We based our methodology on the Medical Research Council's conceptual model on how to evaluate complex interventions [25]. We will use mixed methods, including a quasi-experimental controlled trial (step 1) and semistructured interviews (step 2). For step 1, information on health conditions and medication utilization at study entry and follow-up (ie, after 6 months) will be collected for all included participants. Participating patients will also complete 3 validated questionnaires in either French or English to assess their perceived treatment burden, quality of life, and satisfaction with care. The Multimorbidity Treatment Burden Questionnaire, (MTBQ) [26,27], the EQ-5D-5L [28], and the self-administered Physician Enabling Skill Questionnaire (PESQ) [29] will be used. Pharmacists will also report their interventions and suggestions during the 6-month follow-up. In step 2, we will invite all health professionals who implemented the intervention (ie, pharmacists, nurses, and physicians) to participate in a semistructured interview where they will be asked about the ease to integrate the intervention into their practice, their satisfaction with it, and their perception of its impacts for older adults and their families.

First, nurses and physicians in intervention FMGs (exposed) will be asked to refer all older adults undergoing cognitive evaluation OR referred to a memory clinic OR recently diagnosed (<12 months) with a MNCD to the FMG pharmacist. Moreover, home care teams will be asked to refer older adults receiving home care for MNCDs to the FMG pharmacist. FMG pharmacists will perform a medication review. A medication review is a structured and comprehensive evaluation of a patient’s pharmacotherapy with the aim of identifying and resolving problems and improving health outcomes [30]. This medication review involves an interview with each referred older adult, including their caregiver, if applicable, to establish the best possible medication history (BPMH) [31]. The BPMH is a complete documentation of medication therapy, including the name, dose, administration route, and medication administration frequency. To this end, medication information will be validated with at least 1 other reliable data source (eg, community pharmacy records) and documented in the patient’s chart. The pharmacist will then detect medication-related problems [32] by analyzing this information and clinical data using their own judgment or validated criteria [33]. They will establish a care plan and the follow-ups needed with the patient in collaboration with the health care team and the community pharmacist. Pharmacists will also document the number and type of interventions and recommendations made to optimize patients’ pharmacotherapy for 6 months following the medication review. We will offer training and mentoring support to these teams, including pharmacist support for more complex interventions and monthly web-based interdisciplinary meetings.

The study will comprise participants from an exposed group of approximately 5 FMGs with attending pharmacists who will implement the intervention and a control group composed of participants from 5 others FMGs without involved pharmacists. FMGs and home care teams will participate at their convenience within 2 Integrated (University) Health and Social Services Centers territories: (1) the Centre intégré universitaire de santé et de services sociaux (CIUSSS) du Nord-de-l’Ile-de-Montréal (NIM) in a metropolitan area and (2) the Centre intégré de santé et de services sociaux (CISSS) de Chaudière-Appalaches (CA) in a mixed—urban and rural—area. We used this selection process to facilitate implementation as the intervention teams need to be motivated toward its application and training and mentoring activities. To be eligible for participation, intervention FMGs must have at least 1 participating pharmacist, interested partner nurses, and physicians. Similarly, intervention home care teams must have access to a participating FMG pharmacist and a home care nurse. In contrast, in control FMGs and their home care teams, no FMG pharmacist will be systematically involved in the eligible patients’ trajectory. However, the FMGs must be willing to allow data collection and delay implementing similar interventions within their teams.

For the first step, the intervention targets 2 patient subgroups at turning points: (1) older adults undergoing cognitive evaluation or recently diagnosed with MNCDs in FMGs and (2) older adults with MNCDs followed at home by the home care teams.

For the second step, FMG professionals and home care teams implementing the intervention are also targeted.

In step 1, we will use an intention-to-treat approach to assess the GPS intervention’s effectiveness. The principal analysis will compare the difference (1) in the average number of medications per person between the unexposed and exposed groups according to the study period (baseline and 6 months) and (2) the average number of potentially inappropriate medications per person between the 2 groups, according to the study period, by mixed Poisson repeated measures regression. This type of analysis considers intra-FMG and intraindividual correlations. Contrasts will be built to determine if the average number of medications per person over time differs between the exposed and the unexposed group. Secondary results will be analyzed similarly to compare the MTBQ, PESQ, and EQ-5D-5L scores between the groups exposed and unexposed to the GPS intervention.

Medication data missing at measurement time periods after the initial data collection will be imputed using the last-observation-carried-forward method. Using this method, the potential benefits of the intervention should be underestimated. For participants lost to follow-up regarding the questionnaires, no measures will be imputed at follow-up, but sensitivity analyses may be used in which baseline questionnaires scores could be imputed for follow-up scores. We will perform all analyses with SAS statistical software (SAS Institute Inc).

In step 2, we will record the interviews and transcribe them verbatim to then perform qualitative thematic content analysis. The first stage of this analysis consists of coding the data using NVivo qualitative analysis software (QSR International). We will develop a codebook based on Patton’s Theoretical Model of Change [34] for this purpose. Two research agents will independently test the first version of the codebook by coding excerpts of a few interviews. They will discuss and improve this version and test it again by coding excerpts of other interviews. This procedure aims to obtain a rigorous coding of the interviews and an accurate description of the experiences. The final sequential analysis will allow the integration of the quantitative results on the intervention’s impacts with the qualitative results describing the experiences of the FMG and home care teams. We will build and present an integration matrix and discuss it with the members of the research team and the stakeholders with whom the project was initiated, as well as with patient partners. This process will identify crucial elements for improving the intervention for future implementation.

According to mixed methods research methodology, the quantitative results obtained at step 1 will be interpreted and discussed in the light of information obtained from step 2, so that lessons can be learned on how to best implement such an interdisciplinary practice change, including the increased roles of some of the players [38].

Ethical approval was obtained from the Ethics review board of the CISSS CA and CIUSSS NIM (project number: MP-23-2020-732; latest amendments approved in January 2022).

The participant’s consent to participate in the study will be obtained in writing or verbally, recorded using an audio device. All participants will be able to withdraw from the study at any time without giving reasons. All patient data will be collected and recorded into REDCap, a secure and confidential data entry software. Nominal patient data will be entered into a data collection sheet and confidentially stored in an ongoing computer database of the research center (CHU de Québec).