From March 2022 to April 2022, we recruited participants with T2D from a community hospital in Yangzhou City.

The inclusion criteria were as follows: (1) age 65 years and older; (2) type 2 diabetes mellitus: diabetes was diagnosed based on self-reports with verification (current treatment, medical records, confirmation from health care providers, fasting plasma glucose of ≥126 mg/dL, or symptoms of hyperglycemia with a plasma glucose level of ≥200 mg/dL); (3) informed consent and voluntary participation in the study; and (4) capability for self-care, no communication barriers, and no cognitive impairment.

The exclusion criteria were as follows: (1) serious physical illnesses (eg, malignant tumors; severe cardiac, hepatic, or renal insufficiency; and cerebrovascular accidents), (2) serious diabetes mellitus–related comorbidities, and (3) current participation in other clinical trials or plans to participate in other clinical trials.

Researchers screened eligible participants through electronic health records within the community hospital health care system and contacted them by phone. Participants who agreed to participate in the study were invited to the community hospital, where researchers met them face to face to discuss the details of the intervention program again, following informed consent. Laboratory testing and questionnaire collection were completed for eligible participants. A total of 198 participants were recruited. At 3 and 6 months of intervention, nurses reminded all participants to visit the community hospital for laboratory examinations and to complete the corresponding questionnaire collection.

Using Microsoft Excel to generate random numbers, eligible study participants were randomly divided into an intervention group and a control group, each comprising 99 individuals, thus maintaining a 1:1 ratio. Recruitment personnel, data collectors, and data analysts were kept unaware of the intervention allocation. Recruitment, data collection, and data analysis were conducted by the respective research assistants, researchers, and statisticians. The physician conducting the lab tests was also unaware of the group assignments. Due to the nature of the intervention, the health care provider delivering the intervention was not blinded.

Upon their initial inclusion in the study, all participants received a diabetes self-management handbook based on the HBM. The multidisciplinary team, consisting of physicians, nurses, and pharmacists, had undergone specialized training in diabetes management intervention based on the HBM, equipping them with expertise in diabetes management and health coaching. Each group of 25 intervention participants was assigned to a team that provided face-to-face group education sessions every 2 weeks for 3 consecutive months (40‐50 min each). These sessions took place in the training room of the community hospital. The intervention group was divided into 4 groups, each comprising approximately 25 individuals. Each group attended educational sessions scheduled for the same week. The educational content was structured around the dimensions of the HBM and covered topics, such as the definition, symptoms, and risk factors of T2D (perceived susceptibility); benefits of using smart devices for diabetes management (perceived benefits); impact of T2D on quality of life and physical health (perceived severity); diabetes self-management through smart devices (cues to action); discussion of the patient’s experience in self-management and use of smart devices, providing solutions to reduce and remove barriers (perceived barriers); and mobilizing family support, showcasing role models, and increasing self-efficacy (self-efficacy). Each session featured a lecture and discussion on the week’s topic, incorporating photos, videos, presentations, and other relevant materials.

Participants’ baseline data were collected using a researcher-designed individual health information form, which included age, sex, living arrangements, educational level, marital status, medical insurance coverage, employment status, monthly household income, duration of diabetes, treatment methods, complications, comorbidities, and other demographic information.

This study used HbA_1c values as the primary outcome indicator. We set the reference values based on previous literature [25]. The mean HbA_1c in the treatment group was 9% (SD 1.9%). In the control group, the mean HbA_1c was 10% (SD 2.7%). We set a 2-sided α of 0.05 with 80% power. Using G*Power 3.1 software (Heinrich-Heine-Universität Düsseldorf), the sample size for the treatment group (N1) was calculated to be 87 cases, and for the control group (N2), it was 87 cases. Accounting for a potential dropout and refusal rate of 10%, we determined that a minimum of 96 participants in each of the intervention and control groups was needed, totaling a minimum of 192 participants.

Statistical analysis in this study was conducted using SPSS 27.0 (IBM Corp), R (version 4.0.3, R Core Team), and GraphPad Prism (version 9.4.1). A P value <.05 was considered statistically significant.

Descriptive statistics were used to describe the demographic distribution and scores for the SDSCA, SED, and HBM. Variables with normal distribution were expressed as mean (SD), categorical variables were expressed as frequencies (percentages), and non-normally distributed variables were expressed as medians (quartiles). Continuous values were compared to baseline data using the independent samples t test or Mann-Whitney U test.

The study used an intention-to-treat analysis to analyze the data. With no pattern in the missing data, the 10% missing values were assumed to be missing at random. The predicted values of the missing data were estimated using Markov Chain Monte Carlo multiple imputation to obtain a complete dataset. For the primary outcome (6-month change in HbA_1c), linear mixed-effects models were used with HbA_1c as the dependent variable. This model included a random intercept with an unstructured covariance matrix, indicator variables for time and study group, and group-by-time interaction terms. The same approach was used for the secondary outcomes.

The study protocol adhered to the Declaration of Helsinki and received approval from the Ethics Committee of the School of Nursing and Public Health at Yangzhou University (approval number YZUHL2022001). The trial has been developed and reported in agreement with the CONSORT (Consolidated Standards for Reporting Clinical Trials) checklist (Checklist 1). Written informed consent was obtained from all participants.

To ensure privacy and security during the use of the device, each participant must log in to the device with a unique account and password. After data collection, it is anonymized and stored in the cloud server, strictly following advanced data protection regulations. At the same time, the system will perform encrypted backup operations as needed to ensure that data and programs are protected.

A total of 219 potential participants were assessed for eligibility in this study. Out of the 210 who were eligible, 12 were excluded due to scheduling conflicts and a lack of interest. Consequently, 198 individuals agreed to participate and were randomly assigned to either the intervention group (n=99) or the control group (n=99) after providing informed consent. A total of 198 out of 191 (96.5%) participants completed the follow-up assessments (Figure 1). The ages of the 198 participants ranged from 65 to 87 years. The duration of diabetes varied from 1 to 41 years, with a mean HbA_1c level of 7.05% (SD 2.41%, range 6.20%-8.25%). In addition, 76.8% (198/152) of the participants were taking oral hypoglycemic drugs to control their blood glucose, 22.7% (198/45) had complications, and 29.3% (198/58) had 2 or more comorbidities. There were no significant differences in baseline characteristics between the 2 groups (P>.05), as shown in Table 1.

Participants’ HbA_1c levels decreased after 3 months of intervention (P=.02), with a more significant improvement noted at 6 months (P<.001). In the intervention group, participants’ HbA_1c levels showed better control during follow-up (P<.001). The decrease in HbA_1c levels at 6 months in the control group was not statistically significant (P=.06). The interaction between group and time was not statistically significant (P=.51), as shown in Table 2. In the intervention group, HbA_1c levels decreased from a baseline of 7.87% to 7.28% at 3 months and further to 6.88% at 6 months. In the control group, HbA_1c levels decreased from a baseline of 7.72% to 7.30% at 6 months. These changes are illustrated in Figure S1 in Multimedia Appendix 1.

At 3 months of intervention, mean SDSCA scores increased from baseline levels in both the intervention (17.13, SD 24.50 vs 19.59, SD 21.54) and control groups (16.97, SD 22.52 vs 18.43, SD 17.58). At 6 months, mean SDSCA scores were higher than baseline levels in the intervention group (17.13, SD 24.50 vs 22.93, SD 14.59; P<.001), whereas mean SDSCA scores were lower than baseline levels in the control group (16.97, SD 22.52 vs 16.93, SD 22.49). The difference in scores between the 2 groups was statistically significant (P<.001).

Mean SED scores also improved from baseline in both groups at 3 months: the intervention (25.38, SD 25.38 vs 29.51, SD 28.25) and control groups (26.18, SD 11.16 vs 26.44, SD 7.55). At 6 months, mean SED scores were higher than baseline levels in the intervention group (25.38, SD 25.38 vs 34.99, SD 17.95; P<.001), while mean SED scores were slightly higher than baseline levels in the control group (26.18, SD 11.16 vs 26.24, SD 8.75). There was a statistically significant difference between the 2 groups (P<.001) at both 3 and 6 months.

At 3 months, HBM scores were higher than baseline in both the intervention group (69.91, SD 56.44 vs 77.89, SD 29.95) and the control group (1.55, SD 22.12 vs 73.35, SD 19.73). At 6 months, HBM scores were significantly higher than baseline levels in both the intervention group (89.65, SD 18.44) and the control group (77.04, SD 23.05), and the intervention group showed a greater improvement in HBM scores. There was a statistically significant difference between the 2 groups (P<.001) at both 3 and 6 months, as detailed in Table 3 and Figure S2 in Multimedia Appendix 1.

No serious adverse clinical events were reported from enrollment to the 6-month follow-up. However, some operational issues with the devices, such as Bluetooth pairing problems between blood glucose meters and smart devices, were reported as inconvenient by participants. Indeed, it has been demonstrated that such issues can lead to lower satisfaction and reduced use of remote technology [26]. Before the start of the study, all participants were informed about the potential for these problems and the availability of telephone assistance.

In this study, we tested a home telemedicine intervention and health education program based on the HBM among community-recruited older patients with T2D. Screening, recruitment, and participation outcomes indicated that, although the number of participants was small, the vast majority were willing to participate and complete the program. While adherence was high in both groups, adherence to self-management varied significantly. Overall, this program has shown potential in helping older patients with T2D achieve safe and effective home diabetes self-management under the supervision and guidance of health care providers.

Longitudinal analyses demonstrated significant improvements in HbA_1c levels at 3 and 6 months, as well as significant enhancements in diabetes self-management, self-efficacy, and health belief scores in the intervention group. In contrast, in the control group, only health beliefs showed significant improvement at 6 months. HbA_1c levels were significantly lower following the intelligent blood glucose management in the intervention group compared to the control group. This study underscores the effectiveness of intelligent blood glucose management as a means of delivering behavioral health interventions to older adults with T2D [27]. At 3 months, HbA_1c levels in the intervention group decreased significantly by 0.59% from baseline, compared to 0.13% in the control group; however, this difference was not statistically significant. At 6 months, HbA_1c levels in the intervention group further decreased by 0.40% compared to 0.29% in the control group, indicating sustained improvement. This finding is consistent with research by Sun et al [28], suggesting that older adults may require time to familiarize themselves with telemedicine systems. Furthermore, after training and remote guidance from the health care team, participants were better able to use smart devices independently [29]. In this study, glycemic management improved in both groups, which may be attributed to several factors. First, the Hawthorne effect [30] could explain improvements, as control participants also had frequent contact with researchers and received additional attention, including regular tests of their glycosylated hemoglobin every 3 months. Second, all participants received a diabetes self-management handbook and specialized care from health care professionals, potentially allowing for more intensive glucose management than usual [31]. However, sustained and effective long-term effects are challenging without ongoing telemedicine oversight and support [32]. Therefore, long-term follow-up of older adults with T2D is crucial.

The significant and long-lasting improvement in self-management in the intervention group, compared to the control group, suggests that our intervention may promote improved self-management behaviors, aligning with findings from previous studies [33]. However, one study indicated that the difference between groups using intelligent blood glucose management alone was not significant, possibly due to older adults’ lack of experience with and incomplete knowledge of smart devices [34]. After incorporating group health education based on the HBM, participants in the intervention group significantly improved their self-management skills. This underscores the vital role that face-to-face diabetes health education plays in helping older adults overcome the challenges of using smart devices and enhancing their self-management. A systematic evaluation noted that an essential component of diabetes self-management is the active participation of the patient in their care [8]. One effective method is through health education, which facilitates the learning process of self-management. Ahmad et al [35] found that older adults with T2D may require regular face-to-face counseling to address the challenges they encounter during telemedicine interventions. Providing support and education through a dedicated community-based diabetes health care team can help participants achieve optimal self-management [31]. Therefore, combining telemedicine with face-to-face health education appears to be a promising approach for enhancing diabetes self-management.

In this study, the health belief scores of participants in the intervention group were higher than those in the control group. The longer the intervention period, the more pronounced the differences in health belief scores became. This suggests that telemedicine-based health education is more effective in improving the health beliefs of older participants with T2D compared to traditional health education [36]. Health education makes participants aware of the complications, severity, and harms associated with their condition, as well as the benefits of adopting a healthy lifestyle. This encourages them to take the initiative in managing their health and deepens their understanding of diabetes. In addition, the feedback and reminders provided by telemedicine play a positive role in enhancing participants’ health beliefs [37], which strengthens their health motivation and helps them establish a correct and comprehensive set of health beliefs. The results of this study show that health education based on the HBM within the context of telemedicine has a positive impact, increasing patient self-efficacy, reducing perceived barriers in the intervention group, and facilitating an increase in patient health behaviors.

We also found that the intervention improved self-efficacy in older adults with T2D, similar to the results of other studies [38]. Given that increased health beliefs are associated with higher diabetes self-efficacy [20], participants in the intervention group experienced higher self-efficacy as their level of health beliefs increased. Elevated levels of self-efficacy are often linked with decreased barriers to engaging in health behaviors, which in turn leads to increased motivation to engage in desired activities [39].

First, while studies have shown the short-term impact of telemedicine on glycemic management in older adults with T2D, the long-term validity of these effects remains unassessed. Second, the participants who chose to join this study likely demonstrated some level of readiness for digital health literacy change. Therefore, strategies have yet to be developed for those uninterested individuals who did not participate in the study. Finally, the study relied on structured questionnaires for participants to self-assess their self-efficacy, health beliefs, and diabetes self-management skills, which may have introduced recall bias into the results.

The findings of this study suggest that health education and telemedicine interventions based on the HBM can improve HbA_1c levels and other related outcomes in older adults with T2D. The telemedicine system provided timely access to the health status of older participants with T2D, facilitated communication and feedback between doctors and participants, and enabled the inclusion of health care provider counseling to enhance patient motivation. This, in turn, improved the effectiveness of medical care, leading to better glycemic management in the intervention group. Future research should further evaluate the long-term effects of this approach compared to routine community diabetes care. In addition, it should focus on providing better support for lifestyle changes through enhanced interventions and counseling to improve patient motivation and confidence, ultimately offering more comprehensive diabetes management for participants.