Machine-learning and hemodynamics-guided decision support to personalize blood pressure medicines in primary care

Implementation of Machine Learning and Hemodynamic Profiles Based Clinical Decision Support Systems for Personalized Guideline Accordant Antihypertensive Regimens in Primary Care: a Pragmatic Cluster Randomized Controlled Trial

This trial will test whether two computer decision tools—one using guidelines plus machine learning and the other adding hemodynamic measurements—help primary care doctors choose better blood pressure medicines for adults with hypertension.

Quick facts

• Phase: Not applicable
• Study type: Interventional
• Enrollment: 2160 (estimated)
• Ages: 35 Years to 79 Years
• Sex: All
• Sponsor: China National Center for Cardiovascular Diseases (Government)
• Locations: 1 site (Shenzhen, Shenzhen)
• Trial ID: NCT06828692 on ClinicalTrials.gov

What this trial studies

This is a prospective, multicenter, single-blind, pragmatic cluster randomized trial that randomizes 51 primary care sites in China 1:1:1 to a guideline+ML decision support system, a guideline+impedance cardiography+ML decision support system, or usual care. Randomization is stratified by region and type of primary healthcare institution (community health centers, community health stations, and township health centers). Adult patients aged 35–79 with systolic blood pressure ≥140 mmHg who regularly attend the participating clinics and meet medication criteria are enrolled at the clustered sites. The trial compares effectiveness and safety of the two CDSS approaches versus usual care for guiding antihypertensive regimens in routine primary care.

Who should consider this trial

Good fit: Adults aged 35 to 79 with hypertension who routinely attend participating primary care clinics, have a screening SBP ≥140 mmHg, and are taking no more than two classes of antihypertensive medications as defined by the protocol are the intended participants.

Not a fit: Patients with suspected secondary hypertension, very high blood pressure at screening (SBP ≥180 mmHg or DBP ≥110 mmHg), white-coat hypertension, a history of coronary heart disease or heart failure, or intolerance to multiple antihypertensive classes are unlikely to benefit or are excluded from participation.

Why it matters

Potential benefit: If successful, the intervention could help clinicians select more effective, personalized antihypertensive regimens leading to better blood pressure control and fewer side effects.

How similar studies have performed: Previous decision-support systems and guideline-based interventions have shown modest improvements in guideline adherence and blood pressure control, but combining machine learning with impedance cardiography for personalized antihypertensive selection is relatively novel and not yet widely proven.

Eligibility criteria

Inclusion Criteria:

1. 35 ≤ Age \< 80 years.
2. Regularly attend the clinic for hypertension treatment during the study period, without plans of traveling during the study period.
3. Diagnosed with hypertension, with SBP ≥140mmHg at screening.
4. Currently taking 0 or 2 classes of antihypertensive medications (A/C/D), with or without concurrent use of class B.
5. Willing to participate in the trial and capable of providing written informed consent.

Exclusion Criteria:

1. Physician-diagnosed or suspected secondary hypertension (e.g., hypertension caused by renal disease, renal artery stenosis, aortic stenosis, primary aldosteronism, pheochromocytoma/paraganglioma, Cushing's syndrome, thyroid dysfunction, intracranial disease, drug-induced, or rare monogenic genetic disease).
2. SBP ≥180mmHg and/or DBP ≥110mmHg at the screening visit.
3. Suspected or diagnosed white coat hypertension.
4. History of coronary heart disease.
5. History of heart failure.
6. Intolerance to 2 or more classes of A, C,D antihypertensive medications.
7. Currently taking antihypertensive medications other than class A,B,C and D.
8. Diagnosed chronic kidney disease, estimating Glomerular Filtration Rate (eGFR)\< 60 ml/min·1.73m2, or receiving dialysis.
9. Serious medical conditions (e.g., malignant cancer and hepatic dysfunction)
10. Currently in an acute episode of disease (e.g, new-onset cardiovascular and cerebrovascular disease occurred within 3 months).
11. Cognitive or communicative disorders.
12. Currently pregnant or breastfeeding or planning a pregnant or breastfeeding during the study.
13. Reluctant to take antihypertensive medications or have poor compliance with previous treatment.
14. Participating in other clinical trials.

Where this trial is running

Shenzhen, Shenzhen

• Fuwai Shenzhen Hospital, Chinese Academy of Medical Sciences — Shenzhen, Shenzhen, China (Recruiting)

Study contacts

• Principal investigator: Xin Zheng — Fuwai Hospital, China Academy of Medical Sciences, National Center for Cardiovascular Diseases
• Study coordinator: Xiaofang Yan

Conditions Hypertension; CDSS; cluster randomized controlled trial