Randomised controlled trials: the importance of sample sizeBMJ 2015; 350 doi: https://doi.org/10.1136/bmj.h1586 (Published 23 March 2015) Cite this as: BMJ 2015;350:h1586
- Philip Sedgwick, reader in medical statistics and medical education1
- 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
The effects of pelvic floor muscle training on pelvic floor symptoms in a primary care population were investigated using a randomised controlled trial. The intervention of pelvic floor muscle training was provided individually and combined with home exercises. The control treatment was watchful waiting. Participants were women aged 55 years and over with symptomatic mild pelvic organ prolapse. The primary outcome was change in bladder, bowel, and pelvic floor symptoms, as measured by the Pelvic Floor Distress Inventory-20 (PFDI-20), three months after the start of treatment. The PFDI-20 is patient completed and the scores range from 0 to 300; a higher score indicates a greater symptom burden.1
The sample size calculation was based on detecting a reduction in symptom burden as indicated by a difference of 15 points between treatment groups in the mean change from baseline in PFDI-20 scores. It was assumed that the mean PDFI-20 score at baseline would be 60 points and that the standard deviation of the change in scores at three months from baseline would be 36 points in each group. To detect this treatment effect with a power of 80% and a critical level of significance of 0.05, 92 women were needed in each treatment arm. To allow for an estimated 15% dropout rate it was necessary to recruit 216 women. In total, 287 women were recruited and randomised to pelvic floor muscle training (n=145) or watchful waiting (n=142). Follow-up questionnaires were completed by 250 (87%) women.
The intervention group showed a significant reduction in symptom burden compared with the control group at three months of follow-up. The mean reduction in PFDI-20 scores from baseline was on average 9.1 (95% confidence interval 2.8 to 15.4) points greater in the intervention group than in the watchful waiting group (P=0.005). …