique UMLS codes were identified. “Pain” (15.5%), “sick” (8.1%), and “tired” (4.6%) were the top three ADE mentions. Pain rates agreed with ADE rates in compendia (6 to 20%) but “sick” and “tired” were not specifically reported as such. Disproportionality analysis resulted in proportional ADR reporting ratio (PRR) of 0.011, 0.013, and 0.011 and lift of 0.012, 0.014, and 0.012 for the top three respectively. CONCLUSIONS: ADRMine identified frequently mentioned ADEs and found reporting rates of some UMLS concepts similar to ADEs catalogued in drug compendia. This study suggests that automatically mining social media and resulting disproportionality metrics can yield promising results for further quantification of ADEs.