Study design, materials and methods
A secondary analysis of data from a prospective cohort study evaluating the effect of a freely available app that supports pelvic floor muscle training [1]. Adult, non-pregnant, non-postpartum women who aimed to treat their urinary incontinence and had completed the Patient Global Impression of Improvement (PGI-I) at 3-month follow-up, were included. We registered their answers to the International Consultation on Incontinence Questionnaire (ICIQ-UI SF) at baseline and follow-up. The change in score was analysed for correlation (Spearman) and using one-way ANOVA to determine a MID. The MID in general and for each sub-group was set to the mean change of the women who reported being a little better.
Results
In total 1,733 women with all types of urinary incontinence were included. Of those, more than half (54.8%) reported having symptoms of stress urinary incontinence only, 31.7% reported symptoms of mixed urinary incontinence and 10.8% of only urgency urinary incontinence. The mean age was 46.5 years (range 18-87). The mean baseline ICIQ-UI SF score was 8.92 (range 3-21). When divided into severity categories, 354 (20.4%) had a slight severity, 1,059 (61.1%) moderate, 300 (17.3%) severe and 20 (1.2%) very severe.
There was a significant correlation between the PGI-I and change in ICIQ-UI SF, Spearman rho -0.323 (p<0.001). The one-way ANOVA revealed significant differences between PGI-I categories (p<0.001) and the MID was set to 1.46 (95% Confidence Interval [CI] 1.26-1.67). Sub-group analysis could not determine a MID for the group with slight severity. For moderate severity the MID was set to 1.33 (95% CI 1.10-1.57) and for severe/very severe to 3.59 (95% CI 3.08-4.09). There was no difference in MIDs between different types of incontinence.
Interpretation of results
The results showed that women with a larger reduction in symptom score experienced a greater impression of improvement. For women who seek to self-manage urinary incontinence via eHealth, a reduction of 1.46 points on the ICIQ-UI SF score would be the minimum important difference. For the group with slight incontinence, a MID could not be determined despite the large sample size. For the group with moderate incontinence, 1.33 points was a clinically relevant reduction and for the group with severe incontinence 3.59 points was clinically relevant. We found no differences in MID between the different types of urinary incontinence, but this finding should be interpreted with caution.