This is an open-label, cluster-randomized, controlled study that compared digital interventions for hypertension management to a control group. The study was conducted at 8 rural health care sites in China. We selected trial sites based on the following three criteria: a high prevalence of poorly controlled hypertension, geographic remoteness and limited access to health care, and the village physician’s willingness to participate. Patients were recruited between June and July 2024, and the follow-up began in December 2024 and ended in January 2025.

The inclusion and exclusion criteria are presented in Textbox 1.

Researchers screened eligible participants by reviewing medical records from 8 rural health care sites involved in the trial. Randomization (1:1) was performed using SPSS software (version 22.0; IBM Corp.) to randomly assign 8 villages to the digital intervention group or the control group. The randomization process included (1) generating a unique random number for each participating village; (2) sorting these random numbers in ascending order; and (3) allocating the first 4 villages to the intervention group, whereas the subsequent 4 villages were assigned as the control group based on the sorted random number assignment.

The village physicians contacted the identified eligible individuals via phone calls or WeChat to invite them to the local health care sites to learn about the study. Following this, the eligible people received printed education booklets containing information on the definition, causes, and basic treatment guidelines for hypertension (including lifestyle changes and medications), which were uniformly explained to them by the researchers. After resting for at least 15 minutes, the researchers measured their BP using an electronic BP monitor (Omron, HEM-7121). Three BP readings were taken at intervals of 1 minute. Those who met the criteria and wished to participate were asked to provide informed consent (Multimedia Appendix 1), after which their baseline data were collected. Participants in the digital intervention group could directly enter their information via the WeChat mini-program. The collected baseline data included demographic characteristics, physical examination findings, medical and personal history, laboratory test results, medication usage, and psychological assessments (including the Generalized Anxiety Disorder 7-item scale (GAD-7) Anxiety Screening Scale, Patient Health Questionnaire-9 item (PHQ-9) Depression Screening Scale, and Pittsburgh Sleep Quality Index).

The targets of BP control were set according to the latest hypertension guidelines [13,14] (participants aged <80 years: <140/90 mm Hg; participants aged ≥80 years: <150/90 mm Hg; and those with diabetes, chronic kidney disease, coronary heart disease, or history of stroke, or transient ischemic attack: <130/80 mm Hg). The primary end point was the hypertension control rate at 24 weeks. The secondary end point was the change in mean SBP and DBP from baseline to 24 weeks.

This study aimed to investigate the effects of digital interventions on hypertension control. Two weeks after enrollment, participants were contacted by telephone to assess their ability to properly use the BP monitor. BP readings were not recorded at this time. Previous literature indicated that achieving BP control within 6 months was associated with a reduced risk of adverse outcomes compared to later control [12]. This time frame also allowed for sufficient adjustments in BP management. Therefore, follow-up visits were scheduled at 24 weeks post randomization and were conducted by researchers and village physicians. The main methods of follow-up included clinic visits or telephone calls, and participants’ BP was recorded.

The sample size was calculated based on the following assumptions: based on previous studies, the hypertension control rate in the control group was set at 10% [14], whereas the estimated control rate for the digital intervention group was 40%. Referring to prior literature, assuming an intracluster correlation coefficient of 0.01 for the primary outcome [15,16], with a test power of 90% and a 5% type 1 error probability, a sample size of 42 patients per group was required to detect a difference in hypertension control rates between the digital intervention and control groups. Considering a potential 10% dropout rate, each group should initially include 47 participants (12 participants in each cluster).

All analyses were performed based on the full analysis set population. Patient characteristics at baseline were described using mean (SD) or median (quartiles) for continuous variables and number (proportion in percentage) for categorical variables. Normally distributed continuous variables were compared using Student t test with unequal variances, and continuous nonparametric variables were compared using Kruskal-Wallis test. Categorical variables were compared using the chi-square test. The primary end point, the hypertension control rate at 24 weeks, was compared between the digital intervention group and the control group using the chi-square test. To comprehensively assess the impact of digital interventions on the hypertension control rate, logistic regression analysis was conducted. The secondary end point was analyzed using analysis of covariance, adjusted for baseline BP, to avoid confounding the 24-week BP change. In addition, a 2-factor, 2-level repeated measures analysis was used to assess the impact of the digital interventions and the pre- and post-measurement time points on the change in mean SBP and DBP from baseline to 24 weeks.

Statistical analyses were performed using SPSS software (version 22.0; IBM Corp.) and R software (version 4.4.0; R Core Team), with the ggplot2 package (version 3.5.1; Posit, PBC) for creating figures. P values <.05 were considered statistically significant.

This study was reviewed and approved by the Medical Ethics Committee of the Sichuan Academy of Medical Sciences and Sichuan Provincial People’s Hospital, with the ethics review approval number: Ethics Review (Research) No. 47, 2024 (Multimedia Appendix 4). The trial was conducted in accordance with the Declaration of Helsinki and all relevant laws and guidelines in China. In addition, this study was reported in compliance with the CONSORT-EHEALTH (Consolidated Standards for Reporting Clinical Trials of Electronic and Mobile Health Applications and Online Telehealth) checklist (Checklist 1). All participants provided written informed consent before enrollment in the study. They retained the right to refuse participation or withdraw at any time during the research. No public reports of study results will disclose participants’ personal identities. As stipulated by the study protocol, and to compensate participants for their time and effort, each individual received an electronic BP monitor and a printed hypertension education booklet, provided free of charge. These items were intended for use throughout the study and to encourage ongoing health self-management thereafter. The combined approximate retail value of these provisions is US $35. This compensation plan, including the nature and value of the provided items, was explicitly reviewed and approved by the Medical Ethics Committee.

Between June and July 2024, a total of 95 participants were enrolled and allocated to the digital intervention group (n=48, 51%) or the control group (n=47, 49%). Of these, 87 (92%) of 95 participants were followed for BP monitoring over 24 weeks (Figure 2). On average, study participants were aged 63.8 (SD 9.7) years, with 48% (n=42) being female participants, and the time since hypertension diagnosis averaged around 7.6 (SD 6.8) years. More than half of the participants had a primary school education and low incomes. In the digital intervention group, 91% (40/44) had a self-reported history of hypertension, and 77% (34/44) participants took antihypertensive medications, whereas in the control group, 81% (35/43) had a self-reported history of hypertension, and 77% (33/43) participants took medications. More than half of the participants were taking CCB (n=26, 59% in the digital intervention group and n=25, 58% in the control group); more details are presented in Table 1 and Multimedia Appendix 5.

With a follow-up of 24 weeks, a higher percentage of participants (25/44, 57%) in the digital intervention group achieved controlled BP compared to the control group (21/43, 49%), although this difference was not statistically significant (P=.60; Table 2). Logistic regression analysis showed that the digital intervention did not significantly increase the hypertension control rate (odds ratio 0.73, 95% CI 0.31‐1.69; P=.46). In the digital intervention group, SBP decreased from 158.0 (SD 18.4) mm Hg at baseline to 137.5 (SD 13.0) mm Hg at 24 weeks, and DBP decreased from 93.8 (SD 10.3) mm Hg to 85.3 (SD 11.6) mm Hg. In the control group during that same period, SBP decreased from 161.1 (SD 18.2) mm Hg to 139.6 (SD 13.2) mm Hg, and DBP decreased from 99.2 (SD 9.2) mm Hg to 83.4 (SD 12.1) mm Hg.

The average reductions in SBP and DBP from baseline to 24 weeks were both significantly greater in the digital intervention group (SBP −21.8 mm Hg, 95% CI −25.7 to −17.9; and DBP −10.5 mm Hg, 95% CI −14.0 to −6.9) and the control group (SBP −20.2 mm Hg, 95% CI −24.1 to −16.2; and DBP −13.8 mm Hg, 95% CI −17.4 to −10.2). After adjusting for baseline SBP or DBP, the mean change from baseline to 24 weeks was comparable between the digital intervention and control groups for both SBP (between-group difference −1.6, 95% CI −7.2 to 3.9; P=.56) and DBP (between-group difference 3.3, 95% CI −1.8 to 8.5; P=.21; Figure 3).

Repeated measures analysis revealed that there was a substantial difference in SBP from baseline to 24 weeks in the 2 groups (between-time difference −21.0, 95% CI −25.3 to −16.7; P<.001), but without consideration of measurement time, no significant difference in SBP was observed between the digital intervention and control groups (between-group difference −2.6, 95% CI −7.8 to 2.6; P=.33). Due to the interaction between measurement time and group (P=.01), the amplitude of the change in DBP from baseline to 24 weeks differed between the 2 groups (the digital intervention group: between-time difference −8.5, 95% CI −12.6 to −4.4; P<.001; and the control group: between-time difference −15.9, 95% CI −20.0 to −11.7; P<.001). Additionally, baseline DBP was notably different between the 2 groups (between-group difference −5.4, 95% CI −9.6 to −1.3; P=.01), but no significant difference in DBP at 24 weeks was found (between-group difference 1.9, 95% CI −3.1 to 7.0; P=.45; Figure S1 in Multimedia Appendix 5).

Three adverse events were reported across the two groups: 1 participant in the digital intervention group was hospitalized for uncontrolled hypertension, whereas 2 participants in the control group were hospitalized—one for uncontrolled hypertension and the other for acute cervical trauma. Hypotension did not occur in either group. None of the adverse events were considered to be related to the digital intervention system.

The main finding of this study was that while the digital intervention did not show statistically significant results, it still played a role in improving BP control. Moreover, providing convenient BP measurements and health education to patients in rural areas substantially improved the hypertension control rate, achieving 57% in the digital intervention group versus 49% in the control group at 24 weeks. Notably, both the digital intervention and control groups exhibited considerable BP reduction. Specifically, the digital intervention group showed a decrease of 21.8 mmHg in SBP and 10.5 mmHg in DBP, whereas the control group experienced a reduction of 20.2 mmHg in SBP and 13.8 mmHg in DBP, respectively.

In China, the control rate of hypertension in rural areas is low, with previous studies reporting <10% [2]. The low hypertension control rate may partially be attributed to a lack of awareness among patients about high BP. Our study findings suggested that self-monitoring of BP and health education could motivate patients to take an active role in managing their hypertension. As demonstrated in meta-analyses, self-monitoring of BP, either alone or in conjunction with co-interventions, has been shown to reduce BP in patients with hypertension [17-19]. A prospective, randomized controlled study suggested that self-BP measurement as an adjunct to office-based BP monitoring improved treatment adherence [20]. While some studies have reported inconsistent results regarding BP reduction and adherence improvement [21,22], several trials still support the benefits of self-measured BP in improving hypertension awareness and control, and it is more effective in improving BP control when combined with education and counseling [23,24]. As home BP monitoring requires active patient participation, it may positively impact patients’ perception of their hypertension, thereby promoting adherence to lifestyle modifications and antihypertensive therapy. Notably, the use of home BP monitoring is recommended by hypertension management guidelines [13,25]. This study highlights the importance of providing free and accessible BP self-monitoring tools for all patients in need. Furthermore, a cluster-randomized trial in China compared nonphysician community health provider–led BP intervention with usual care, indicating that nonphysician community health providers received appropriate training and patients received affordable or free antihypertensive medications, remarkably improved the hypertension control in low-resource settings [26,27]. Therefore, a multifaceted strategy should be promoted in rural areas to reduce the morbidity and mortality associated with hypertension.

Digital therapeutics represents a promising approach to support the management of hypertension [28]. However, previous studies have demonstrated the inconsistent results of digital intervention in managing hypertension. In the study by Kario et al [9], 390 patients with essential hypertension who were not using antihypertensive medication were enrolled. The study found that the digital intervention group (HERB system: an interactive smartphone app [HERB Mobile] with a web-based patient management console [HERB Console] + standard lifestyle modifications) had a significantly greater reduction in 24-hour ambulatory, home, and office SBP at 12 weeks compared to the control group (standard lifestyle modifications only). Another randomized controlled trial from the United Kingdom included 622 hypertensive patients who had been treated but had poor BP control. The digital interventions consisted of self-monitoring of BP, patient education, promotion of healthy lifestyles, and medication guidance. After 12 months, the average BP in the digital intervention group decreased from 151.7/86.4 mm Hg to 138.4/80.2 mm Hg, whereas in the usual care group, it decreased from 151.6/85.3 mm Hg to 141.8/79.8 mm Hg, indicating that digital interventions led to better control of SBP than conventional care [29]. However, a randomized, open-label trial of 297 adults with uncontrolled hypertension showed that while participants randomized to a smartphone coaching app had greater self-confidence in controlling BP, they did not have lower BP at 6 months compared with those receiving a BP tracking app [30]. Conducted in real-world clinical practices in low-resource settings, our study demonstrated that the effectiveness of digital interventions was limited. This may be due to the relatively low educational levels and older age of the study participants, most of whom had only a primary school education, and the mean age of the digital intervention group was 62 years.

Although the role of digital interventions in BP management was not significant in this study, it still holds some implications for hypertension control in rural areas. First, self-measurement of BP and health education should be widely promoted in rural areas to increase patients’ self-awareness in managing their BP. Second, the effectiveness of digital interventions was limited among populations with lower educational levels and older age groups. It is important to better understand the barriers to using digital interventions faced by these individuals. Information on which patients would benefit most from a digital therapeutic intervention would be valuable for targeting the intervention to those patients. In addition, investment in targeted measures is necessary to enable the most vulnerable groups to participate in digital health initiatives and reduce the risk of digital health exacerbating health disparities.

Although we conducted a sample size calculation, the hypertension control rates in the 2 groups were quite similar. A larger sample size or an extended follow-up period may be needed to better understand the effects of the digital intervention. In addition, some patients measured their BP at home during follow-up, and we applied the same criteria as for office BP to determine whether their BP was controlled, whereas current guidelines recommend a difference between the home BP target and office BP target [13]. Moreover, while in-person clinic follow-up was the preferred method, some patients were unable to attend due to various barriers, particularly transportation difficulties. For these individuals, telephone follow-up was provided as an alternative. It is conceivable that these differing follow-up modalities could have impacted the results. Furthermore, measurement errors may have occurred because some patients self-monitored their BP rather than having it uniformly measured by a physician. Beyond these, the study design presented a potential for selection bias, as participation in the intervention group required the ability to use the WeChat mini-program.

Our study demonstrated that digital interventions increased hypertension control rates in rural areas, although this improvement was not statistically significant. Nevertheless, providing convenient BP measurements and health education to these patients notably enhanced hypertension control rates. Further studies are necessary to validate the efficacy of these interventions.