A prescribed app cut incontinence episodes 61%. The control group changed by 1.7%.

Germany is the only large health system that has been prescribing and reimbursing software as a medical treatment at scale for several years. That makes its trials of digital therapeutics unusually consequential: they are not feasibility studies, they are evidence for products already being paid for.

The Lancet Digital Health published one of the larger ones on 16 December. The DINKS trial randomised 194 adult patients assigned female at birth, with stress, urge or mixed urinary incontinence diagnosed by their treating urologist or gynaecologist and at least one incontinence episode per day, to either an app-based therapy plus usual care or usual care alone [s1].

The result

At week 12, the mean relative reduction in incontinence episode frequency was 60.95% (SD 44.32) in the intervention group and 1.69% (SD 43.75) in the control group — a between-group difference of −59.2% (95% CI −71.7 to −46.7; p<0.0001) [s1].

The trial had set a 50% reduction as the threshold for clinical relevance [s1]. The intervention arm cleared it.

Recruitment ran from 30 April to 2 September 2024 across all regions of Germany [s1]. Ninety-six participants were assigned to the intervention and 98 to control, with eight subsequently withdrawing, four from each arm [s1]. Median age was 50 years (IQR 40–57). Of the participants, 121 (62%) had stress incontinence, 43 (22%) urge and 30 (15%) mixed [s1].

The app's content is not exotic: pelvic floor training, behavioural therapy using methods of urinary urge control, and relaxation techniques [s1]. These are the conservative treatments that guidelines already recommend and that the paper describes as often underused [s1]. No relevant treatment-related adverse events occurred [s1].

What the comparator was, and what that costs the interpretation

The control arm received usual care alone [s1]. It did not receive a sham app, an information leaflet, a diary-only condition, or in-person pelvic floor physiotherapy.

That matters because the control group's incontinence frequency barely moved — a 1.69% relative change over twelve weeks [s1]. In a condition where the recommended conservative treatment is supervised pelvic floor training, a control arm that improved by essentially nothing is a description of how little usual care delivered, not a neutral baseline.

The consequence is that the trial cannot separate the app from the therapy the app contains. A 59-percentage-point gap is consistent with software being an effective delivery mechanism for pelvic floor training. It is equally consistent with pelvic floor training being effective and most patients not currently receiving it. Those have different implications: the first argues for reimbursing the product, the second for fixing referral.

Blinding was single, with the study centre masked to assignment [s1]. Participants necessarily knew which arm they were in, and the primary endpoint — a three-day incontinence diary completed by the participant — is self-reported [s1]. In an unblinded trial of an active, effortful intervention, that combination inflates effects by an unknown amount.

The trial was funded by Kranus Health, which makes the app studied [s1]. This is disclosed in the paper and is normal for regulated digital therapeutics, which have no equivalent of an academic sponsor. It is still a reason to weight the effect size cautiously and to want independent replication.

The comparison that puts the number in perspective

The reference point for how much conservative treatment moves incontinence comes from trials with active comparators. A randomised trial published in Annals of Internal Medicine in 2024 compared a twelve-week pelvic-floor-specific yoga programme with a time-matched general stretching and strengthening programme in 240 women aged 45 to 90 with daily incontinence [s2].

Incontinence frequency fell by 2.3 episodes per day with pelvic yoga and 1.9 with general conditioning, from a mean baseline of 3.4 episodes per day — a between-group difference of −0.3 episodes per day (95% CI −0.7 to 0.0) [s2]. Both arms improved substantially. Neither beat the other.

Read together, the two trials suggest that structured, instructed, regular practice of something is what moves this condition, and that the specific content — yoga, general conditioning, an app — matters less than whether the practice happens at all. DINKS shows an app can make it happen. It does not show the app is doing something a physiotherapist would not.

Why it matters

Germany's reimbursement pathway for digital health applications requires evidence of positive care effects, and trials like this are what that requirement produces. The DINKS result is genuinely large and, as far as it goes, clean.

But the size of the gap is partly a statement about the counterfactual. A country where usual care for incontinence produced a 1.69% improvement over three months has a treatment-access problem that predates any software.

What to watch

Whether the effect holds beyond twelve weeks, and whether any trial compares a prescribed incontinence app against supervised pelvic floor physiotherapy rather than against usual care. That comparison would answer the question reimbursement decisions actually turn on.

This article is informational and is not medical advice.

Sources

• [s1] Haferkamp A, Frey L, Duwe G, et al. "App-based therapy for female patients with urinary incontinence in Germany (DINKS): a single-blind, randomised, controlled trial." The Lancet Digital Health, 7(12), published online 16 December 2025. https://doi.org/10.1016/j.landig.2025.100935
• [s2] Huang AJ, Chesney M, Schembri M, et al. "Efficacy of a Therapeutic Pelvic Yoga Program Versus a Physical Conditioning Program on Urinary Incontinence in Women: A Randomized Trial." Annals of Internal Medicine, 177(10), published online 27 August 2024. https://doi.org/10.7326/m23-3051