OBJECTIVES: We present the methodology for a near-real-time Pharmacovigilance & Drug Safety System (PDSS) to detect relationships between indications, drugs and outcomes.  Combinations are user-selectable from a library of clinical scenarios including indications, agents and outcomes.  Exemplary results for drug-outcome pairs from the PDSS are illustrated. METHODS: Deidentified claims data for five million individuals covering a 12-month span was analyzed.  A ‘Heat-Map’ was generated with relative risk (RR) for combinations of 73 DDCs and 39 HOIs.  Clinical indications (on- and off-label), presence of drug, and outcomes were identified as, respectively,1+ ICD-9 claim(s), 1+ NDC codes for a drug refill, and outcome code present in months 6-12 but not in month 1-6.  RR and Chi-Squared were calculated for every combination (SAS 9.3). RESULTS: PDSS generated 47,319 indication-drug-outcome combinations. Run time was less than 48 hours.  At a nominal threshold of at least 5 unique patients chi-square (X2) value of over 5, 13,094 (28%) of the combinations were identified for further inspection.  Example results included: GERD-PPI-Bleeding:0.648 (x2=139); GERD-PPI-Fracture:1.42(x2=11.1); Hypertension-ACEI-CHF:0.81(x2=19.5); Hypertension-metoprolol-Renal failure:1.78(x2=86.7); Non-valvular Atrial fibrillation-Dabigatran-Bleeding:1.1(x2=0.2); Valvular Atrial fibrillation-Dabigatran-Bleeding:1.3 (x2=2.6); mechanical heart valve-dabigatran-stroke:8.76(x2=6.5). CONCLUSIONS: The PDSS rapidly quantifies indication-drug-outcome combinations from real-world data. Such a system can provide prospective and continuous drug surveillance and comparative effectiveness & harms analysis. This can permit more efficient targeting of further analyses. Finally, the PDSS is constructed such that it is rapidly able to incorporate new indication, drug, and/or outcome data from claims or clinical (i.e. EMR) data sets.