Do overactive bladder symptoms and their treatment-associated changes exhibit a normal distribution? Implications for...

Area of Science:

• Urology
• Biostatistics

Background:

• Overactive bladder syndrome symptom reporting frequently assumes normal distribution.
• The use of means versus medians in statistical analysis for these symptoms requires investigation.

Purpose of the Study:

• To evaluate the prevalent use of means versus medians in reporting overactive bladder syndrome symptoms.
• To test the normality of baseline symptom values and treatment-induced changes in large noninterventional studies.

Main Methods:

• Systematic review of studies on overactive bladder syndrome symptoms in major urology journals (2016-2017).
• Normality testing for urgency, incontinence, frequency, and nocturia in two noninterventional cohorts (n=1335 and n=745).

Main Results:

• 86% of reviewed articles reported means, assuming normal distribution.
• Baseline overactive bladder syndrome symptoms and treatment changes significantly deviated from normal distribution (P<.0001).
• Means overestimated absolute values and changes compared to medians, though percentage change differences were minor.

Conclusions:

• The assumption of normal distribution for overactive bladder syndrome symptoms is not supported by empirical data.
• Recommend reporting basal values and absolute symptom changes using medians and nonparametric analysis.
• Means may be appropriate for reporting percentage changes in overactive bladder syndrome symptoms.