Sir,
We read with great interest the paper entitled “Effectiveness of Ultrasound-guided Percutaneous Microwave Ablation for Symptomatic Uterine Fibroids: A Multicenter Study in China” that was published in International Journal of Hyperthermia in November 2016 [Citation1]. The authors aimed to evaluate the clinical efficacy of ultrasound-guided percutaneous microwave ablation therapy (PMWA) for symptomatic uterine fibroids in a multicenter study. We appreciate author’s efforts, but a statistical issue should be considered when interpreting their results.
The authors used independent t tests to determine differences in fibroid volumes, haemoglobin levels, and UFS-QOL score parameters between baseline and each post treatment time point, at 3, 6 and 12 months. Here, they used independent t tests for repeated measures and have done multiple comparisons.
The Type I error rate (alpha) may be inflated (alpha inflation) by analysing the same data more than once. Unfortunately the problem of alpha inflation is a frequently ignored in medical statistics. The common way to deal with this kind of problem is to adjust the Type I error rate below .05, using Bonferroni adjustment. The Bonferroni adjusted level of alpha calculated by: Adjusted alpha = alpha/no. of test. In this study, authors carry out three comparisons for each parameter; therefore an adjusted alpha for this study is 0.05/3 = 0.016. The null hypothesis should be rejected when p < .016, instead of .05. This approach can ensure that the overall chance of making Type I error is less than .05 [Citation2]. To deal with dependency between observations within subjects, a One-way Repeated Measure ANOVA is recommended [Citation3].
Disclosure statement
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.
References
- Liu H, Zhang J, Han ZY, et al. (2016). Effectiveness of ultrasound-guided percutaneous microwave ablation for symptomatic uterine fibroids: a multicentre study in China. Int J Hyperthermia 32:876–80.
- Holm K, Christman NJ. (1985). Post hoc tests following analysis of variance. Res Nurs Health 8:207–10.
- Singh V, Rana RK, Singhal R. (2013). Analysis of repeated measurement data in the clinical trials. J Ayurveda Integr Med 4:77–81.