How well do randomized controlled trial data generalize to 'real-world' clinical practice settings? A comparison of two generalized anxiety disorder studies

The aim of this post-hoc comparison is to compare efficacy and tolerability results from two generalized anxiety disorder (GAD) studies: a placebo-controlled, randomized controlled trial (RCT) and a study conducted in the clinical practice setting, and to evaluate the extent to which results from RCTs in GAD patients can be generalized to clinical practice. In the clinical practice study, GAD outpatients (n=578) were treated with 4 weeks of pregabalin 150-600 mg/day. In the double-blind placebo-controlled RCT, GAD outpatients (n=249) were randomized to 8 weeks of pregabalin (300-600 mg/day), or placebo (only the first 4 weeks are included in the current analysis). Efficacy measures included the Hospital Anxiety and Depression Scale - Anxiety and Depression subscales (HADS-A; HADS-D), a visual analogue anxiety scale (VAS-Anxiety), and the Medical Outcomes Study Sleep Problems Index (MOS-SPI). Baseline HADS-A and HADS-D scores were both higher in the clinical practice study vs. the RCT. In the RCT, treatment with pregabalin resulted in significantly greater Week 4 change vs. placebo in the HADS-A (−5.3 vs. −3.9; P<0.005), VAS-Anxiety (−24.0 vs. −13.3; P<0.02), MOS-SPI (−19.1 vs. −9.5; P<0.01), and HADS-D (−2.7 vs. −1.4; P<0.05). The magnitude of Week 4 improvement on pregabalin in the clinical practice study was numerically larger on the HADS-A (−5.9), VAS-Anxiety (−36.0), MOS-SPI (−22.7), and HADS-D (−5.1), despite use of lower doses. These results suggest that clinical practice patients with GAD may achieve comparable efficacy on lower doses of pregabalin than tested in RCTs, despite having comparable levels of anxiety symptom severity at baseline. The current exploratory comparison also suggests that results from RCTs in patients with GAD may not be directly generalizable to clinical practice.