Impact of Patient Engagement on Blood Pressure Control Among Older Individuals With Hypertension in a Mobile Health Intervention: Longitudinal Analysis Using Latent Growth Curve Modeling

Abstract

Background: Limited research has investigated the influence of patient engagement on the long-term effects of mobile health (mHealth) interventions, particularly among older adults.

Objective: This study aimed to examine the long-term impact of a social media-driven mHealth intervention on blood pressure control among older Chinese individuals with hypertension, through repeated measurements of patient engagement and outcomes at 5 preset time points.

Methods: The study included older Chinese individuals with hypertension between 2017 and 2022. Participants received a hypertension self-management program via the WeChat social media app (Tencent Holdings Ltd), which provided clinically based digital coaching. Blood pressure measurements were taken repeatedly using a home blood pressure monitor (HBPM) connected to the app at baseline, 3, 6, 9, and 12 months. Patient engagement was evaluated based on the frequency of completed measurements at corresponding follow-ups. Latent growth curve models (LGCMs) served to assess the impact of patient engagement on blood pressure among older individuals with hypertension across preset points.

Results: A total of 1723 patients completed the 12-month follow-up (average age 70.1, SD 6.8 years; 890/1723, 51.7% female; and baseline systolic blood pressure 137.2 mm Hg). LGCMs revealed systolic blood pressure decreased significantly over 1 year, notably at 9 months (131 mm Hg, β9=3.244, P<.001), and continued up to 12 months (131.6mm Hg, β12=2.827, P<.001). In addition, a higher frequency of completed measurements was associated with better systolic blood pressure control at 3, 6, 9, and 12 months (β3=-0.016, P=.002; β6=-0.006, P=.02; β9=-0.002, P=.44; β12=-0.003, P=.02). These results remained significant even after accounting for age, sex, and comorbidity status.

Conclusions: This study, using LGCMs and repeated measures data, revealed a significant positive impact of patient engagement on long-term blood pressure control in mHealth interventions targeting older individuals with hypertension. These findings stress the importance of integration of patient-centered engagement approach into mHealth programs designed for chronic disease management in aging populations.

Keywords: aging population; hypertension; latent growth curve model; mHealth; mobile health; patient engagement.

Figure 1.. The diagram of the Sevenguards Health app. (A) Patient registration: after signing up for the Sevenguards Health app, patients input key demographic details like name, sex, date of birth, and comorbidity status. (B) Data upload: patients track their health by uploading data such as blood pressure readings, medication logs, physical activity records, and dietary information. (C) Assignment to health care provider: each patient is linked with a health care provider and receives a personalized management plan. (D) Patient engagement: patients use the app to track blood pressure and plan adherence. (E) Data storage: all collected data are securely stored on the Sevenguards server. (F) Provider support: providers monitor data and use WeChat to communicate, conduct follow-ups, and assist in blood pressure control.

Figure 2.. Changes over time in blood pressure and patient engagement. DBP: diastolic blood pressure; SBP: systolic blood pressure.

Figure 3.. Unconditional latent growth curve model for SBP (N=1723). Model fit statistics were χ²₇=89.351 (P<.001), Tucker-Lewis index=0.957, confirmatory fit index=0.970, standardized root mean square residual=0.112, and root mean square error of approximation=0.083. Observed variables are depicted with boxes. Latent variables are represented by ovals. Unidirectional arrows show the influence of one variable on another. Bidirectional arrows denote correlations. A dotted line indicates nonsignificant paths. Time points are represented as: 0 (baseline), 3, 6, 9, and 12 months post baseline. **P<.01, ***P<.001. SBP: systolic blood pressure.

Figure 4.. Conditional latent growth curve model for systolic blood pressure (SBP) with patient engagement (N=1723). Model fit statistics were χ²₃₂=121.150 (P<.001), Tucker-Lewis index=0.956, confirmatory fit index=0.968, root mean square error of approximation=0.040, and standardized root mean square residual=0.064. Baseline characteristics included age, sex, and comorbidity status. Unidirectional arrows show the influence of one variable on another. Bidirectional arrows denote correlations. A dotted line indicates nonsignificant paths. Time points are represented as: 0 (baseline), 3, 6, 9, and 12 months post baseline. *P<.05, **P<.01, ***P<.001.