Hypothesis / aims of study
Urinary incontinence (UI) is one of the most prevalent health concerns confronting women age 60 and over. Pelvic floor muscle (PFM) training is effective and recommended by international clinical practice guidelines as first-line treatment for stress or mixed UI in women (Evidence Level 1; Recommendation Grade A). However, inadequate human and financial resources limit its delivery. This often results in surgery being used as first-line therapy, even if it is associated with a higher risk of complications, putting pressure on an already stretched health care system. Group-based PFM training could alleviate this problem. To date, however, individual and group-based PFM training for UI in women have not been compared in an adequately powered trial. The aim of this study was to determine if group-based PFM training is not meaningfully less effective than individualized PFM training, for stress or mixed UI in women aged 60 and over, one year after randomization.
Study design, materials and methods
We conducted a multicenter non-inferiority randomized trial. Participants were recruited through advertisements in community centers, newspapers, web and social media initiatives, public conferences in addition to medical clinics. Eligible participants were incontinent community-dwelling women, aged 60 and over with symptoms of stress or mixed UI, who reported at least three episodes of urine loss per week during the preceding three months. Stress/mixed UI was confirmed by the Questionnaire for Incontinence Diagnosis. Exclusion criteria included body mass index ≥ 35, important pelvic organ prolapse (POPQ > 2), physiotherapy treatment or surgery for UI or pelvic organ prolapse in the past year, or any medication and co-morbidities/risk factors interfering with the study.
After an individual session with a physiotherapist to learn how to effectively contract the PFM, participants were randomized to a 12-week PFM training program, either in one-on-one sessions or as part of a group (8 women). Each weekly session lasted one hour and included a 15-minute educational period and a 45-minute exercise component, under the supervision of a trained physiotherapist. The 12-week training protocol comprised three 4-week phases, with gradual addition of increasingly difficult exercises, in terms of duration, number of repetitions and position. In addition, women in both intervention groups were expected to perform PFM exercises at home, five days per week for the first 12 weeks, and three days per week for the remaining nine months.
Blinded assessments were undertaken by trained physiotherapists at baseline, immediately post-intervention and at one-year post randomization. The primary outcome measure was the percent reduction in UI episodes per week, at one-year. Secondary outcomes assessed after 12-weeks and at one-year included: number of urinary leakages recorded on the 7-day bladder diary; amount of leakage on a 24-hour pad test; international Consultation on Incontinence Questionnaire modules on UI symptoms (ICIQ-UI short form) and quality of life (ICIQ-LUTSqol); patient-reported global impression of improvement questionnaire (PGI-I); satisfaction and willingness to undergo subsequent treatment. Complications and adverse events were recorded during the intervention, immediately after treatment and at one year.
Based on current literature and pilot data, we set the margin of non-inferiority at a maximum of 10% difference between the average percent reduction in the weekly number of UI episodes in the individual minus the group-based intervention arms. Accordingly, we calculated the sample size necessary to ensure 90% power for the 95% confidence interval (CI) for the intervention effect to exclude a difference higher than 10%, when assuming that the true effectiveness of the two interventions was equal. Accounting for up to 15% attrition rate at one-year, calculations for non-inferiority trials indicated the need to randomize 182 participants per group. To test our main hypothesis that group-based treatment is not inferior to individualized treatment, primary analysis was performed per protocol as recommended by CONSORT guidelines for non-inferiority trials. Additional sensitivity analyses were carried out using intention-to-treat (ITT) analysis on all randomized participants. As distributions of the primary and most of the secondary outcomes exhibited violation of the normality assumption, main analyses relied on the non-parametric Wilcoxon test.
From January 2013 through May 2017, a total of 362 participants were allocated to either individual (184) or group-based PFM training (178). Among these, 337 (93%) completed the post-treatment evaluation, and 319 (88%) completed the 1-year follow-up. At baseline, no imbalance of demographic characteristics was found between intervention groups. PFM training program attendance was excellent with an average of 11.7/12 (±0.6) sessions for individual training and 11.4/12 (±1.0) sessions for group-based PFM training. Home PFM exercises were performed 4-5 times/week by 89% of the individual PFM training participants and by 86% of the group-based PFM training participants during the 12-week intervention period (P=0.36), and thereafter at least once per week by 68% and 70% of women, respectively (P=0.63).
At one-year, the median reduction in the frequency of leakage episodes per week was 70% (95% confidence interval [CI]; 44% to 89%) in individual PFM training compared to 74% (95% CI; 46% to 86%) in group PFM training (difference -4%; 95% CI -10% to 7%; P=0.58) indicating that group-based physiotherapy is not inferior to individual physiotherapy (Figure 1). Both interventions were highly effective in reducing UI signs, symptoms and improving incontinence specific quality of life outcomes (Figure 2). Frequency of reported impression of improvement on PGI-I (85% vs. 86%) and satisfaction (90% vs. 91%) were high and almost identical in individual PFM training and group-based PFM training, respectively. Need for further treatment was low at 3% vs. 5%, respectively. ITT-based sensitivity analysis confirmed our results. No serious adverse events were reported.
Finally, similarly high proportions of women reached a minimal clinically important change (MCIC) in leakage episode frequency in both individual (72%) and group-based (74%) interventions. For the pad test, a MCIC in the amount of urine loss was somewhat more frequent in the group-based PFM training intervention (74% vs. 64%) but without reaching statistical significance. Finally, for ICIQ-UI short form and ICIQ-LUTSqol, frequency of MICC was almost identically high with 72%/74% - in the individual intervention and 73%/73% -in the group intervention.
Interpretation of results
To our knowledge this is the first and only adequately powered trial to assess non-inferiority of group-based physiotherapy compared to standard, individual physiotherapy for UI, both in the per-protocol and ITT analyses. There was no meaningful difference between the two groups for treating stress or mixed UI in women age 60 and over, one year after randomization. The group-based and individual physiotherapy interventions were very effective for all primary and secondary outcomes, after treatment and at one year. Furthermore, high and identical proportions of participants from both interventions reached MCIC in leakage severity and impact on quality of life at one year. The majority of participants attended most of the PFM training sessions and maintained their home exercise practice in both intervention arms. Loss to follow-up was low and equal, supporting the acceptability of each intervention. Group-based physiotherapy is a conservative management method, which can be offered to large numbers of women. Policy-makers should feel confident that women age 60 or over with stress or mixed UI will attend, adhere and gain benefit from group-based physiotherapy interventions.