It was an open-label, evaluator-blind, randomized controlled study with 2 parallel groups.

Patients aged ≥18 years and scheduled for CRC surgery were included. The exclusion criteria were patients who (1) had American Society of Anesthesiologists (ASA) grade >III, (2) had preoperative immobility or inability to walk unaided affected by neural, muscular, or skeletal diseases, (3) had participated in other clinical studies in the past 3 months, (4) had planned reoperation within 30 days after the initial operation, and (5) had expected to be transferred to the intensive care unit after their operations.

The Ethics Committee of the First Affiliated Hospital of the Air Force Military Medical University approved this randomized controlled trial (approval no. KY20222299-F-1), and the study protocol was registered with the Chinese Clinical Trial Registry (ChiCTR2300068107) on July 2, 2023. All procedures complied with the Declaration of Helsinki and institutional regulations (The Ethics Committee of the First Affiliated Hospital of the Air Force Military Medical University, Medical Ethics Committee Charter) [24]. Written informed consent was obtained from all participants prior to enrollment, which included detailed explanation of the study and their right to withdraw at any time. Upon registration, each patient was assigned a unique study number, and all subsequent data analyses were performed using a deidentified dataset to ensure participant privacy and confidentiality. No financial compensation was provided for participation.

All patients in the trial were randomly assigned (1:1) to a feedback group or a control group. Both groups were set the same target activity time postoperatively and thus performed activities for 2 hours on the first day after operation (POD1), 4 hours on the second day after operation (POD2), and 6 hours on the third day after operation (POD3), according to the ERAS guidelines and our ERAS pathway for CRC [10]. All patients were fitted with a wireless wearable device that continuously recorded their daily ambulation durations (ePM 12M monitor placed at the bedside or ep pod monitor placed on their wrist or blood pressure pod monitor placed on their upper arm).

In the feedback group, patients received real-time visual feedback of the movement time every day through the display of the ePM. According to the wards’ routine care based on ERAS guidelines, patients were reminded to engage in activities combined with ePM/ep pod feedback to meet predefined daily step goals and also informed about the gap between the active time and the target time.

In the control group, patients’ movement time was set to not appear on the display of the ePM. Patients were not given feedback, and they were only encouraged to carry out activities according to the current care standards based on ERAS guidelines.

If the following incidents occur, activity supervision should be suspended immediately: participants experience pain intensity >4 (from 0 to -10), electrocardiogram abnormalities (signs of arrhythmia or acute ischemia), indoor air oxygen saturation <88%, systolic blood pressure <90 mmHg, symptoms of standing intolerance (ie, dizziness, nausea, and blurred vision), exercise weakness (ie, difficulty standing), etc. In the next scheduled mobilization, if the participant is considered to be in stable physical condition after clinical evaluation, a new mobilization attempt will be conducted. If serious adverse events such as falling out of bed or falling occur, in addition to suspending activity supervision, the competent physician should also be asked to evaluate the patient’s injury and terminate the study if necessary. The researcher should immediately fill out the serious adverse event report and report it to the hospital ethics committee.

Patients were randomized using block random groups (sized blocks of 2 and 4) generated by specialized staff. All random assignments were made and sealed to an envelope by intervention staff using Statistical Analysis Software. The randomization process was performed by an independent anesthesiologist at the end of the surgical procedure, separated from the ward’s nursing staff and outcome assessors to reduce potential bias.

It was impractical to blind the intervention staff in view of the intervening measure, but contrary to the outcome evaluators. To ensure blinding during outcome assessments in the surgical ward, a schedule was established to stagger the intervention and evaluation implementation. Ward staff were informed about the activity monitoring study but were not provided with any specific hypothetical details to prevent potential bias from influencing the study results. Given the intervention characteristics, blinding the patients was also deemed to be unfeasible. To minimize performance bias, patients were informed that we were evaluating 2 mobilization schemes on the understanding that neither scheme was expected to confer any potential advantage.

Preoperative baseline data were collected before an operation as follows: demographic data; preoperative medication and abdominal surgery history, and preoperative radiotherapy and chemotherapy history; vital signs; available preoperative examinations; and preoperative physical activity ability assessed by the International Physical Activity Questionnaire (IPAQ) [25]. Operation-related data collected on the day of surgery included surgical duration, intraoperative bleeding, site, type, and catheter placement status. Postoperative data were collected after each operation as follows: complications according to the Clavien-Dindo classification during hospitalization and 1 week, 2 weeks, and 1 month after the operation; the wearable device ep pod record of the activity time; digital pain Numeric Rating Scale (NRS) scores of rest and exercise; length of hospitalization; time of first exhaust and defecation; and opioids consumption; when discharged from hospital, patients in the intervention group were asked about concerns for feedback on activity time and whether they would supplement activities.

The sample size calculation was based on the primary outcome measured by the CCI 30 days after operation. CCI represents the measurement data that follows a normal distribution, as detailed in the previous trials on postoperative complications [33,34]. The average postoperative CCI of CRC was 10 (SD 5). Assuming that the visual feedback of the movement time could reduce CCI by 25%, the average CCI of the feedback group was set to 7.5 (SD 5) [35]. Based on the above effect size, Power Analysis and Sample Size 2015 Software was used to calculate that a minimum of 172 patients (n=86 per group) would provide 90% power to detect a 25% decrease of CCI and an α of .05 (2-sided test), assuming the feedback group was more favorable compared to the control group [35]. To compensate for a potential 20% dropout rate, 216 patients were required to be included (n=108 per group).

We followed the intention-to-treat principle analyzing all random patients in the baseline analysis. The analysis results of the study were based on the modified intention-to-treat (MITT) set. Statistical analyses were performed using SPSS version 23.0. The 2-sided significance level was set at a P value <.05.

Continuous variables with independent access to mean and SD were initially subjected to the Kolmogorov-Smirnov test to assess their normality. Otherwise, the Shapiro-Wilk test was chosen. The Levene test was used for the analysis of homogeneity of variance. Data with skewed distributions are reported as medians with IQR, while data adhering to a normal distribution are given as means with SD. Categorical variables are expressed as frequencies, percentages, and ratios. For baseline comparisons, normally distributed continuous variables with comparable variance were analyzed using Student t test. In contrast, continuous variables with skewed distributions or uneven variance were analyzed with the Mann-Whitney U test. Categorical variables were compared using either the chi-square or Fisher exact tests.

The impact of the primary outcome CCI within 30 days after an operation was analyzed using the Student t test or the Mann-Whitney U test. Except for the MITT analysis, subjects with low device compliance (wearing ePM/ep pod for less than 8 h from 8 AM to 8 PM) were excluded for sensitivity analysis of the primary outcome. In addition, subgroup analyses (preselected age, ASA, preoperative chemoradiotherapy, surgical site, and preoperative mobility) were carried out. Simultaneously, we examined the interaction between the treatment effect and these 5 subgroup variables. In addition, we statistically described the distribution of postoperative complications at all levels in the 2 groups and performed a chi-square test between the groups. As a post hoc analysis, the Spearman/point-biserial correlation coefficient was used to determine the correlation between CCI and the postoperative movement time. In the preliminary analysis, the linear regression model showed a good fitting effect. Based on this result, we conducted a linear regression analysis to assess the relationship between the movement time of POD3 (for patients discharged on POD3, the activity time on POD2 was used to fill in the data) and CCI, adjusted for potential factors such as age, ASA, sex, preoperative chemoradiotherapy, and BMI, all factors which could influence postoperative recovery.

In analyses of secondary outcomes, the Mann-Whitney U or Student t tests were used to analyze the other measurement data except for repeated measurements. The chi-square test was used to analyze the percentage of patients who achieved the predetermined mobilization goal. Binary logistic regression analysis was used to determine the determinants of patient cohorts to achieve mobilization goals. The regression model considered the following variables: age, sex, ASA grade, surgical site, and the IPAQ [25]. Repeated measurement of the NRS pain score, using a generalized linear mixed model with multinomial distribution, was used for within-group comparisons.

In this study, the activity incentive feedback based on a wearable device (ePM/ep pod) did not reduce the CCI of patients with CRC within postoperative 30 days. The differences between groups in terms of daily activity time, length of postoperative hospital stay, intestinal function recovery, and NRS pain scores were also not significant.

There may be the following reasons for the lack of difference in CCI and daily activity durations between the 2 groups. First, this study was conducted in a surgical ward with relatively deepening ERAS guidelines, where preoperative education and routine activity supervision may have facilitated effective postoperative activities even in the absence of the feedback from wearable devices. Second, all the subjects in the present study received the same activity encouragement by medical staff according to the routine care and were set the same daily activity target; therefore, patients in the control group with high compliance might obtain activity feedback through other means. Third, laparoscopy accounted for more than 95% of CRC surgery in the present study, which was similar to the type of surgery used in another study that yielded negative results regarding the hypothesis that early postoperative activity could reduce complications [15]. Future studies should investigate whether such incentive feedback would be effective for patients experiencing more intense surgical stimulation or possessing less ERAS education. Previous studies using step count feedback for mobility-enhancing showed conflicting

The assessment of early postoperative activity achievement in patients with CRC directly reflects the practical outcomes of the intervention strategies employed in this study. The activity compliance rate of patients with CRC on POD1 was 37.4% (in the usual care group) and 43.8% (in the activity monitoring feedback group), which was much lower than expected, indicating that even with the feedback mechanism, there was only a marginal improvement in patients’ adherence to the activity targets. Based on the current 6 hours recommended by ERAS guidelines, this study set a less active target of 2 hours for POD 0. In addition, the maximum activity time on POD1 was 310 minutes, which was much lower than the 6 hours recommended by ERAS guidelines. The relatively low compliance rates suggest that factors beyond the provision of activity feedback may be at play in influencing patients’ postoperative behavior, such as the intensity and appropriateness of the activity targets set. The insights gained from analyzing the achievement of early postoperative activity not only provide a deeper understanding of the limitations of the feedback intervention but also highlight the need for more tailored and patient-centered approaches to enhance postoperative mobilization. The postoperative activity time was measured by an electronic parameter objectively recorded by a wearable device, and the measurement method may be different from the considerations of previous guidelines. This may be the reason why the postoperative activity time was inconsistent with most previous studies.

Previous studies [37,38] have found patients undergoing major visceral (abdominal) surgery may not be able to achieve the mobilization goals in the early postoperative period. An observational pilot study [39] pointed out that only 50% of patients undergoing major visceral surgery could walk until POD2. The cumulative duration of postoperative activity per day ranged from 15 to 155 minutes, and only 40% of patients were satisfied with activity or physical exercise. There is a discrepancy between ERAS goals and clinical practice. Another study [40] reported that only a few patients (23.5%) achieved the mobilization goal of 6 hours (ERAS guidelines) of activity on the first day after CRC surgery. Ramírez et al [41] set the early activity goal for patients with CRC as sitting for at least 6 hours on the first postoperative day. However, the achievement rate was only 44.6%. In general, the mobilization goals described in most studies were either unclearly defined, imprecisely measured, or inaccurately monitored.

In addition, the early activity goals in ERAS did not differentiate between disease types, and radical abdominal tumor surgery is far more invasive than other simple abdominal organ partial removal. Patients with tumors were often elderly and had comorbidities, which can limit physical fitness and motivation. A study by Reed et al [42] on activity tracking after bariatric surgery found a significant negative correlation between age and the steps on POD1. Our study presented similar findings in that age was an independent factor affecting the achievement of activity on POD1. The older the patients were, the lower the possibility of reaching the activity goal. Furthermore, age negatively affected activity duration on POD2 and POD3. All patients with CRC in the present study were 61 years on average, and 42% had preoperative comorbidities, which may also be the reason for the low level of postoperative activity. Also, regression analysis showed that the activity duration of patients with colon cancer on POD2 was significantly longer than that of patients with rectal cancer, being 26.42 minutes longer. Patients with colon cancer were 2.28 times more likely than those with rectal cancer to meet the activity standard on POD2. It can be seen that the mobility of patients with colon cancer on POD2 was higher than that of patients with rectal cancer. We speculate that the reason for this finding may be that colon cancer basically does not receive radiotherapy, the proportion of preoperative chemotherapy may be lower than for rectal cancer, and the proportion of temporary ileostomy of colon cancer being much lower than that for rectal cancer. Our finding also suggests the target of early activity recommended by current ERAS guidelines (6 h in POD1) may be too high for these patients. A further increase in activity did not increase the benefit when activity levels reached a certain level. Considering our result of the regression analysis, 215 minutes might be considered as the target for activity time on POD3 to decrease CCI in patients with CRC.

In addition, several studies [43-45] have reported that early postoperative mobilization (within 24 h) was associated with fewer postoperative complications, fewer major complications, faster intestinal function recovery, and shorter hospital stays. The regression analysis in this study found that patients with high levels of physical activity before surgery were about 27 minutes more active the next day compared to patients with moderate levels of physical activity before surgery, and about 37 minutes more active than patients with low levels of physical activity before surgery. Therefore, perhaps by exploring the extent of postoperative activity decline caused by different surgical trauma, the approximate postoperative activity amount of the patient can be inferred and then combined with the preoperative activity level of the patient, a personalized postoperative activity plan can be formulated.

This study had some limitations. First, the 216 patients were all admitted to one hospital; therefore, the conclusion still needs to be verified in more hospitals where ERAS guidelines are less practiced. The low proportion of ASA 3 patients in this study may limit our comprehensive assessment of the benefits of patients with high comorbidities from feedback from postoperative activities. Future research may consider including more ASA3 patients to better understand the response of this special population to postoperative activity feedback intervention. Second, since it was not possible to blind the patients in this study, the 2 groups of patients may have influenced each other. Due to practical difficulties such as high patient mobility and limited follow-up time, this study failed to collect information about the views of patients and clinicians with the wearable device and how motivated they felt by wearing this. In the future, the study can be expanded to multiple wards to achieve cluster randomization, which may reduce bias. Future research can also be combined with qualitative items to provide a more comprehensive understanding of patient and clinician experiences and perceptions of wearable devices. Third, as 96% of patients in this study underwent laparoscopic surgery, the results may not be generalized to open surgery or other types of surgery. Although our inclusion criteria were broad, we excluded patients who had difficulty with preoperative activity or were in a poor physical condition. These patients may benefit from the incentive of activity feedback. A final limitation was there may be a difference between self-reported preoperative mobility assessed by IPAQ, which was not designed for hospital patients, but for assessing physical activity in daily life, and actual quantitative mobility.

In the context of ERAS, this randomized controlled trial did not find evidence that activity stimulation based on feedback from ePM/ep pod wearable devices could reduce 30-day postoperative CCI in patients who underwent CRC surgery. However, ePM/ep pod could accurately record daily activity duration, which may be negatively correlated with CCI on POD3.