This study adopted an explanatory sequential mixed-methods design, which integrates quantitative and qualitative evidence in formative digital health research [17,18]. Quantitative pre-post changes were collected through a 24-week single-group pilot intervention of the LeYi app. Subsequently, postintervention usability questionnaires and semistructured interviews were conducted to supplement and interpret numerical findings. This standard sequential mixed-methods procedure helps enrich the understanding of intervention performance and user experience. The study was reported following the CONSORT-EHEALTH checklist (Checklist 1).

Based on methodological guidance for pilot and feasibility studies [19], prespecified feasibility criteria were set as follows: (1) recruitment rate ≥50%; (2) retention rate ≥80%; (3) daily monitoring adherence ≥70% (days with at least one BP reading/168 days); and (4) app usability ≥4.0 on the user version of the Mobile App Rating Scale (uMARS; overall quality score; scale 1‐5). These criteria were used to interpret the feasibility and acceptability of the intervention.

The LeYi mHealth app was developed using the Intervention Mapping protocol, a systematic 6-phase framework that emphasizes theory- and evidence-based intervention design [20]. This process ensured that all app features were directly aligned with the identified needs of the target population and grounded in established behavioral science theory.

This study was conducted from September 1, 2024, to October 22, 2025, in the cardiology outpatient clinic of a tertiary care hospital in Leshan, China.

The inclusion criteria were as follows: (1) age 18‐59 years; (2) diagnosis of essential hypertension according to the Chinese Guidelines for the Prevention and Treatment of Hypertension [21]; (3) current prescription of at least one antihypertensive medication; (4) owning a smartphone and proficiency in basic operations (eg, installing apps and receiving messages); and (5) voluntary participation and providing written informed consent.

The exclusion criteria were as follows: (1) presence of secondary hypertension; (2) history of severe psychiatric illness or cognitive impairment that would preclude study participation; (3) history of drug or alcohol abuse; (4) plan to become pregnant, current pregnancy, or breastfeeding during the study period; (5) concurrent participation in other clinical trials that could interfere with this study; and (6) any other conditions deemed unsuitable for participation by the researchers (eg, life expectancy of less than 1 year).

A convenience sampling method was used, and a total of 30 eligible patients were enrolled. According to sample size guidelines for pilot studies, a sample of 12 to 30 participants per group is considered appropriate for pilot studies primarily focusing on feasibility assessment and effect size estimation [22]. Moreover, this study adopted an explanatory sequential mixed-methods design, with in-depth interviews planned for all participants. A sample size of 30 patients also provides a robust foundation for achieving thematic saturation in qualitative data analysis. All enrolled participants completed the 24-week intervention and subsequent assessments, with no dropouts, yielding a final analytical sample of 30 participants.

All participants were instructed to install the LeYi app on their smartphones. Prior to the intervention, trained researchers provided standardized, face-to-face, one-on-one instructions (approximately 15‐20 min), focusing on the operation of core functions. The 24-week intervention was delivered entirely through the app, which was developed following the Intervention Mapping framework and systematically integrated nudge strategies, including default options, immediate feedback, and social norms, to guide health behaviors by optimizing the digital decision-making environment.

The intervention’s core mechanism involved decomposing self-management goals into daily micro-tasks with low cognitive load and reinforcing positive behavior through real-time feedback and visualization. According to the Chinese Guidelines for the Management of Hypertension (2024 Revision), the recommended home BP monitoring target is systolic BP (SBP) <135 mmHg and diastolic BP (DBP) <85 mmHg. In this study, BP control was defined accordingly. Participants were guided to use 5 core functional modules, each designed based on the aforementioned theoretical principles (see Table 1 for strategy mapping and detailed feature descriptions). Screenshots of the app interface are provided in Multimedia Appendix 1. The key interaction logic operated as follows: establishing medication routines through standardized adherence support; enabling visual tracking of health data via structured BP monitoring; translating health goals into concrete actions through contextualized lifestyle guidance; facilitating early risk identification and professional intervention through a tiered alert and clinician response mechanism; and providing ongoing information and support through closed-loop patient-provider communication and education. Together, these 5 modules constituted a continuous management cycle encompassing monitoring, alerting, intervention, and support.

To ensure smooth initiation of the intervention, the research team provided proactive technical adaptation support via telephone or in-app messages during the first 2 weeks, addressing initial barriers such as login difficulties and operational challenges. Throughout the intervention period, the app operated autonomously according to preset logic (eg, medication reminders and report generation). The critical early warning workflow functioned as follows: when a user-entered BP value reached the tiered alert threshold defined in accordance with the Chinese Guidelines for the Prevention and Treatment of Hypertension, the system issued an alert to the user and simultaneously transmitted the warning information to the responsible clinician’s terminal. The clinician then conducted a manual assessment and delivered personalized guidance via the secure in-app messaging module. The research team reviewed anonymized, aggregated BP reading availability on a weekly basis to monitor overall engagement at a macro level.

The app was designed to function as a “digital bridge” connecting patients’ daily self-management with professional medical support. By integrating intelligent assistance with human collaboration, it aimed to enhance the feasibility, safety, and sustainability of self-management.

Patients submitted home BP readings through the app 4 times daily (before meals and at bedtime). The system automatically graded BP levels using color-coded labels: red for BP ≥140/90 mmHg and orange for BP 130-139/80‐89 mmHg. When an alert was triggered, the clinical team reviewed the patient’s data, assessed the reading, checked for potential measurement errors, and delivered individualized guidance within 24 hours. For readings suspected to be affected by improper measurement technique, patients received retraining on standard measurement methods and were asked to remeasure. Risk grades were reassessed based on the remeasurement results. For confirmed abnormal readings, clinicians provided advice through the secure in-app messaging module. In cases of severe hypertension or symptoms, medical staff contacted the patient immediately by telephone and recommended nearby medical attention. All patients underwent BP reassessment after the intervention. Patients with controlled BP were placed under long-term follow-up, while those with recurrent red alerts were managed as high-risk cases and arranged for offline outpatient visits.

Two research assistants were responsible for baseline data collection. They assisted participants in operating their smartphones and successfully installing the LeYi app. Pretest questionnaires were completed in a quiet, private room within the cardiology outpatient clinic.

After the 24-week intervention, the same research assistants collected posttest questionnaire data at the same venue. Subsequently, one-on-one semistructured interviews were conducted with each participant by a cardiovascular nurse who had received training in qualitative research methods. No relationship was established with participants prior to study commencement; participants were informed only of the researcher’s professional role and the study purpose. Only the researcher and participant were present during the interviews.

The interview guide was developed based on a literature review and was pilot tested with 2 patients (not included in the final sample) to refine the questions. Each interview lasted approximately 30‐45 minutes and was audio-recorded. Field notes were taken during and after each interview. All interviews were transcribed verbatim, and transcripts were returned to participants for member checking to verify accuracy. Data collection continued until thematic saturation was achieved, meaning no new themes emerged from subsequent interviews. Interviews focused on participants’ experiences with the app and their perceived changes in self-management behaviors.

This study was approved by the Institutional Review Board of the authors’ affiliated institution (approval number: LYLL[2024]KY107). All participants provided written informed consent prior to enrolment. Participants were assured that their identity and all personal information would be kept strictly confidential. To ensure confidentiality, unique identification numbers were assigned to each participant and used throughout the data analysis process in place of their names.

Statistical analyses were performed using IBM SPSS Statistics for Windows, version 27.0 (IBM Corp). All participant data were entered and systematically verified for consistency. Categorical variables are presented as frequencies and percentages. Continuous variables, which were not normally distributed as determined by the Kolmogorov-Smirnov test, are reported as medians with IQRs. Within-group comparisons between pre- and postintervention were conducted using the nonparametric Wilcoxon signed-rank test. Statistical significance was set at a 2-tailed P value of <.05.

Although the primary inferential analysis used nonparametric Wilcoxon signed-rank tests due to nonnormal distributions, means, 95% CIs, and Cohen d values are also reported as descriptive effect size measures to facilitate comparison with other studies. These estimates are presented as supplementary descriptive statistics and should not be interpreted as parametric inferential results.

For qualitative data, thematic analysis was used to systematically examine the interview transcripts. Two researchers independently performed the analysis; both had substantial knowledge of the patient population through clinical and research experience, as well as expertise in qualitative methodology. Coding and categorization were facilitated using Microsoft Word and Excel. An inductive and iterative approach was adopted. The researchers familiarized themselves with the data by repeatedly reading the transcripts, after which initial codes were generated. These codes were subsequently collated into potential themes [26]. Themes were then reviewed, defined, and named following verification of coding accuracy and data completeness. Throughout the process, the 2 researchers worked independently and held regular discussions to compare coding structures, refine thematic classifications, and reconcile interpretations until consensus was achieved.

Quantitative and qualitative data pertaining to the usability of the LeYi app were analyzed separately within the mixed-methods framework. An integrated interpretation of these findings will be presented in the Discussion section.

All prespecified feasibility criteria were met or approached. The recruitment rate was 75% (30/40), exceeding the 50% criterion. Retention was 100% (30/30), exceeding the 80% criterion. Daily monitoring adherence was 67.5%, approaching the 70% criterion. App usability was favorable, with a median uMARS overall quality score of 4.5 (IQR 4.3‐4.7), exceeding the 4.0 criterion. These findings suggest that the intervention was feasible and acceptable in this pilot context.

Of the 40 individuals invited to participate, 30 (75%) provided written informed consent and were enrolled. No intervention-related adverse events were reported during the study period.

Table 2 shows the baseline characteristics of the 30 enrolled participants. The mean age was 39.17 (SD 7.89; range 27‐58) years, with the largest subgroup aged 31‐40 years (17/30, 57%). Of the 30 participants, 19 (63%) were male and 11 (37%) were female. The majority were married (27/30, 90%). Regarding educational attainment, over half of the participants (17/30, 57%) held a bachelor’s degree or higher, whereas 27% (8/30) had completed high school or vocational secondary education or below. Current smokers comprised 33% (10/30) of the sample, nonsmokers comprised 60% (18/30), and former smokers comprised 7% (2/30). Moreover, 53% (16/30) reported current alcohol consumption. With respect to antihypertensive regimens, half of the patients (15/30, 50%) were prescribed a single agent, 14 (47%) were taking 2 or 3 agents, and only 1 (3%) was prescribed 4 or more agents.

As shown in Table 3, self-management behavior showed a significant pre-to-post improvement. The median score increased from 68.0 (IQR 63.0‐75.5) at baseline to 108.5 (IQR 105.5‐111.0) postintervention. The mean change was 38.93 (95% CI 35.05-42.81), with a large effect size (Cohen d=3.75; z=4.788; P<.001). Similarly, eHealth literacy showed a significant improvement. The median score increased from 15.5 (IQR 14.0‐24.5) to 34.0 (IQR 33.0‐36.0), with a mean change of 15.67 (95% CI 12.88-18.46) and a large effect size (Cohen d=2.10; z=4.544; P<.001). However, due to the uncontrolled single-group design, these improvements cannot be attributed specifically to the intervention. Graphs for self-management behavior and eHealth literacy are provided in Figures 1 and 2, respectively.

This pilot study, using an explanatory sequential mixed-methods design, suggests that the nudge theory–based LeYi app exhibits good feasibility and shows preliminary trends in self-management among young and middle-aged patients with hypertension. Quantitative findings revealed notable changes after the intervention: after 24 weeks, patients’ self-management behaviors (z=4.788; P<.001) and eHealth literacy (z=4.544; P<.001) showed significant pre-to-post increases relative to baseline. Moreover, the app received a favorable usability rating, with a median uMARS overall quality score of 4.5, indicating that its high information quality and ease of use provided a foundational basis for the observed changes. Consistent with previous research examining self-management in elderly patients with hypertension [27], our study extends these findings to a younger population and explores the potential of theory-driven mHealth interventions to support self-management behaviors.

The joint display (Table 5) illustrates how the qualitative themes consistently supported and contextualized the quantitative improvements. Participants who reported greater self-management improvement also described how visualized health monitoring (theme 1) helped them internalize healthy behaviors, while those with higher eHealth literacy gains frequently mentioned the value of real-time feedback and information support (theme 4). These findings were concordant and mutually reinforcing. However, a partially discordant case also emerged. Despite the overall improvements in self-management and eHealth literacy, some participants with lower daily monitoring adherence reported that daily data entry felt burdensome (theme 5). This suggests that while the app was beneficial for many, the requirement for manual entry may pose a barrier to long-term engagement for a subset of users. This discordance highlights the need for future iterations to reduce user burden, for example, by integrating automated data synchronization.

Importantly, the qualitative findings elucidated the mechanisms behind these quantitative results. Thematic analysis of the interview data revealed 6 key dimensions related to intervention performance: at the cognitive level, visualized monitoring facilitated internalization of health awareness; at the behavioral and motivational levels, goal-oriented design strengthened motivation for sustained engagement; and at the support and contextual levels, continuous patient-provider interaction and information feedback optimized the support system, while technological burden emerged as a notable challenge. These insights collectively point toward future optimization directions, including the demand for expanded functionality. The qualitative evidence thus complements and corroborates the quantitative results, together constructing a comprehensive evidence chain that illustrates how this theory-driven mHealth intervention supports self-management through multidimensional pathways.

The core findings of this study provide empirical support for the application of nudge theory in chronic disease self-management, particularly within digital health interventions targeting young and middle-aged populations. The notable changes in patients’ self-management behaviors directly corroborate a central tenet of nudge theory: that optimizing the “choice architecture,” rather than imposing mandates, can effectively guide individuals characterized by “bounded rationality” toward healthier decisions [13]. More importantly, this study demonstrated how these abstract nudge principles can be operationalized into a coherent set of actionable digital intervention strategies, specifically tailored to address the behavioral challenges prevalent among young and middle-aged patients with hypertension, such as limited cognitive bandwidth and a propensity for immediate gratification.

The design of the LeYi app systematically integrated multiple categories of health-related nudge tools, as categorized by Blumenthal-Barby and Burroughs [15]. Commitment devices were embedded through goal setting and daily check-ins; environmental priming and immediate feedback were facilitated via data visualization and alerts; and decision costs were minimized by simplifying the recording process through default options. Qualitative insights from user interviews provided direct evidence regarding these strategies’ roles in overcoming common behavioral inertia and execution difficulties. Participants frequently reported that “reminders helped me build a habit” and that “visualizing the data made me clearer about what to do.” These findings resonate with the broader body of systematic evidence regarding nudge strategies and health-related behaviors, including guideline adherence among health care professionals [28], thereby underscoring the generalizability of these principles across varied contexts and actors.

The findings of this study offer a promising approach to address the specific self-management challenges encountered by young and middle-aged patients with hypertension. Due to their fast-paced lifestyles and pronounced health perception biases, sustaining engagement and ensuring long-term adherence in this population have remained persistent challenges in the field of mHealth [29]. The observed trends in this study show that an intervention grounded in nudge theory can help address these challenges [13,14]. Its value stems from the translation of behavioral economic principles into actionable digital strategies that systematically reshape patients’ decision-making environments.

First, by simplifying decision-making processes—through features such as rapid data entry—the cognitive and operational costs of initiating behavior are reduced, thereby accommodating individuals’ fragmented schedules and limited cognitive bandwidth [10,13]. Second, by providing immediate and intuitive feedback (eg, trend charts), the intervention transforms long-term, often imperceptible health benefits into short-term, visible signals, thereby counteracting inherent perception biases [30]. Finally, the integration of gentle, external reminders as preconfigured environmental cues actively bridges the intention-behavior gap [13,15]. The core of this multicomponent strategy is neither coercive nor didactic; rather, it creates a digital environment in which healthier choices become more convenient, salient, and accessible, thereby gently guiding users to overcome bounded rationality and behavioral inertia.

The high ratings for “functionality” and “ease of use,” together with users’ qualitative feedback emphasizing the app’s “simplicity of operation,” are not incidental findings. Rather, they represent direct outcomes of the design philosophy outlined above, collectively confirming that a design ethos centered on “reducing burden and enhancing efficiency” aligns closely with the pragmatic needs of this population, who prioritize utility and practicality [10]. Notably, the relatively modest scores for “engagement” (eg, entertainment value) did not affect overall performance. On the contrary, this observation suggests that for the young and middle-aged patients with hypertension targeted in this study, a tool that is efficient, reliable, and seamlessly integrable into daily routines is valued more highly than one featuring elaborate entertainment functions [11,29]. This insight advances our understanding of design priorities for digital health interventions tailored to this specific demographic.

This study provides preliminary evidence regarding the potential of a nudge theory–based mHealth intervention for supporting self-management among young and middle-aged patients with hypertension. By reshaping the decision-making environment to bridge the “knowledge-attitude-practice” gap, this approach demonstrates clear advantages in terms of accessibility, acceptability, and potential sustainability. Consequently, it offers both a valuable theoretical foundation and a viable digital tool prototype for designing public health interventions aiming to help improve hypertension control rates in this population [31].

Specifically, this model could be explored as a complementary component within existing hypertension prevention and control frameworks. In response to the limited reach of traditional in-person services among young and middle-aged adults, this intervention attempts to supplement these services through digital means. Its daily monitoring functionalities may contribute to earlier identification of health risks, while structured reminders and support could potentially facilitate adherence during the initial phases of treatment [29]. Furthermore, continuous data tracking and light-touch interactions may offer additional support for some patients during intervals between conventional outpatient visits. Naturally, these potential benefits warrant further validation in broader populations and over extended follow-up periods [31].

Moreover, this mHealth intervention may have potential for scalability. However, its cost-effectiveness remains theoretical, as no economic data were collected in this pilot study, and formal health economic assessments are needed in future research. mHealth interventions typically entail low marginal costs following initial development, providing an economic rationale for large-scale implementation [32]. Given the serious consequences of poorly controlled hypertension and its substantial socioeconomic burden, the potential benefits of preventive digital interventions merit careful consideration [33]. It should be noted that, as an initial pilot exploration, the primary value of this study lies in establishing the feasibility of the intervention and generating preliminary evidence to inform future larger-scale research, including rigorous health economic evaluations [34].

In summary, the thoughtful integration of such behaviorally informed digital tools into comprehensive hypertension management strategies—subject to further research and evaluation—may offer insights into addressing certain gaps in current service delivery. This approach also contributes to accumulating evidence and experience for developing innovative chronic disease prevention and control strategies suited to the digital era.

The findings of this pilot study offer several practical implications for cardiovascular nurses working with young and middle-aged patients with hypertension.

First, the LeYi app can serve as a tool that extends nursing care beyond clinical settings. Nurses can introduce the app during discharge education, guiding patients to use medication reminders and BP monitoring features. During follow-up telephone consultations, nurses can review the app’s data dashboard with patients to track adherence and identify areas needing additional support.

Second, the visualized health data generated by the app, such as BP trend charts and medication adherence logs, can facilitate nurse-patient shared decision-making. By reviewing these visualizations together, nurses can help patients better understand the relationship between their daily behaviors (eg, diet and physical activity) and BP control, thereby reinforcing motivation for sustained self-management.

Third, the tiered alert system integrated with clinician notification demonstrates how nurses can leverage technology for proactive intervention. When patients receive automated alerts for abnormal BP readings, nurses can follow up promptly via the app’s secure messaging module, providing timely guidance and preventing complications.

However, the technological adaptability challenges reported by some participants highlight an important nursing role: providing initial guidance and ongoing technical support. Nurses should assess patients’ digital literacy at baseline, offer personalized instructions on app use, and remain available to address operational difficulties, particularly for those less comfortable with smartphone technology.

By incorporating such theory-driven digital tools into routine practice, nurses can play a pivotal role in bridging the intention-behavior gap and promoting sustainable self-management in this challenging population.

A principal strength of this study lies in the development of the LeYi app, a locally tailored mHealth intervention for young and middle-aged patients with hypertension. Designed to fit their lifestyle and usage needs, the app uses the local language, which improves cultural adaptability and user acceptability.

This study also has several limitations. First, it adopted a single-group pre-post design with no control group. Observed changes are only preliminary within-participant trends and cannot be attributed entirely to the intervention. The Hawthorne effect, regression to the mean, and concurrent routine care may also have contributed to the observed changes. Second, this was a single-center pilot study with a small sample size (n=30), limiting the generalizability of the findings. Large-sample multicenter randomized controlled trials are needed for further validation. Third, outcomes mainly depended on self-reported scales, and the 24-week follow-up was too short to assess long-term behavioral changes and clinical endpoints. More objective indicators and longer follow-ups are required in future work. Fourth, participants needed basic smartphone skills, which may have caused inequity due to differences in digital access and literacy.

Despite the above limitations, the findings of this pilot study indicate that theory-driven mHealth tools can support hypertension self-management. Moreover, the study provides practical experience for shifting hypertension care toward proactive digital prevention.

The findings of this pilot study suggest that the LeYi app, a mobile health intervention grounded in nudge theory, may be feasible and acceptable for supporting self-management behaviors and eHealth literacy among young and middle-aged patients with hypertension. With features, including simplified decision-making, real-time feedback, and strengthened patient-provider interaction, the app applies behavioral economic principles to build a user-friendly digital choice architecture. This design helps users cope with behavioral inertia and cognitive biases in self-management and offers a noncoercive way to guide health behaviors. Given the single-group design without a control arm, these findings are only preliminary trends. Larger randomized controlled trials are required to further verify the intervention’s actual efficacy.

This study provides preliminary evidence for adopting theory-driven digital health tools as a complementary approach in hypertension management. Such tools may help optimize hypertension control. Further studies with larger samples and longer follow-ups are needed to explore the generalizability and long-term performance of this intervention. Overall, this work offers useful theoretical references and practical implications for developing user-oriented, localized mHealth tools for hypertension and promotes digital development in chronic disease prevention and management.