This single-center, single-blind, 2-group randomized controlled trial was conducted at CHUM between August 2022 and September 2023. Patients were divided into 2 groups: one group received perioperative support through a personalized telehealth platform (LeoMed), while the other group received standard care. The study protocol was published in April 2023 [9].

All patients over the age of 18 undergoing elective day surgery at CHUM were eligible. The inclusion criteria required patients to have internet access and be able to understand both written and spoken communication in either French or English. Patients were excluded at two stages: (1) before randomization, if they did not meet the inclusion criteria or chose not to participate, or if surgeries were postponed, and (2) after randomization, if patients withdrew their consent, met exclusion criteria not identified during the telephonic initial approach, were hospitalized on the day of surgery, or if surgeries were postponed at the last minute.

Patients were initially approached by phone 2 days before their surgery, based on the operating schedule. To manage the resources available in the research unit, we limited the number of patients included each day to a maximum of 5. During the assessment, a research assistant verified the inclusion criteria and provided verbal information about the study. If the patient agreed to participate, they could provide written consent either electronically via email or give verbal consent and sign in person on the morning of their surgery upon arrival at the day surgery unit. All enrolled patients were electronically randomized in a balanced 1:1 ratio into either the intervention group or the control group, using blocks of 2-4. The randomization list was created with the MedSharing Randomizer for Clinical Trials web application. Patients were not informed of their randomization group; however, due to the nature of the intervention, they were not blinded to their group allocation.

The intervention consisted of perioperative support through a personalized telehealth platform (LeoMed Technologies Inc). Specifically, patients in the intervention group received a link the day before their surgery, prompting them to access the platform to prepare for their procedure. Preoperatively, the platform directed patients to CHUM’s digital portal, which centralized health sheets outlining the perioperative process. In addition, customized psychological support videos, developed in collaboration with CHUM’s communications team, were made available to address patient concerns and provide comprehensive guidance. Postoperatively, the platform sent electronic follow-up forms to participants’ phones daily during the first 4 days and again on the seventh day, prompting them to evaluate their health status. These forms included questions about potential surgical site infections and symptoms such as redness, pus, fever, and wound pain. Patient responses were analyzed automatically by an artificial intelligence algorithm, generating alerts visible to CHUM’s telehealth service team. This service, staffed by approximately 50 specialized nurses trained in remote health care delivery, provides remote support and coordinates interventions with CHUM’s medical and paramedical teams. As part of this project, nurses contacted patients within a maximum of 8 hours following an alert to ensure appropriate care was delivered.

Participants assigned to the control group received the hospital’s standard care, which included a preoperative phone call from a scheduling agent on the day before surgery. Postoperative follow-up for control group patients depended on the surgical specialty, with the timing of follow-up visits varying by surgeon and type of surgery, typically ranging from a few days to several weeks. However, like patients in the intervention group, control group patients were required to log in to the platform the day before surgery. At that time, they completed the preoperative quality of life assessment score (EQ-5D-5L VAS), as in the intervention group.

In both groups, the platform sent follow-up evaluation forms on postoperative days (POD) 14 and 30 to assess variables needed for the medico-economic analyses. These included the postoperative EQ-5D-5L VAS score, health care usage, and patient satisfaction.

To ensure compliance in the intervention group, we developed a Compliance Index to assess patient engagement with the telehealth platform during the immediate postoperative period, specifically the first 4 days. Compliance was measured by calculating the frequency of questionnaire responses, which served as a proxy for engagement. A participant who responded to all 4 questions received a compliance score of 1, while a participant who did not respond to any received a score of 0.

Demographic data (age, gender, type of surgery, and type of anesthesia) were extracted from CHUM’s electronic medical record (EMR) system (Oacis; Telus Health). Economic and health-related variables were collected from the telehealth platform via electronic questionnaires completed by patients preoperatively and on POD 14 and 30. To ensure the accuracy of the information provided, all patients who reported complications or events, such as any health care usage, were systematically contacted by our research assistants via phone. Patients who did not complete the electronic questionnaires also received a reminder call. For each patient, the data were systematically recorded and securely stored on a local server before being extracted into 2 Excel (Microsoft) databases. These data were cross-referenced with those extracted from the EMR.

Data analysis was conducted using R. Demographic data were summarized using descriptive statistics. Categorical data were presented as frequencies and percentages, while continuous variables were reported as means with standard deviations. Health care service usage was analyzed alongside associated costs and carbon dioxide (CO₂) emissions. All financial calculations were expressed in Canadian dollars. QALYs were calculated using EQ-5D-5L scores converted into utility values based on the Canadian Time Trade-off Value Set [10]. Utility scores were derived using a predefined function and adjusted for the number of days since surgery (14 or 30 days) divided by 365. This method accounts for differences across all EQ-5D-5L dimensions. Our analysis primarily followed a modified intention-to-treat (mITT) principle, whereby all randomized participants were included except those with completely missing outcome data for a given analysis. No imputation was performed, and participants with incomplete medico-economic data were excluded from the corresponding analyses. To detect differences between the 2 groups, as well as gender differences, we used the chi-square test, Mann-Whitney U test, or independent Student t test, as appropriate. This study adhered to the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) 2013 Statement [13] and the CHEERS (Consolidated Health Economic Evaluation Reporting Standards) guidelines [14] (the CONSORT [Consolidated Standards of Reporting Trials] checklist is provided in Checklist 1).

Our sample size calculations were based on the study of Jaensson et al [15], who suggested that better postoperative recovery—indicated by a reduction in symptoms and complications—would likely result in fewer postoperative calls to health information lines, visits to the emergency department, unplanned readmissions, and associated medical visits, thus lowering overall costs [15]. In their study, they found that patients in the intervention group, who used a recovery assessment mobile app, exhibited better recovery outcomes (as measured by the global Swedish Web Version of the Quality of Recovery [SwQoR] score) compared with those receiving standard care on POD 7 (effect size 0.21; P<.001).

Using a significance level of .05 and a statistical power of .90, the necessary group size was calculated by G*Power (Heinrich Heine University) to require 357 participants per group. To accommodate a 20% loss to follow-up and potential protocol deviations, we planned to recruit 500 patients for each arm of the study, for a total of 1000 patients.

The study adhered to the principles of the Helsinki Declaration and was approved by the institutional review board and research ethics committee at CHUM (21.110). The trial was registered with ClinicalTrials.gov (NCT04948632).

Between August 2022 and September 2023, 1411 patients were assessed for eligibility. After excluding 197 patients, 1214 participants were randomized into 2 groups: 608 in the control group and 606 in the intervention group (Figure 1). After randomization, 108 participants from the control group and 106 from the intervention group were excluded, primarily due to the following reasons: withdrawal of consent (n=34), surgery cancellations or postponements (n=68), technical problems (n=46), conversion to inpatient hospitalization (n=40), or other reasons (n=26). A total of 55 patients in the control group and 64 in the intervention group did not return follow-up questionnaires or respond to telephonic contacts, resulting in a loss to follow-up. Figure 1 illustrates the trial flowchart. Baseline characteristics are described in Table 2. Compliance was estimated as 0.89 in the intervention group.

No significant differences were observed in unanticipated postoperative health care usage between the groups (Table 3). The mean cost of health care usage was CAD $370 (US $275) in the control group compared with CAD $323 (US $239) in the LeoMed group (P=.60). No significant difference was observed for the ‘Costs+Transportation’ variable (US $374 vs US $323; P=.33).

At POD 14, the intervention group demonstrated a statistically significant QALY gain of 0.002 compared with the control group (P=.01). Among the 5 dimensions of the EQ-5D-5L, autonomy, daily activities, and pain showed significant improvements in the intervention group at POD 14 (Table 4). At POD 30, the difference was no longer statistically significant (QALY gain=0.001; P=.14). No significant differences were observed between the groups for the EQ-5D-5L VAS scores at POD 14 and 30 (Table 4).

There were also no significant differences in GHG emissions between the groups, with the intervention group emitting an average of 0.870 kg CO₂-eq compared with 1.055 kg CO₂-eq in the control group (P=.52). However, patient satisfaction was significantly higher in the intervention group at both POD 14 and POD 30, with mean scores of 8.9 versus 8.4 (P=.02) and 8.8 versus 8.1 (P<.001), respectively.

This single-center, 2-group randomized controlled trial highlights the potential utility of telemedicine platforms in enhancing postoperative follow-up for ambulatory surgeries. While we observed no statistically significant reduction in unplanned health care usage, our study demonstrated a gain in QALYs at POD 14 in the intervention group. With willingness-to-pay thresholds for QALYs ranging from CAD $47,048.84 to CAD $73,936.87, the estimated economic benefit range from CAD $94 to CAD $148 per patient [16].

Several factors may explain these findings. Although the study did not demonstrate a significant difference in cost-effectiveness, we observed a trend toward reduced health care visits in the intervention group, similar to results reported by Temple-Oberle et al [17], who found no significant differences in unplanned health care contacts in a randomized trial evaluating a smartphone app for postoperative monitoring. This lack of significant results may reflect limited statistical power due to the sample size. Nevertheless, our findings contrast with previous studies that reported more favorable outcomes associated with app-based follow-up on health care usage. For instance, Armstrong et al [18] demonstrated a significant reduction in in-person visits following ambulatory breast reconstruction using a mobile app [18], while Lee et al [19] reported low rates of unplanned emergency department visits in a same-day discharge protocol for minimally invasive colorectal surgery supported by mobile health tools. Despite these differences, the observed QALY improvement at day 14 in our study suggests that patients in the intervention group may have experienced a faster subjective recovery during the initial postoperative period. At day 30, recovery levels were similar between groups, possibly reflecting natural convergence in postoperative recovery trajectories and a psychological adaptation to residual symptoms such as pain, fatigue, or reduced autonomy. While the QALY gain observed at postoperative day 14 (0.002) was statistically significant, its clinical interpretation remains modest. In the literature, the minimal clinically important difference (MCID) for QALYs varies widely depending on the population and condition. As emphasized by Revicki et al [20], the interpretation of QALY gains should consider the nature and duration of the intervention as well as the size of the target population. However, small utility gains may still be meaningful when considered at a population level, particularly in large-scale procedures such as ambulatory surgeries. In this context, even a modest per-patient gain could represent substantial cumulative benefit, especially given the approximately 2.3 million ambulatory procedures performed annually in Canada [2]. Future studies should aim to define MCID thresholds specific to short-term digital perioperative care interventions.

Despite the promise of telemedicine in postoperative care, few studies have demonstrated clear benefits of digital health solutions [21,22]. This is partly due to the heterogeneity of available platforms and the challenges of conducting large-scale multicenter trials with standardized solutions. Our study also showed higher patient satisfaction in the intervention group, corroborating existing literature suggesting that digital follow-up addresses a critical gap in postoperative care [23]. For instance, Day et al [24] demonstrated that automated messaging during the perioperative period enhanced satisfaction, improved health literacy, and fostered collaborative decision-making between patients and care teams. In addition, a meta-analysis of telemedicine interventions found that over 80% of patients receiving remote follow-up reported high satisfaction levels [25]. In line with these results, the high compliance index observed in our study (0.89) reflects strong patient engagement with the telehealth platform, suggesting both good implementation feasibility and user acceptability.

The single-center design of this trial limits the generalizability of findings to other settings, particularly those involving minor surgical procedures. However, this limitation underscores the broader challenges of implementing standardized telehealth platforms across multiple institutions. The observed trend toward reduced unplanned health care contacts and costs might have reached statistical significance with a larger sample size. In addition, the lack of participant blinding, inherent to the nature of the intervention, may have introduced response bias in subjective outcomes such as satisfaction and perceived recovery. This limitation is compounded by the use of self-reported measures which, although based on validated instruments, remain susceptible to perception and recall biases that could influence reported satisfaction or quality-of-life ratings. Furthermore, since all participants in the intervention group received access to the full LeoMed platform, we were unable to determine which specific components such as psychological support videos or educational modules contributed most to the observed benefits.

Moreover, health care costs may have been underestimated. For instance, hospitalization costs were calculated based on single-night stays and excluded extended emergency department visits that were not classified as formal admissions. Diagnostic testing costs associated with emergency visits were also not included.

Finally, we were unable to identify subgroups that might benefit more from the platform due to heterogeneity across age groups and surgical specialties. Although randomization was expected to balance baseline characteristics, some potential confounders such as previous experience with telehealth were not systematically collected and could not be adjusted for. Future research should focus on identifying high-risk populations that might derive the greatest benefit from telemedicine interventions.

In conclusion, with the expansion of ambulatory surgery and the inclusion of more comorbid patients and increasingly complex procedures, ensuring effective follow-up care is becoming increasingly important. Our study demonstrated promising results regarding the cost-utility benefit of telemedicine platforms for perioperative monitoring. However, further large-scale, multicenter studies are needed to better determine the exact medico-economic impact and the clinical benefits for the broader patient population.