Back to Annual Meeting Posters
Observational Studies Overestimate the Benefit of Breast Reduction Surgery when Compared to Randomized Controlled Trials
Sophocles Voineskos, MD, Christopher J. Coroneos, MD, Forough Farrokhyar, BSc, MPhil, PhD, Mohit Bhandari, MD, MSc, PhD, FRCSC, Achilleas Thoma, MD, MSc, FRCSC.
McMaster University, Hamilton, ON, Canada.
A methodologically sound single large Randomized Controlled Trial (RCT) is the study design that is believed to provide the most accurate and valid estimate of a treatment’s effect. Plastic surgical RCTs, however, are less common and observational studies, which are of a lower level of evidence, are typically favored because of their simplicity. Previous literature from drug trials has shown that well designed prospective cohort studies can derive similar results to randomized trials. The purpose of this study was to determine whether the quality of life effect size (benefit magnitude) provided to a patient by breast reduction surgery in a prospective observational study was similar to that seen in a randomized controlled trial.
Change in Health Utilities Index-3 scores (HUI3, a valid, reliable, and responsive measure of quality-of-life) in patients who underwent an inferior-based pedicle breast reduction mammaplasty were analyzed in two study designs. The change from pre-operative to 12-month post-operative HUI3 scores from a McMaster University observational study with 49 patients were compared to change scores in a McMaster University RCT in which 127 separate patients underwent the same type of procedure. Generalized linear regression was performed to adjust for the baseline characteristics of the patients (conservative α < 0.1 was required for inclusion in the model). An independent samples t-test was performed to compare the mean change of 12-month HUI3 scores between the two study designs (two-tailed test with α = 0.05). Effect Sizes were classified as “Small” (0.20-0.49), “Medium” (0.50-0.79), and “Large” (greater than 0.80).
A total of 176 patients (observational study = 49 patients, RCT = 127 patients) were included from both studies. Patients enrolled in the RCT were older, more educated, and had a lower BMI than those in the observational study. After adjusting for baseline differences, reduction mammoplasty when evaluated in an observational study led to a large treatment effect (effect size=0.89) for mean change in 12-month HUI3 (health utility) scores. However, the same technique, when evaluated in an RCT led to a medium effect (effect size= 0.61). The differences in effect between the RCT and the observational study were not statistically significant (p = 0.15); however, observational study trended to a larger reported effect of reduction mammoplasty on 12-month health utility scores.
In these breast reduction mammaplasty studies, we found that while both types of study design provided patients with improvements in quality of life, the observational study data demonstrated an inflated treatment effect. Therefore, plastic surgeons need to be aware of this possibility when evaluating the benefit of breast reduction mammaplasty techniques, especially when both randomized controlled trials and observational studies are available.
Back to Annual Meeting Posters