This study was approved by The Joint Chinese University of Hong Kong – New Territories East Cluster Clinical Research Ethics Committee (2022.264), Hong Kong SAR. All participants were informed of the voluntary and anonymous nature of the study. Written consent was obtained from all participants. Participants were given a document to explain the study’s purpose and procedures. All data collected were anonymized to ensure that there was no personally identifiable information, encrypted, and stored electronically.

This randomized controlled trial was conducted over a period of 1 year in 2024, involving 165 participants, with a follow-up time of 8 months. A total of 83 participants were assigned to the intervention group and 82 participants to the control group using a computer-generated randomization process. Patients were categorized into 2 groups: those who had adopted the health management module and those who had not. The intervention group received the eHealth app, which was installed on their phones with detailed instructions from our team members at baseline. Reminders were sent weekly to ensure the use of the app among participants in the intervention group. During the designated reporting period, 3 to 6 phone calls were conducted to remind participants for blood tests. The enrollment process adhered to the CONSORT (Consolidated Standards of Reporting Trials) guidelines [16].

This trial was performed in a designated general outpatient clinic with a high volume of patients with diabetes attending for follow-up. Patients’ eHRSS registration status was documented in the clinical management system. Eligible participants included (1) adult patients aged 18 years or older, (2) patients with physician-diagnosed diabetes who had undergone diabetes mellitus complication screening before their installation of the eHRSS health management module, (3) regular visitors of the study clinic, and (4) patients diagnosed with type 2 diabetes. Exclusion criteria were the presence of medical conditions that impaired their ability to understand and participate in the study and previous use of the module.

Eligible participants were recruited in the clinic and underwent a subsequent diabetes mellitus complication screening to assess the potential complications of diabetes. Social demographic information was collected, including information on the patients’ lifestyles, such as dietary habits, physical activity, cigarette smoking, and alcohol consumption. Their clinical parameters were also assessed, including BMI, blood pressure levels, HbA1c levels, urinalysis (albumin/creatinine ratio), and the presence of diabetes complications.

Data were collected at baseline and at follow-up visits at 4 and 8 months. Recruited participants underwent assessments for various diabetes-related health parameters, including blood pressure, HbA1c, renal function tests, and urine albumin/creatinine ratio. Additionally, lifestyle data including dietary consumption, physical activity, cigarette smoking, and alcohol consumption were collected. Dietary consumption and physical activity were measured using a 4-point Likert scale from 1 (never) to 4 (every time), with higher scores representing healthier lifestyles. Information on diabetes-related complications (macrovascular complications such as coronary artery disease, peripheral arterial disease, and stroke and microvascular complications such as diabetic neuropathic, nephropathy, and retinopathy) and prescribed medications was also collected. App usage at 4 and 8 months were measured using a 6-point Likert scale from 0 (never) to 5 (every time), with higher scores representing higher levels of use.

Block randomization was adopted in this study to ensure balanced group sizes throughout the trial using blocks of size 4. This approach allowed random assignment of participants within each block, helping to maintain equilibrium between treatment and control groups for potential confounding variables.

Baseline characteristics were compared between the intervention and control groups. Student t tests for continuous variables and chi-square tests for categorical variables were performed. Each clinical parameter was analyzed between the intervention and control group to examine the potential impact of the use of the eHealth chronic disease management module on various health outcomes. All analyses were performed using SPSS version 26.0 (IBM Corporation). P values <.05 were considered statistically significant.

A total of 165 patients were recruited, with 161 completing at least 1 follow-up (Figure 1). During the study, 4 participants were lost to follow-up, resulting in a default rate of 2.4%. In the end, 79 participants in control group and 82 participants in intervention group were included in the analysis. The average age of the participants was 66.58 (SD 9.75) years in the control group and 66.49 (SD 8.45) years in the intervention group, with no difference between groups. Details of patient sociodemographics are shown in Table 1.

At the 4-month follow-up visit, the intervention group exhibited better diabetes control, as indicated by a significant difference in HbA1c levels. The intervention group had a mean HbA1c of 6.76% (SD 0.64%), compared to 7.09% (SD 0.82%) in the control group (P=.007). However, for other outcome measures, such as renal function tests, estimated glomerular filtration rate, and urine albumin/creatinine ratio, no significant differences were observed between the 2 groups. Additionally, subsequent follow-ups did not reveal any significant differences (Table 2).

In terms of the proportion of participants who achieved optimal glycemic control, the results indicated a significant difference in HbA1c levels at the 4-month follow-up between the intervention and control groups. Specifically, 75.95% (60/83) of the intervention group achieved an optimal result, compared to 58.75% (47/82) in the control group (P=.02). No significant differences were found for the other outcome parameters, such as estimated glomerular filtration rate and urine albumin/creatinine ratio, and no further differences were observed in the subsequent follow-up (Table 3).

App usage also declined over time, with the intervention group showing a mean usage of 1.88 points (SD 0.81 points) at the 4-month follow-up compared to 1.39 points (SD 0.72 points) at the 8-month follow-up (P<.001), suggesting a decrease in engagement over time (Table 4).

This study found a significant improvement in HbA1c levels among participants in the intervention group at the 4-month follow-up and no significant difference between participants in the intervention group and control group at the 8-month follow-up. This finding suggests that optimal glycemic control could be achieved for short period of time with app use. However, the compliance of users with the module could be a barrier to the long-term success of app use on glycemic control and other health outcomes.

This study showed short-term improvement of glycemic control, aligning with some previous studies. Tobacman et al [17] conducted a prospective cohort study and revealed a significant improvement in care service for the intervention group that had documentation of their own health records. Furthermore, better clinical outcomes were also observed from another randomized controlled trial study conducted by Bourgeois et al [18]. However, the phenomenon was not observed in month 8; this drop in the effect of the module could explain the decrease in medication adherence. Arshed et al [19] studied 23 studies on cardiovascular disease between 2000 and 2021 and revealed 26% of studies showed that mobile health interventions failed to show continuous impact on medication adherence. Pouls et al [20] conducted a systematic review and revealed that 17 out of the 29 randomized controlled trials included showed positive effects in medication adherence through use of SMS or interactive voice response, mobile apps, and phone calls as feedback, strategies to teaching, and facilitating communication or decision-making. The future of the management module could focus on developing features that facilitate increased adherence by patients, eventually improving clinical outcomes.

These findings were compatible with results from previous studies [21]. Many studies have pointed out that the use of electronic health programs can help patients improve their health behaviors and achieve better disease control, which may be attributed to the ability of the eHealth chronic disease management module to enhance self-control and self-efficacy, such as illness recognition and the stimulation of self-reflection through continuous monitoring of health data [22].

However, the lack of significant improvements in other clinical parameters and the diminished effect at the 8-month follow-up point to challenges in sustaining long-term engagement with the app. Our study also found that user engagement with the app declines over time. Previous studies also pointed out that users’ willingness to use the eHealth management module is often highest in the beginning stage, and the attrition rate may increase over time, potentially due to waning motivation, declining interest, and nonusage attrition [23,24]. This pattern may explain why the initial benefits observed at 4 months were not sustained at 8 months. Therefore, the sustainability of the apps’ long-term impact and continuous use need to be taken into consideration [25].

Users of the eHealth app, particularly the eHealth management module, demonstrated better blood glucose control compared to nonusers. Older users showed better blood glucose control, despite the majority of app users being younger. Numerous factors, such as sociodemographic characteristics, medical conditions, duration of diabetes diagnosis, and medication, could have influenced the overall HbA1c outcomes. This study revealed no significant difference between the 2 groups in term of BMI, blood pressure, HbA1c, and urine albumin/creatinine ratio in patients with diabetes over a 1-year period. Developers of eHealth apps should focus on increasing the adherence of users to eHealth app usage and improving app features that may help patients to manage their weight, blood pressure, and chronic disease such as diabetes.

This study demonstrated the positive health outcomes in glycemic control achieved by using the eHealth chronic disease management module. However, this study has some limitations. First, despite reminders being sent routinely to participants in intervention group, participants may not have fully adhered to the modules, leading to biased results. Second, the statistically significant difference in HbA1c levels between the intervention and control groups may not represent a clinical difference. Further follow-up and investigation in real clinical settings are needed to confirm the finding. Third, based on the outcomes of this study, longer follow-up periods—such as 12 or 24 months—are suggested in order to generate a comprehensive result. Fourth, individuals assessed for eligibility were not included in the study, as recruitment occurred in the clinic without maintaining formal records of those assessed during the eligibility screening process.

This randomized controlled trial demonstrated the potential short-term benefits of an eHealth chronic disease management module for improving diabetes control. However, the decline in app usage and lack of sustained improvement in other clinical parameters at the 8-month follow-up highlight the need for strategies to maintain user engagement over time. This study identified positive clinical outcomes from the module and the need for improvement in the compliance of patients. Future iterations of the eHealth program should incorporate features that enhance long-term user motivation, such as regular feedback, ongoing support, and educational interventions, particularly targeting older adults or those with limited technology proficiency. Ongoing evaluation and adaptation of the eHealth module will be critical to maximize its long-term impact on diabetes management.