Background. Knowledge about real-world use of duloxetine and venlafaxine XR to treat depression in the UK is limited. Aims. To identify predictors of duloxetine or venlafaxine XR initiation. Method. Adult depressed patients who initiated duloxetine or venlafaxine XR between January 1, 2006 and September 30, 2007 were identified in the UK's General Practice Research Database. Demographic and clinical predictors of treatment initiation with duloxetine and venlafaxine XR were identified using logistic regression. Results. Patients initiating duloxetine (n = 909) were 4 years older than venlafaxine XR recipients (n = 1286). Older age, preexisting unexplained pain, respiratory disease, and pre-period use of anticonvulsants, opioids, and antihyperlipidemics were associated with increased odds of initiating duloxetine compared to venlafaxine XR. Pre-period anxiety disorder was associated with decreased odds of receiving duloxetine. Conclusion. Initial treatment choice with duloxetine versus venlafaxine XR was primarily driven by patient-specific mental and medical health characteristics. General practitioners in the UK favor duloxetine over venlafaxine XR when pain conditions coexist with depression.