App-based treatment for urinary incontinence in women: A pragmatic, randomized, controlled, non-inferiority trial in primary care setting

HYPOTHESIS/AIMS OF STUDY: Over 100 apps for urinary incontinence are available in the app store and google play, but evidence on their effective-ness is scarce. Available studies focused on the treatment of stress urinary incontinence only (1). Mobile applications may support women with urinary continence, but it is unclear if the effect of such treatment equals care as usual. In many countries, general practitioners (GPs) provide frst line treatment of urinary incontinence. We studied if the effectiveness of an app-based treatment for stress-urgency-and mixed urinary incontinence is non-inferior to care as usual provided by GPs after 4 months. To assess the beneft of an app in real clinical practice, we chose a non-inferiority comparison to care as usual. STUDY DESIGN, MATERIALS AND METHODS: This trial was a pragmatic, block-randomized, two arm, parallel, non-inferiority design with participants receiving app-based treatment or care as usual for stress-, urgency-or mixed urinary incontinence (UI). We recruited participants through primary care practices, lay press and social media. Adult woman, with ≥ 2 urinary incontinence episodes per week, with access to a smartphone or tablet and willing to receive treatment were eligible. We excluded woman with conditions possibly related to stress-, urgency-or mixed incontinence, those who had previous surgery for incontinence or any other treatment for incontinence in the previous year, or those that were unable to complete a questionnaire in Dutch or had a terminal or mental illness. Participants signed an informed consent form before entering the study. The app contains a step-by-step program for the self-management of urinary incontinence. The app ofered information on incontinence and lifestyle advices, exercises focusing on awareness of the pelvic foor muscle and exercises to train the pelvic foor muscle and/or bladder. The type of incontinence directed treatment advices. Care as usual consisted of any or more of the following: instructions on pelvic foor muscle and/or bladder training; prescribing a pessary, drugs, or absorbent products; or referral to a continence nurse, a pelvic physiotherapist, or to secondary care. After inclusion and baseline assessment, the randomization was carried out using the validated web-based computer program ALEA to ensure full allocation concealment. The primary outcome was diference in change in urinary incontinence severity (assessed with the ICIQ-UI-short form) from baseline to 4 months. We used linear regression to compare the change of ICIQ-UI-SF symptom score between groups, in a per-protocol analysis, which is the most conservative for the non-inferiority design. For this, we included women from the app-group that logged in in the app at least once and women from the care as usual group that visited their GP for incontinence at least once. The non-inferiority margin is set at 1.5 points diference of change in UI severity between groups. This is based on the 1.58 points between-treatment minimum important difer-ence in change (MID) for the ICIQ-UI-SF, previously identifed among participants with stress UI (2). Assuming a power of 0.80 with a one-sided type I error of 0.025, we needed a sample of 100 evaluable participants per group. RESULTS: Participants were recruited through 88 GPs (n=201) and through (social) media (n=149). Of 350 participants screened, 262 were eligible and randomly allocated to either App-based treatment (n=131) or care as usual (n=131). The mean age of included women was 53 years (range 20-86), median duration of incontinence was 7 years (Interquartile range 4-14). Fifty percent of participants reported a mixed type of incontinence, 42% stress incontinence and 8% urgency incontinence.The incontinence severity was slight (10%) moderate (64%) or severe (26%). The per-protocol analysis included 96 (app-group) and 75 participants (care as usual). Both groups showed improvement after treatment with a mean change in incontinence severity of-2.15 (SD 2.56) points in the app group and-2.75 (3.62) points in the care as usual group. The mean diference of change of incontinence severity between both groups was 0.071 points (95% CI:-0.837 to 0.979), which did not reach the pre-specifed boundary of 1.5 points but did cross zero (Figure1). This means that the app-based treatment effect was non-inferior to the care as usual treatment effect, but not superior. INTERPRETATION OF RESULTS: This RCT shows that an app-based treatment for woman with stress-, urgency and mixed urinary incontinence is at least as effective as the best current treatment provided by GPs at 4 months. This is the frst study with a pragmatic comparison of an app-based treatment for urinary incontinence to care as usual and we believe our results provide a realistic reffec-tion of the treatment effect in a real life setting. Symptom improvement shown was clinically relevant in both groups. The three main types of incontinence can be treated by the same app, therefore increasing the applicability and relevance of app treatment. CONCLUDING MESSAGE: This is the frst study on an app-based treatment for women with stress-, urgency-, and mixed UI, showing that it is non-inferior compared to care as usual in primary care after 4 months. Further research is needed to study long-term effectiveness, and the cost-effectiveness of app-based treatment, and to understand the process underlying treatment success.