A multimodal and multidisciplinary program to prevent loss of mobility in patients aged over 70 years: study protocol of a multicenter cluster randomized study in primary care (the PRISME-3P study)

PRISME-3P is a national, multicenter, and cluster randomized trial conducted in France (Fig. 1). GPs will be randomly assigned to the intervention group or to the control group. Recruitment and inclusion will be performed by GPs and the intervention will take place at the hospital and home of the patient. The planned inclusion period is 24 months with 12 months follow-up.

In both groups, GPs will perform a consultation for inclusion at baseline and two evaluations at 6 and 12 months thereafter. In the intervention group, an initial blood test will be performed (blood cell count, CRP, pre-albuminemia, and albuminemia). Follow-up will include two visits to the allocated study hospital outpatient clinic (within 2 weeks and at 3 months from inclusion) and 3 visits to the GP (at 3, 6, and 12 months from inclusion). All data collected will be notified by the GPS on the case report form (CRF).

First, we have searched for hospital that could participate in the study (having a geriatric unit with a dietician). We have invited 8 geriatric units to participate. Second, we have contacted the university general practice department of each hospital areas, i.e. Lyon (Lyon, Villefranche-sur-Saône, Bourg-en-Bresse, Annemasse), Nantes (Nantes), Saint-Etienne (Saint Etienne) and Nice (Menton). In each department a coordinator will be designated with responsibility to recruit GPs near the hospital.

A total of 60 GPs will be recruited in 8 areas of France (6 in Nantes, 6 in Dijon, 19 in Lyon, 8 in Villefranche-sur-Saône, 5 in Bourg-en-Bresse, 5 in Annemasse, 6 in Saint-Etienne, and 5 in Menton). They will be assigned to the intervention group or to the control group using the survey select procedure of SAS® (SAS Institute Inc., Cary, NC, USA). The sampling will be stratified on the referring geriatric center (i.e. city). GPs working in the same practice will be considered as a unique sampling unit to avoid risk of contamination. Each collaborating university general practice department will organize training for the screening and assessment of patients required by the study protocol, including the provision of information on the eligibility criteria and study organization.

All subjects aged over 70 years old will be eligible except patients with functional or locomotor disabilities, a life expectancy less than 6 months, a progressive chronic disease affecting participation in the program, or a MMSE< 20.

For each consecutive patient over 70 years of age without exclusion criteria, GPs will screen patients with medical interview to detect 2 or more inclusion criteria described in Table 1. These criteria are partly derived from the sarc-f screening tool for sarcopenia [28].

In the presence of 2 or more inclusion criteria, GPs will assess the patient using the Short Physical Performance Battery (SPPB) [29]. In the event of a SPPB score < 10 (out of 12), patients will be given 9 questionnaires grouped in a unique folder: Activity Scale for the Elderly (PASE) [30], weekly sedentary time inspired by Gardiner [31], Activity of daily living scale (ADL), Instrumental activities of daily living scale (IADL) [32], SF-12 scale [33], Mini-Nutritional Assessment (MNA) [34], EPICES score (Evaluation de la précarité et des inégalités de santé dans les centres de santé (in French)) [35, 36], Charlson index [37], Geriatric Depression Scale (GDS) [38] (Table 2) that measure their autonomy, nutrition and physical activity, and will be proposed another medical appointment for inclusion in the study. If a patient refuses to participate in the study, this will be noted in the CRF.

At the hospital, a three-step intervention are planned. First, participating geriatricians will perform a standardized geriatric assessment of each patient to rule out underlying diseases that might cause exhaustion or weight loss and hamper participation in the study. Second, a dietician from the study center will evaluate the nutritional status of patients (Mini Nutritional Assessment, weight, Body Mass Index, weight loss) and analyse the dietary intake reported on a three-day survey given to the patients to complete prior to the visit. On the day, the dietician will also perform calorimetric measurements to estimate dietary needs of each patient. A dietary strategy will be planned and agreed with the patients. The dietician will give written advice to each patient about calorie, protein, calcium, omega-3 fatty acid, and vitamin requirements (Fig. 2). Third, the patient will meet a personal trainer, who is specialized in physical exercise in the older age. A physical training program has been developed for the purpose of the study. It is based on existing validated training programs, with sets of exercises dedicated to three specific targets: muscle reinforcement, balance/flexibility, and endurance. It will consist of daily exercises lasting approximately 30 min and performed 5 days a week. These exercises are simple and require no specific equipment. Three sessions are dedicated to muscle reinforcement (lower limbs, upper limbs, and chest), and two to balance/flexibility and endurance. Intensity will increase over time to improve muscle strength, tailored to the abilities of individual patients. This program will be supported by a dedicated personal trainer, who will evaluate the patients: the personal trainer will explain and demonstrate each exercise to the patient (and caregiver), then the patient will practice each exercise in front of the monitor and be given the opportunity to rectify, and once the monitor has checked that the exercises are fully understood, he/she will hand the patient and caregiver a written handbook with illustrated explanations of the exercises for home training. Patients will be asked to report their activity and adherence to the program in the handbook. A monthly telephone follow-up will be set up, by which the monitor will be able to provide some advice and encouragements to the patients.

By contrast with the control group, patients in the intervention group will visit the GP at 3 months, who will perform an intermediate SPPB assessment in order to adapt the intensity of the intervention. If SPPB improves by one point or more, no change will be planned, if not they consult the same geriatrician and dietician, and will receive again the physical training program from the personal trainer who will demonstrate the exercises and propose individual (home) or group training sessions (hospital) twice a week for 10 weeks.

The inclusion of 28 GPs per arm will provide a power of 80% to detect a difference of 0.8 points in SPPB score change between arms (1.2 points increase in the control group, versus 2 points in the intervention group), with 5 patients for each GP, an intra-cluster correlation coefficient of 0.15, and with a standard deviation of inter-individual variability of 1.7. Accounting for loss to follow-up, it is planned to randomize 30 GPs in each group, giving a total of 60 GPs and 300 patients (5 patients for each GP). The power of the study was calculated using the method described by Donner and Klar [40] and with the package CRTsize in the R program, version 3.3.1 (R foundation, Vienna, Austria).

The patient characteristics at inclusion will be described according to each group (intervention and control groups). A summary table will be provided in order to ensure the homogeneity of the prognostic factor distributions according to the groups. A second summary table of clusters (GPs) will also be provided. The method and the presentation of the results will be according to the guidelines for cluster trials [41].

The analysis used for the primary outcome will be used to evaluate the change of the secondary outcomes between the inclusion and the 6-month follow-up visit.

The analysis of the change of secondary outcomes in the interventional group will be performed using a longitudinal mixed effect linear model with a fixed period effect and as random effects: the center, the GP, and the patient nested within the GP. This analysis will be adjusted on the age at inclusion.

A univariate analysis will be performed using a longitudinal mixed effect linear model to identify the factors associated with the SPPB result at 6 months in the interventional group. The model will include as random effects the center, the GP and the patient nested within the GP, and as fixed effect the corresponding factor. This analysis will provide the effect estimation of each factor with its 95% confidence intervals. The independent factors found significant in this univariate analysis will be included in a multivariate analysis.

All analysis will be performed using the statistical analysis program SAS® version 9.4 (SAS Institute Inc., Cary, NC, USA), and the R program, version 3.3.1 (R foundation, Vienna, Austria).

The study protocol was approved by the Sud Est 4 Ethics Committee on October 18, 2016 and cover all sites involved in this study. The research carried out will be on accordance with the Helsinki Declaration and ICH GCP Guidelines. The study complies with the principles of the data protection act in France. Each GP had to collect a writer consent at the beginning of the procedure. This consent is retained in the CRF. The patient can stop the study at any time with an oral information at his GP.

The study will be randomized and multicenter, and will investigate the effect of a specific clinical pathway with both the effects of the intervention and, simultaneously, the implementation of the program in current practice conditions, across various areas of France. Rather than an individual randomization, a cluster design was chosen in order to limit the risk of contamination, but also for reasons of feasibility and ethics; it would have been difficult to justify why two different patients in the same practice would not have access to the same management. Likewise, we chose to stratify randomization to ensure that GPs working in the same practice would be randomized to the same group. The high number of clusters (60 in total) and low number of patients within each cluster should help reduce the risk of bias due to intra-cluster correlation.

Despite the cluster design, the effects of the intervention will be determined individually. An open-label design is necessary given the nature of the intervention. The effects of the intervention will be evaluated using an objective outcome, the SPPB, which has shown to be reproducible [29, 39]. An SPPB score below 10 is predictive of all-cause mortality [42] and reveals a weakness of lower extremity function and higher mobility disabilities [43]. Clinically, an improvement of one point in the SPPB score is associated with substantial and clinically relevant changes in physical performance [44]. SPPB is a reliable tool to screen for patients likely to benefit from a physical exercise program [41] and it is generally considered as a valid predictive tool to assess physical performance and predict risk of mobility loss [11, 39, 42]. The GPs will screen systematically each consecutive patient over 70 years of age to limit selection bias. Furthermore, they will note each refusal to participate in order to evaluate the acceptability of the program from the patients’ perspective.

Standard training will be provided to the participating GPs for the recruitment and to physical trainer to ensure harmonized delivery of information. All GPs will be given specific training in group sessions for the loss of mobility screening assessment, prior to randomization. This will ensure that the screening and inclusion procedure is the same for all patients, and ensure comparability between groups for the measurement of the primary outcome. Moreover, GPs in the intervention group will not be involved directly in the intervention itself.

Patients with an estimated life expectancy of less than 6 months will be excluded as they would derive no benefit from intervention. In France, patients choose their GP and generally trust him/her. We believe this privileged relationship might help acceptance and adherence to the program, and also limit the rate of patients lost to follow-up. Conversely, part of the intervention will take place in a hospital setting and this might limit the acceptability of the program. This was decided for feasibility reasons, as this limited the number of personal trainers required and helped to optimize their time by avoiding multiple home visits.

The recruitment of patients from primary care will be facilitated by the collaboration and institutional support of the university general practice departments in each participating area. To facilitate GP adherence and acceptability of a screening, GP needs a valid, simple and rapidly administered tool. The FRASI-Study found that the use of SPPB is feasible in daily practice in primary care [45]; it is a simple tool that has good performance to predict incident ADL disability, worsening mobility, and death [39, 43, 46]. We have estimated the time required to deliver the SPPB to be less than 20 min, which corresponds to the mean duration of a regular GP consultation in France. To minimize the duration of the inclusion visit, GPs will screen and assess patients and give them a questionnaire during a routine visit. They will be invited to read this questionnaire and answer the questions to the best of their abilities, which will be then completed with the help from GPs during a dedicated inclusion visit.

To the best of our knowledge, this will be the first study in France in which a multimodal intervention with cooperation between geriatricians, general practitioners, dieticians, and personal trainers will be implemented. We will be able to evaluate the implementation of a collaborative program that may help setting-up effective pathways for prevention of loss of mobility in primary care.