Multimodal AI-Driven Out-of-Hospital Management for Hypertension and Cardiometabolic Risk

High blood pressure often occurs together with abnormal blood glucose, abnormal blood lipids, or abdominal obesity. Managing these conditions outside the hospital can be difficult. This study will test whether a digital health program that combines artificial intelligence (AI) with review and support by doctors and nurses improves health outcomes compared with standard out-of-hospital guidance. The study will enroll 150 adults aged 18 to 70 years who have hypertension and at least one additional cardiometabolic risk factor that is not at target. Participants will be randomly assigned in a 1:1 ratio to either the hybrid AI-assisted management group or the standard-guidance group. The intervention group will use a patient mini-program and connected devices to monitor blood pressure, glucose, and lifestyle information. The AI system will propose personalized medication, diet, exercise, and monitoring plans, which the healthcare team will review and confirm before delivery. The control group will receive routine printed health education and usual clinical follow-up. All participants will be followed for 6 months. The main outcome is the proportion of participants achieving blood pressure control together with control of at least one glycemic or lipid indicator. Other assessments include treatment adherence, cardiovascular events, hospitalizations, medical costs, quality of life, and system usability.

Inclusion Criteria: * Aged 18-70 years. * Confirmed hypertension plus at least one of dyslipidemia, hyperglycemia, or abdominal obesity, with at least one condition not at target. Hypertension: BP ≥130/85 mmHg and/or diagnosed hypertension under treatment. Hyperglycemia: fasting glucose ≥6.1 mmol/L or 2-hour post-load glucose ≥7.8 mmol/L and/or diagnosed diabetes under treatment. Abdominal obesity: waist circumference ≥90 cm in men or ≥85 cm in women. Dyslipidemia: fasting triglycerides ≥1.7 mmol/L, LDL-C ≥4.1 mmol/L in the general population, or failure to meet the ASCVD risk-stratified LDL-C target (\