A prospective, randomized controlled trial was conducted in 2 university-affiliated hospitals (Huashan Hospital of Fudan University in Shanghai and The First Hospital of Fujian Medical University Hospital in Fujian, China). Randomization was 1:1 into the digital health group (intervention group) or the rehabilitation manual-based group (control group). They were requested to complete 2 circles of assessments with baseline as the first time before surgery and subsequent assessments after 3 weeks’ intervention.

This study was intended for patients with PLC after hepatectomy. Participants aged 35 to 75 years who met the following inclusion criteria were approached to participate: (1) patients with PLC expected to undergo hepatectomy 2‐4 liver segments, progressive PLC (stage 2 or 3) for open or laparoscopic hepatectomy; (2) Child-Pugh A; (3) patients with an oncology history of PLC that has not recurred within 3 years after surgery; and (4) ability to operate a mobile phone. Notably, patients were excluded if they had exercise, psychological, or cognitive disorders before hepatectomy.

Appropriate participants were jointly identified by the chief rehabilitation therapists and chief surgeons. Participants were given ample time to consider and voluntarily choose to participate in the study, with written informed consent obtained the day before surgery. Before obtaining consent, researchers thoroughly explained the research procedures and detailed the preoperative and follow-up data to be collected in the ward. The methods and specific data content were clearly outlined in the informed consent form. All participants received standard postoperative care in the operating room, postanesthesia care unit, and ward.

The R Plus Health app (Recovery Plus), a medical device cleared by the Chinese FDA, served as the underlying software and hardware for the intervention in this trial, including the basic app structure and the target heart rate belt to guide and supervise training exercises. Based on this, a 3-week exercise program specifically designed for patients with PLC after hepatectomy was developed by an oncology exercise rehabilitation team, which consisted of 3 hepatobiliary surgical oncologists, 2 rehabilitation doctors, 2 nurses, and 3 physiotherapists. After the preliminary literature review and expert demonstration meeting, a set of exercise rehabilitation plans for patients with PLC surgery after discharge was discussed [11]. Reference to the framework in the guide for "FITT Principle” (Exercise prescriptions include the 4 elements of exercise frequency, intensity, time, and type), aerobic exercise, resistance exercise, and stretching training are in the exercise type. The exercise frequency was 30‐50 minutes each time, 3‐5 times per week, and 90‐150 minutes each week. The oncology rehabilitation program consists of 4 parts: warm-up movements, whole-body exercises, stretching exercises, and finishing exercises. Exercise intensity is recommended to choose low-intensity movements during the warm-up and finishing stages, and moderate intensity during full-body activities. The details of the 4 modules of the oncology rehabilitation program are described in Table 1. This program was subsequently integrated into the app by the software manufacturer.

Participants in the intervention group were provided with a QR code to download the app onto their mobile phones (available for both Android and iOS) before discharge. Upon approval by the researchers, participants were wirelessly connected to a chest-worn heart rate monitor, which measured exercise frequency, intensity, duration, and progression and ensured safety. The exercise intensity was performed at 64%‐76% of the age-predicted maximal heart rate (220-age) [12]. When participants first logged into the app, a tutorial guide video was displayed and saved in the tutorial section of their personal account. The exercise prescription was posted to each participant before discharge, teaching them how to follow the video action training. The content of the training was consistent with intervention sessions for exercise rehabilitation (Table 2). Every time when patients finished the rehabilitation prescription, the APP provided the questionnaire scale (Borg Breathing Scale) and the option for the patient to record the discomfort symptoms and feelings, combined with the record of the heart rate during exercise, which was transmitted to the doctor’s platform in time. The doctors adjusted the order and number of movements in the exercise program according to the patient’s information (Figures 1 and 2). In this trial, different patients had varying exercise capacities. Weaker patients could choose to skip movements they were unable to complete, and the system informed the doctor to make appropriate adjustments. Patients initially unable to complete 30‐50 minutes could exercise in smaller increments throughout the day but were encouraged to exercise for 90‐150 minutes per week. The software automatically uploaded the amount of exercise and daily exercise time.

From discharge from the hospital to 3 weeks until follow-up, patients received a paper-format rehabilitation manual in exercise rehabilitation, the content of rehabilitation training was consistent with that of the digital health group, and rehabilitation maneuvers were guided before discharge patiently until patients could master them. Patients were asked to write down their daily exercise content in the exercise diary. After discharge, patients were supervised via the web once a week. If there were any problems with the exercise, participants could contact the researchers by phone and receive instructions until the outpatient follow-up visit. Participants were followed up in the outpatient clinic at the end of the 3 weeks after discharge from the hospital.

Data were collected between January 2022 and October 2023. After signing the paper consent form, all participants completed the demographic and baseline physical outcomes before randomization (T0). Participants in both groups also completed the assessment at 3 weeks (T1) when they followed up. If participants did not return to the hospital, researchers contacted them by phone to collect related information.

Data were analyzed using IBM SPSS (version 26.0; IBM Corp) and Stata BE (version 17; Stata Corp). Primary and secondary outcomes were analyzed as follows: using the Kolmogorov-Smirnov test to determine whether continuous data follows a normal distribution. Count data were presented as frequencies and percentages (%). Continuous data following normal distribution were expressed as mean and SD, while nonnormal distributions were expressed as median (quartile spacing). Because the outcome variables were measured twice before and after intervention, a difference in differences approach (linear regression) was used to control for baseline levels. The independent samples t test was used for comparison between groups of normally distributed continuous data. The Mann-Whitney U test was used for comparison of nonnormally distributed continuous data. Count data were expressed as the number of cases (rate) and tested using the chi-square test or Fisher exact test. Mean differences in the change from baseline to week 3 were also calculated. If subsequent multi-factor analysis is required, multiple linear regression or logistic regression models are selected according to continuous-type or subtype dependent variables. Multiple interpolations were used for the analysis of missing data. P<.05 was considered significant.

The study used mixed-effects linear regression models to examine the preliminary impact of the intervention while adjusting baseline characteristics. The model incorporated fixed effects for age, sex, education levels, hypertension status, drinking status, and time. To account for the hierarchical structure of the data (patients nested within 2 hospitals), the model included random intercept and slope at the hospital level. At the patient level, random slope and intercept for a time were included to allow variations between individual patients in terms of initial status and trajectories over time on the outcome variables. The model is estimated using maximum likelihood estimation [23,24].

This study was approved by the institutional review board of Huashan Hospital, Fudan University (approval number 2021-794). The research procedures were conducted in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and the World Medical Association Declaration of Helsinki. All participants provided written informed consent prior to their participation in the study. To ensure privacy and confidentiality, all collected data were anonymized and securely stored. Access to the study application was protected by individual private usernames and passwords assigned to each participant. No financial compensation was provided to participants for their involvement in the study, and all study materials were provided free of charge. The study was registered at ClinicalTrials.gov (Identifier: ChiCTR2100052911).

Of the 254 patients screened, 99 patients did not meet the inclusion criteria. Specifically, 72 patients were diagnosed with benign liver tumors, 17 patients were expected to undergo hepatectomy involving fewer than 2 liver segments, and 10 patients were unable to operate a mobile phone. Additionally, 28 patients declined to participate without specifying reasons, 6 patients were not enrolled due to lack of family support, 2 patients refused because they were uncertain about the trial’s usefulness, 12 patients lacked interest, and 7 patients declined for other unspecified reasons.

Of the 100 patients who met the criteria, 50 patients were randomly assigned to the digital health group (n=45, 90% male patients; the mean age was 54.6, SD 10.6 years) and 50 to the control group (n=40, 80%) male patients; the mean age was 56.4, SD 9.7 years). At follow-up, 5 patients in the digital health group were lost: 1 patient lost contact, 1 patient cited time constraints, 1 patient withdrew voluntarily, and 2 patients withdrew due to an allergic reaction to the heart rate monitor. In the control group, 4 patients were lost to follow-up: 2 patients lost contacts, 1 patient died, and 1 patient withdrew voluntarily.

A total of 91 participants completed all assessments, with a dropout rate of 10% (5/50) in the digital health group and 8% (4/50) in the control group. No significant differences were detected between those who were lost to follow-up and those who completed the assessments, as none of the P values reached statistical significance. For more details, refer to Figure 3 for the participant timeline.

Participants’ ages ranged from 35 to 75 (average 55.5, SD 10.2) years. No significant differences were detected in the baseline characteristics between the digital health and control groups (P>.05). The majority (69/100, 69%) of the patients were diagnosed with hepatocellular carcinoma. No significant differences were detected in the surgery characteristics between the digital health and control groups (P>.05). Details are listed in Table 2.

The 6MWT demonstrated a significant improvement in the intervention group compared with the control group. At baseline, the mean 6MWT distance was 515.3 (SD 127.6) m in the intervention group and 520.0 (SD 73.7) m in the control group. After 3 weeks, the mean 6MWT distance in the intervention group decreased to 501.0 (SD 114.1) m, whereas the control group showed a more pronounced decline to 464.0 (SD 100.7) m. The mean difference in the change from baseline between the groups was 70.21 (95% CI 0.730-82.869) m (P=.05), indicating a statistically significant improvement in the intervention group over the control group.

The chi-square test was used to compare the exercise compliance between the 2 groups, and the results showed that compared with the control group, the exercise compliance of the experimental group was significantly higher than that of the control group at 3 weeks after discharge (χ²₂=15.87, P<.001), and the difference was statistically significant, as shown in Table 4.

In recent years, rapid advancements in the treatment of PLC have significantly improved overall survival rates for patients [25]. Rehabilitation, increasingly recognized as a vital component of postoperative care, is delivered through a multidisciplinary team. The benefits of exercise are well-documented by authoritative bodies [26]. However, exercise rehabilitation remains a valuable yet underused treatment strategy for patients with PLC. In this trial, the digital health rehabilitation program was rigorously validated by a multidisciplinary team and specifically designed for patients with PLC posthepatectomy in China. This pilot study aimed to compare the efficacy of a digital health intervention with a conventional rehabilitation manual on physical fitness and symptoms in patients with PLC following hepatectomy over a 3-week intervention period. The findings demonstrated a statistically significant improvement in cardiopulmonary fitness (as measured by the 6-minute walk distance) in the digital health group compared with the control group. However, no statistically significant effects were observed for secondary outcomes. Importantly, exercise adherence was significantly higher in the digital health group.

The findings of this randomized controlled trial conducted in China suggested that novel digital intervention in exercise rehabilitation could improve exercise capacity and physical fitness among patients with PLC after hepatectomy. The intervention has the potential to increase access to rehabilitation treatment and reduce adverse symptoms in patients in the postsurgical recovery period.