This study uses a retrospective, before-and-after cohort design involving patients who benefited from the EPOCA RPM system between February 2022 and August 2024.

Participants aged 70 years or older with polypathology (ie, having multiple chronic diseases, 2 or more affected systems) and who have been part of the long-term care pathway within EPOCA (at least 1 year) were included in the analysis.

The EPOCA RPM system [28] is a telemedicine solution offering secure home support for chronically ill older adults. It includes a 24/7 RPM platform for continuous remote medical monitoring and emergency assistance. The system facilitates coordination between care providers, GPs, and geriatricians to create personalized care plans. It consists of 2 main elements: a connected device (such as a wristband) for transmitting health data and an RPM center staffed by qualified and trained nurses and medical practitioners who can communicate with both patients and caregivers when needed. Every patient enrolled in the RPM system receives an initial home visit, during which their needs and health status are thoroughly assessed. They also receive teleconsultations and personalized installation of remote monitoring devices, as well as training and advice on the use of these devices. Further details about the EPOCA intervention can be found in our previous study, in which the geriatrician is the operator of remote care [26].

The outcome measures for this study were the total number of unplanned hospital admissions, including ED visits, the hospitalization rate per patient, and the cumulative length of hospital stay in days for both the total population and per patient. Outcomes were compared between the year preceding the patient’s enrollment in the RPM system (Y–1) and after 1 year of follow-up (Y).

Patient data are collected by medical and paramedical teams and compiled in the EPOCA RPM database at the inclusion and throughout the follow-up period. Data related to the study objectives for patients included are extracted from the EPOCA database for (Y–1) and the follow-up year (Y). To minimize potential bias in Y–1 data collection, a systematic and thorough approach is adopted when including patients in the RPM program, using medical records and direct communication with the patient’s care team. All data were anonymized before extraction and analysis. All analyses were conducted in a secure HDS (health data hosting) space.

Descriptive statistics were used to provide detailed information about the study cohort. Continuous variables (ie, sociodemographic and clinical data) were expressed by mean and SD. Categorical variables were reported as numbers and percentages.

We first used a Shapiro-Wilk test to assess the normality of the data distribution. We then applied the Wilcoxon test, a nonparametric approach, to compare outcomes in the year before (Y–1) and after (Y) the intervention. Statistical significance was defined at P<.05. All statistical analyses were performed using Visual Studio CodeTM (Microsoft), version 1.84.2, in Python language (version 3.11.5). No data imputation was required for the statistical analysis since all outcome data were fully available for comparison between the baseline year (Y–1) and the follow-up year (Y).

A generalized linear model with a negative binomial distribution and a log link function was used to account for overdispersion in the count outcome. Model selection was based on the theoretical relevance and statistical significance of predictors.

This study follows the MR004 framework of French research methodology; it requires a CNIL (Commission Nationale de l’Informatique et des Libertés) declaration and patient consent. However, ethics committee approval—equivalent to institutional review board approval—was not required, given the retrospective nature of the data analysis. In this context, the minimum dataset was declared to the French authorities under CNIL authorization number 2234081, in accordance with the General Data Protection Regulation (GDPR). All patients enrolled in the EPOCA RPM system systematically receive written information indicating that data collected from the EPOCA RPM database system, including medical records and follow-up outcomes, could be used retrospectively for research purposes. They also provide written consent for the use of their anonymized data following their enrollment in the EPOCA RPM system. No compensation was provided to participants.

Of the 235 eligible study participants, only 80 met the inclusion criteria (ie, aged 70 years or older, with polypathology and more than 1 year of follow-up) and were included in the analyses (see Figure 1). Patients’ characteristics are presented in Table 1. The mean age of participants was 86.7 (SD 8.3) years, and the mean of comorbidities was 6.7 (SD 3.7). Females represented 69% (n=55) of the sample size. Approximately 68% (n=54) of participants were assisted by a nonprofessional caregiver, and 55% (n=44) were living alone at home. The most prevalent referrals for the intervention were from nurses by 49% (n=39), followed by hospitals by 22% (n=18), EDs by 20% (n=16), long-stay units by 5% (n=4), and GPs by 4% (n=3). Further details on patient characteristics are provided in Multimedia Appendix 1.

Overall, the results of our study showed a significant improvement in all outcome measures between Y–1 and the follow-up year (Y) by the EPOCA RPM system. This was observed with significant reductions in health care use, detailed in Table 2. The overall number of unplanned hospital days dropped by 49%, which resulted in an average reduction of 9.7 days spent in the hospital per patient each year. The number of unplanned hospital admissions decreased by more than half (57%), while emergency room visits declined by two-thirds (62%). During the Y–1 period, 60 (75%) of individuals required hospitalization at least once. This number dropped to 21 (26%) over the follow-up period (Y).

To identify predictor factors associated with the number of hospitalization days in the follow-up period (Y), we conducted a secondary analysis using a generalized linear model. The results, presented in Table 3, identified several variables significantly linked to hospitalization duration, R²=0.84. For instance, compared to females , males were linked to a significantly lower number of hospital days (β=−3.862, P<.001). Similarly, greater autonomy, which can be seen through higher GIR (Groupe Iso-Ressources (iso-resources group)) scores (β=−.827, P<.001), and an increased number of hospitalization days in the Y–1 (β=−.037, P<.001) were both associated with shorter stays. Interestingly, the increased number of comorbidities was also a variable significantly linked to fewer hospitalization days (β=−.203, P<.001). In contrast, other variables such as age, functional status (Activities of Daily Living), and the number of professional caregivers did not show a significant association with the hospitalization days.

To further explore whether the impact of the RPM system varied based on patient characteristics, we conducted additional stratified analyses on the study’s main outcomes. These analyses were based on stratification by (1) the number of hospitalizations in the period Y-1 (Y–1: 0, 1, or 2+ hospitalizations), (2) the disability level using the GIR score (severe disability GIR ≤3 and no or moderate disability GIR ≥4), and (3) the presence or absence of a nonprofessional caregiver. The results are presented in Multimedia Appendix 2

Overall, the stratified analysis showed significant relative risk reduction across all groups, regardless of patient characteristics. Interestingly, when stratifying by hospitalization history, patients with 2 or more hospitalizations in the Y–1 period experienced significantly greater reductions across all outcomes compared to patients with 1 or no prior hospitalizations. Similarly, patients with severe disability (GIR ≤3) showed greater relative reductions compared to those with no or moderate disability. Finally, when we stratified by the nonprofessional caregiver status, both groups revealed significant reductions.

This retrospective study involved a before-and-after comparison of health care resource utilization among 80 polypathological older adult patients, who were followed for at least 1 year with the EPOCA RPM system when the GP was the system operator. Compared with the previous year (Y–1), a significant reduction, by half, was observed in the total number of unplanned hospitalization days and admissions, as well as ED visits. The main contribution of this study lies in providing additional evidence supporting RPM systems’ potentials in improving health outcomes and optimizing care delivery for older adult patients with complex chronic conditions.

Our results are similar to those of a before-and-after analysis carried out on 120 polypathological older adult patients (6.1 comorbidities, mean age 86.8; further details are presented in Multimedia Appendix 3), followed by the EPOCA RPM system, and when the system operator was the hospital geriatrician [26]. A reduction of 63%, 48%, and 48% was also observed in the total number of unplanned hospitalization days, admissions, and ED visits. This comparability of results suggests 2 main implications. First, the effectiveness of the EPOCA RPM system, as it is associated with a consistent improvement in patient outcomes across different settings in the French context. This was seen through a reduced use of health care resources, in particular the rate and duration of unplanned hospitalizations and ED visits, with 2 different system operators (ie, geriatricians and GPs). However, a large-scale randomized control trial (RCT) is required to confirm the effectiveness of our system compared to a control arm. Second, in the context of a limited number of geriatricians in France [7], such an at-home RPM solution, involving geriatric expertise, could be beneficial for the geriatric population. Particularly when the majority of patients aged 70 years and older have a GP. Only 5.6% of patients aged 70 to 79 and 5.7% of patients aged 80 years and older do not have a GP [29]. Therefore, our study underscores the potential benefits of adopting RPM solutions, not only by geriatricians, but also by GPs and local health care institutions, as a useful strategy to manage complex conditions, prevent clinical deterioration, and thus reduce burdens on the French health care system.

Our results also align with previous findings suggesting a positive impact of RPM systems on patient clinical outcomes and overall health care improvement [30,31], particularly in reducing hospitalization risk, as demonstrated in both national and international studies. In previous RCTs, the risk was lowered by 57% [23], 28% [22] in intervention groups compared to conventional care groups, and 33.2% in a before-after cohort [32] (Table 4). However, in terms of ED visits, our findings differ from those of another French RCT, the eCOBAHLT study, which showed an increase in ED visits uncorrelated with any increase in hospitalizations. This increase was attributed to the unavailability of treating physicians, leading patients with manageable health conditions to seek care in the ED [22]. Contextual and methodological challenges may also explain these inconsistencies.

Multiple predictors were significantly associated with reduced hospitalization duration in the follow-up year (Table 3). The strongest predictors include male sex, patient disability level, number of hospitalization days in the Y–1 period, and, interestingly, number of comorbidities. Combining these findings with complementary stratified analyses, the results showed that the greatest effect of the RPM system was observed in patient subgroups with a history of 2 or more hospitalizations, severe disability (GIR score ≤3), or without a nonprofessional caregiver. These observations suggest that high-risk patients may benefit the most from the EPOCA RPM system, likely reflecting its structured and coordinated care model. This interpretation is supported by the fact that the RPM team works closely with external and ambulatory health care professionals (mean of 2.4 professionals per patient) [26]. This approach allows proactive management of patients with complex health conditions, thus stabilizing their health status, preventing complications, and avoiding hospitalizations [33]. It is important to note that, among the subgroups, for patients without hospitalization history in the Y–1 period, the effect of the RPM was difficult to measure. In future research, this variable, history of hospitalization, could be considered as an inclusion criterion to minimize bias and improve the precision of effect estimates. Other variables such as GIR score and caregiver status could be used for more detailed stratified analyses in a large-scale study.

Finally, while the finding of a correlation between the number of comorbidities and a reduced hospitalization duration seems counterintuitive, it can be explained by the specific profile of patients followed by the EPOCA RPM system. These patients are predominantly polypathological, with an average of 6.7 comorbidities. Such complex profiles represent the ideal target population who may benefit the most from close home-based care and personalized health management through RPM systems like EPOCA’s. Looking ahead, further in-depth research on RPM interventions is therefore essential to fully understand and measure its impact on care systems and health outcomes of polypathological older adult patients.

Several risks and challenges may be associated with RPM interventions, including overuse of digital monitoring and sustainability issues in patient care. Previous findings suggest that over-monitoring can be associated with psychological, physical, and other systemic risks, which may entail staff stress, reduced performance, and turnover [34]. Alarm fatigue, here, in health care, driven by excessive and often false alarms, as well as over-surveillance, may endanger staff mental health and patient safety [35]. Although these aspects were not examined in this, they merit particular attention and should be explored in future research. On the other hand, the long-term sustainability of RPM interventions, particularly in older adult patient care, can be a major challenge. We need to ensure their capacity to use resources wisely while maintaining high-quality care. To achieve this, it is essential to facilitate its adoption within health care systems, for example, through starting better reimbursement models that support their use and impact. Therefore, to help integrate the RPM system into older adult care on a larger scale, we suggest the following practical policy changes (Textbox 1).

Several limits must be acknowledged in this study. First, it is based on retrospective data, possibly leading to information gaps and potential biases, particularly regarding the data collected for the Y–1 and Y periods. To mitigate this, multiple measures have been taken to ensure comprehensive data collection, especially for the Y–1 period, using patient medical records and direct contact with their care team.

Second, the study sample size was relatively limited, as we only included patients with more than 1 year of follow-up to facilitate a more robust comparison between the 2 periods. This inclusion criterion was not considered in the first study, where patients with less than 1 year of follow-up were included in the analysis. Although the results were adjusted, this difference could potentially affect the comparability of the results in terms of different operators of the system (ie, geriatrician or GP).

Third, some factors such as differences in access to care, disparities in socioeconomic status, and adherence to RPM interventions were not specifically addressed in our analysis. Our primary focus was on measuring the impact on unplanned hospitalizations; we also provided secondary analysis related to patient characteristics, including age, sex, GIR score, comorbidity, and others. Looking forward, variables such as care accessibility, adherence to RPM, as well as cost and sustainability issues, could be further explored in future studies.

Fourth, this study covers the period between 2022 and 2024. It is important to acknowledge that following the COVID-19 pandemic, health care systems across Europe, including in France, have experienced an accelerated adoption of digital health solutions, such as remote monitoring systems. We believe that the observed effectiveness of the EPOCA RPM system in supporting the care management of complex older adult patients is reflecting the broader shifts in health care practices.

Finally, the absence of a specific control group, such as a usual or standard care group, which may not be feasible retrospectively, limits the generalizability of our findings. To address these limitations, a large-scale randomized controlled trial is set to begin in 2025 (NCT06845917) to generate more robust, high-level evidence.

This study showed that, regardless of the system operator, the EPOCA RPM system can help in reducing unplanned hospitalization days and numbers, as well as ED visits, potentially alleviating hospitals’ burdens. These findings highlight the possibility of a new organization of at-home care for complex older adult patients and suggest RPM systems such as EPOCAs, as a potential alternative solution, especially in the context of limited geriatric resources. In perspective, further evidence regarding the effectiveness of our RPM system can be obtained through a large-scale randomized controlled trial.