OBJECTIVES: Many types of pain medications are commonly used in treating osteoarthritis (OA). The purpose of this study was to identify the demographic and clinical factors associated with the choice of pain medications between duloxetine and other pharmacological agents among OA patients. METHODS: A retrospective cohort study was conducted using a claims database. OA patients who initiated duloxetine, anticonvulsants, antidepressants, muscle relaxants, non-steroidal anti-inflammatory drugs (NSAIDs), opioids, or steroids between November 1, 2010 to March 31, 2011 were selected. The dispense date of the first initiated medication preceded by at least a 90-day gap in medication supply was defined as the index date. Comorbidities and prior medication use was assessed during 6 months prior to the index date. Multiple logistic regression models were performed to identify predictors of initiating duloxetine versus other pain medication. Models were estimated with different sets of predictors: 1) Charlson comorbidity index (CCI) score and use of selected medications, and 2) individual comorbidities and number of unique pain medications in the prior six months, respectively. RESULTS: We identified 96,666 OA patients (mean age 65.2 years; 64.5% female) who initiated opioids (41.6%), NSAIDs (21.8%), steroids (15.4%), muscle relaxants (7.7%), antidepressants (7.5%), anticonvulsants (3.9%), or duloxetine (2.1%). Patients who initiated anticonvulsants and duloxetine had the highest CCI scores, were on more medications and had more inpatient stays prior to initiation than patients initiating other medications. Males were in general less likely to initiate duloxetine than the other medications. Higher CCI (except for comparing with anticonvulsants) score and prior use of pregabalin, gabapentin, antidepressants, muscle relaxants, and steroids were associated with a higher likelihood of initiating duloxetine. Patients who had more than 10 unique medications, depression, or anxiety were more likely to initiate duloxetine. CONCLUSIONS: Presence of selected demographics, comorbidities and prior use of medications were associated with duloxetine initiation among OA patients.