Long-term effects of motherfit group therapy in pre-(MOTHERFIT1) and post-partum women (MOTHERFIT2) with stress urinary incontinence compared to care-as-usual: study protocol of two multi-centred, randomised controlled trials

The study consists of two multi-centred, randomised controlled trials (RCTs), namely MOTHERFIT1 and MOTHERFIT2. MOTHERFIT1 focusses on investigating PFMGT pre-partum and MOTHERFIT2 on PFMGT post-partum. Participants will be recruited in the southern part of The Netherlands from the Maastricht University Medical Center (MUMC+), Zuyderland MC (Heerlen/Sittard), Laurentius Hospital (Roermond), Maxima MC (Eindhoven) and surrounding midwifery practices. Except for Maxima MC, all obstetric centres are part of the Obstetric Consortium Limburg, a first-, second- and third-line obstetric midwifery maternity care collaboration. In every region, a registered PPT will provide motherfit group therapy. This protocol has been prepared in accordance with standard protocol items; recommendation for interventional trials (SPIRIT), a completed SPIRIT checklist is included as Additional file 1.

Women will be included if they meet all of the following criteria: (1) aged ≥ 18 years, (2) UI (stress or mixed with predominant stress UI factor, according to Haylen et al. [6], (3) a score of > 3 on the International Consultation on Incontinence Modular Questionnaire-Urinary Incontinence-Short Form (ICIQ-UI-SF) questionnaire [27], (4) are motivated for participation in the motherfit programme, (5) are competent to speak and understand the Dutch language and to read and fill in forms independently and (6) have a mobile app (mApp) on a tablet (Apple or Android) available.

Exclusion criteria are: (1) UI prior to first pregnancy, still existing during pregnancy, (2) high-risk pregnancy, resulting in a contra-indication for performing intensive pelvic-floor-muscle (PFM) exercises (e.g. placenta praevia, vaginal blood loss, preterm uterine contractions), (3) suffering from significant exercise limitations or co-morbidities (physical or psychological) that would restrain a woman from participation in motherfit group therapy, (4) a history of chronic neurological disorders or diseases related to UI (e.g. multiple sclerosis, cerebrovascular accident, diabetes mellitus (during ≥ 1 year with glycated hemoglobin (HbA1c) > 10 mmol/l)), (5) urinary tract infection (urine-sediment, urine culture), (6) a history of anti-incontinence or urogynaecological surgery, (7) women who are expected to be lost to follow-up (e.g. because of a planned change of residency), (8) have had recent pelvic physiotherapy (< 6 months) and (9) refusal to use a mApp.

After signing the informed consent, the participant will receive an email with a link to the electronic baseline questionnaires. Once the questionnaires are completed, block randomisation (block size is 4) will be done by a computer-generated sequence in a 1:1 ratio on the individual patient and location level. Allocation in blocks of 4 is concealed and done using a central computer. Participants are either allocated to the motherfit programme (intervention) or the CAU (control group).

Due to the nature of the interventions, the participants and PPTs cannot be blinded. During the trial the coordinating researcher is not blinded. However, once the participant has completed the questionnaires, it is not possible to make changes in the data due to locking of the questionnaires. Moreover, before the statistical analyses all participants will be appointed a new study number for which the coordinating researcher is blinded. Therefore, analyses will be done blinded for treatment allocation.

In The Netherlands, pelvic physiotherapy is a specialisation within the field of physiotherapy and has its own registration in order to guarantee quality [30].

Preceding the inclusion period, information, instruction and training on the standardised assessment and group therapy protocol will be provided to involved PPTs during a 2-h workshop. The PPTs receive a set of laminated A4 papers with a detailed description for each therapy session, containing: topics to discuss, PFM and homework exercises.

All data (electronic case report forms and questionnaires at baseline and follow-up) of the participants and case managers will be collected in a (web-based) digital central database. Demographic variables and personal characteristics will be registered by the Nederlandse Vereniging voor Obstetrie & Gynaecologie vragenlijst (NVOG-q) at baseline for MOTHERFIT1 and MOTHERFIT2.

MOTHERFIT1: data will be collected at baseline, 34 weeks of gestation, 6 weeks and 6 and 18 months post-partum.

MOTHERFIT2: data will be collected at baseline and 4, 9 and 18 months post-partum.

The case manager fills in a first survey after the inclusion of a participant. For MOTHERFIT1 these questions include, among others, expected delivery date and current medication use. Two weeks after delivery, case managers receive a second survey regarding delivery variables. For MOTHERFIT2 the case manager fills in identical surveys, except the question on expected delivery date.

Participants in the intervention group fill in a training diary and three questions regarding their general physical activity level, weekly. The PPTs will register attendance of the participant during the intervention period and send it by postal mail to the researcher.

Tables 1 and 2 show the schedule of assessments for MOTHERFIT1 and MOTHERFIT2. The primary outcome measure is self-reported UI based on the severity sum score of the ICIQ-UI-SF. The ICIQ-UI-SF is a brief (four questions) and robust measure for evaluating the frequency of symptoms and impact of UI [34]. The total score ranges from 0 (not affected) to 21 (severely affected). The ICIQ-UI-SF is divided into the following four severity categories: slight (1–5), moderate (6–12), severe (13–18) and very severe (19–21) [35]. The questionnaire is translated in Dutch [36]. Therapy success is defined as absence of UI or change from baseline of at least 3 points on the ICIQ-UI-SF at 18 months post-partum [37].

Patient-reported improvement: the Patient Global Impression of Severity (GPE) questionnaire assesses patients’ self-reported improvement [38]. It is an accepted and reliable scale for incontinence, consisting of one question and seven response options [39, 40].

Quality of life outcomes: the Incontinence Impact Questionnaire-7 (IIQ-7) contains seven items that reliably assess the impact of UI on health-related quality of life (HRQL) [41, 42]. It determines UI impact on four domains: mobility, physical functioning, emotional health and embarrassment and ranges from 0 to 100.

General activity level: the diary has to be filled in weekly. Next to a question regarding the number of days PFM exercises have been executed, it contains three questions regarding general activity level. The questions on general activity level are modified from the Dutch healthy exercise norm (Nederlandse Norm Gezond Bewegen). This norm is based on publications of the American College of Sports Medicine [43].

Only participants in the intervention group register their performance of requested pelvic-floor-muscle exercises, including their general physical activity, weekly, at home in the training diary. The training diary is a data entry form and, if scanned, an Excel file will be computer generated.

For the purpose of the economic evaluation, a study-specific cost questionnaire has been developed. Participants’ resource use ((in) direct costs related to SUI) is collected from the societal and health care perspective. Furthermore, the EuroQol instrument (EQ-5D-5 L) will be administered, a validated, generic health-state measure [43, 44] widely used in economic evaluations. The five-level version (EQ-5D-5 L) is proposed by the recently updated Dutch guideline for economic evaluations in health care [45] and consists of the EQ Visual Analogue Scale and a descriptive system. The descriptive system comprises five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension can be rated at five levels: no problems, slight problems, moderate problems, severe problems and extreme problems.

A study-specific questionnaire has been developed to evaluate patient satisfaction of motherfit group therapy (part 1, 10 items) and satisfaction with the use of the mApp (part 2, seven items). Questions on motherfit group therapy were, e.g. on whether the participant liked training in a group and if there were enough opportunities to ask the motherfit group therapist questions. Questions regarding satisfaction of the mApp were, e.g. on ease of use and whether participants would continue using the mApp after the intervention period. Each item ranges from 1 (strongly disagree) to 5 (strongly agree).

Assuming that the average score of the primary outcome measure (ICIQ-UI-SF; range 0–21) of MOTHERFIT1 will lie at 8 and for MOTHERFIT2 at 9 (which is also the expected mean ICIQ-UI-SF score at 18 months post-partum in the CAU group; in contrast, the expected mean ICIQ-UI-SF score at 18 months post-partum in the experimental group is 5 (for MOTHERFIT1 and MOTHERFIT2) together with a relatively high standard deviation of 5 at baseline (because of the non-normality of the measure), participants will – with 97.5% probability – vary at baseline within the ranges of 0 to 19. The minimum acceptable score of participants to be treated is set at 3, so the range lies approximately between 3 and 19. From earlier studies [44], it became clear that the success of the PFMT exercises will be considerable and will be clinically relevant if the gain is higher than half the standard deviation of the baseline, presumably 3 with a somewhat smaller standard deviation of 3, because of the homogenising effects in the experimental arm. In contrast to MOTHERFIT1 (women remain stable), in MOTHERFIT2 it is assumed that the condition of CAU participants at 18 months will worsen with an average ICIQ-UI-SF score going from 9 in the baseline to 10 (SD 5).

Assuming two-sided testing, a power of 90% (beta = 0.10) and a significance level of 0.95 (alpha = 0.05) in each arm of the trial in MOTHERFIT1, minimally 60, and in MOTHERFIT2, minimally 35 participants will have to be included without taking into account that participants may drop out of the study during the 18 months of observations. Using a 20% dropout, in MOTHERFIT1 each arm will need 75 participants, 150 in total, and in MOTHERFIT2 each arm will need 45 participants, 90 in total.

Analysis of the participants will be reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) Statement [44]. Data will be analysed according to the intention-to-treat principle. By preference, multiple-imputation techniques are used for missing values.

Firstly, descriptive, univariate statistics will be reported. In case of metric, normally distributed variables, mean and standard deviations are presented. If not normally distributed, medians and percentiles are presented. The Shapiro-Wilk test will be used to assess normality.

In both studies, the main hypothesis concerns differential changes in ICIQ-UI-SF within time between two randomised groups of participants. (Repeated measurements) analysis of covariance (ANCOVA) will be performed with baseline measurements (T0) as covariate. Transformations of original scores will be attempted if the ICIQ-UI-SF shows a non-normal distribution at T0. Randomisation groups (motherfit group therapy versus CAU) are regarded as a between factor. Next, the within-participants linear trend in time of the outcome will be calculated with the weights from the first orthogonal polynomial contrast and this is used as a dependent variable in a multiple (dummy-)regression analysis. It concerns repeated measurements from T0 to T4 (MOTHERFIT1) and T0 to T3 (MOTHERFIT2). Next to the baseline covariate measurement and the randomisation groups’ dummy variable, other possible confounding variables will be used in this multiple linear regression analysis of the linear trend in time of the ICIQ-UI-SF.

The following potential confounding variables are considered to be used in the multiple linear regression analysis: Body Mass Index (BMI) before pregnancy (> 25), parity, maternal age (> 35 years) and the ability to perform a PFM contraction at baseline.

Forward selection and backward elimination techniques will be used to determine the best fit of the data to a final regression model. Testing of interactive relationships between statistically significant effects of predictors in the final model will be done, especially if it concerns the experimental between-randomisation groups’ factor. Listwise deletion of missing cases will be used in all linear regression modelling. This may be in case of loss-to-follow-up because of a succeeding pregnancy during the follow-up period of 18 months. For the final best-fitting regression model, a residual analysis will be done on the standardised Studentised z-scores and a screening will be performed on outliers to ensure the legitimacy and validity of the use of parametric statistics in analysis by testing the normality of distribution of the linear trend in ICIQ-UI-SF.

Statistical analysis on the secondary outcomes of the study, such as the IIQ-7, the GPE and the EQ-5D-5 L will be handled in the same way as the primary outcome measure ICIQ-UI-SF. Process and structure indicators will be analysed with descriptive statistics and presented quantitatively as numbers and absolute and proportion data.

The cost-analysis will be performed from both a societal and health care perspective. Resource use will be measured in natural units and will be valued in monetary terms by multiplying these units by the costs per unit. If available, standardised, national cost prices (e.g. specified by the recently updated Dutch guideline for cost research in health care will be used [46]. Costs are distinguished into motherfit programme costs including the group sessions and home-based part and costs of the mApp (initial and replacement costs for ICT hardware and software), health care costs (e.g. use of incontinence materials, visits to the general practitioner, gynaecologist, midwifery costs, visits to the PPT, surgery, etc.), non-health care costs (e.g. travel costs and productivity losses) and patient and family costs (time spent on the programme, informal care costs). Data on (health care) resource utilisation associated with SUI will be prospectively recorded during the study by the participants. Other health care, non-health care and patient and family costs will be collected by means of a standardised cost questionnaire to be filled out by patients. Costs occurring 12 months after study inclusion will be discounted at 4% according to the Dutch guidelines for economic evaluations health care [47].

The outcome for the cost-utility analysis (societal perspective) is defined in terms of QALYs from inclusion up to 18 months post-partum. The number of QALYs is derived from the adjustment of survival data with HRQL. HRQL will be measured with the EuroQol-5D (EQ-5D) instrument, which provides a descriptive health profile and a Dutch valuation set for obtaining utility scores from the EQ-5D [45]. The outcome for the cost-effectiveness analysis (health care perspective) is based on the proportion of women with clinically relevant reduction in UI at 18 months post-partum. Outcomes occurring 12 months following study inclusion will be discounted at 1.5% according to the Dutch guidelines for economic evaluations of health care [47].

Next to the trial-based EE, a model-based EE will be performed, as it is expected that the economic impact of motherfit is best investigated by means of a long-term decision analytical model. First a structure and working model will be created that will facilitate the necessary analysis to be performed throughout the project. This model will be able to incorporate the values of all input parameters (both point estimates and uncertainty). Once the structure of the model is established, four essential types of data will be required: probabilities, costs, survival and health utilities (QALYs). Short-term costs and effectiveness data are readily available from the trial-based EE, whereas longer-term data may require synthesis of available evidence in the literature. Estimates of the economic impact will first be made using fixed estimates of probabilities, costs and health outcomes. Subsequently, a probabilistic sensitivity analysis will be performed which will address the joint uncertainty of the model inputs. As for the trial-based cost-effectiveness analysis, cost-effectiveness acceptability curves will be constructed. As with the trial-based EE, the model-based EE will address the cost per QALY (societal perspective) and cost per UI prevented (health care perspective). We will express uncertainty by means of confidence intervals and by creating cost-effectiveness acceptability curves. The appropriate time horizon will be agreed upon during the study but is expected to be lifetime.

A BIA will be performed according to the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) guidelines [48]. The BIA addresses the financial stream of consequences related to the implementation of motherfit group therapy and thus its affordability. The budget impact will depend, e.g. on patient acceptability of the programme, the uptake of the programme by health care professional and the target group, the cost-increase due to increased implementation of motherfit group therapy, and the cost savings due to preventing or reducing UI, i.e. reduced cost-of-illness. The structure and some data input of the decision analytical model developed for the EE will be adapted for the BIA. Input parameters will be based on results of the trial, national prevalence data, unit prices and tariffs obtained in the trial-based EE, and from the available literature when necessary. The analyses will be performed from different perspectives, including a health care budgetary perspective and a health insurers’ perspective. The model will take changes in the adoption/implementation of the programme, and patient acceptability/uptake into account and will compare different scenarios as regards to the swiftness and extensiveness of the uptake. In order to test the robustness of the results, sensitivity analyses will be performed. The time horizon will be varied from 1 year up to 5 years. No discounting will be applied.

Subjects can leave the study at any time and for any reason, if they wish to do so, without any consequences. The investigator can decide to withdraw a subject from the study for urgent medical reasons.

All women enrolled in the study will be followed and accounted for. Women who are unwilling or unable to commit themselves to the study plan and follow-up schedule (i.e. serious illness, during pregnancy, e.g. premature rupture of membranes, blood loss, severe high blood pressure, pre-eclampsia, movement out of the local area, etc.) may be withdrawn from the study. Women who will become pregnant again during the follow-up period of 18 months will be handled as dropout cases. Upon withdrawal of a subject, all documentation is available immediately for the investigators through the electronic case report file.