Machine Learning to Reduce Hypertension Treatment Clinical Inertia

Among individuals with an uncontrolled BP at the current visit, the objective of this study is to compare clinical management of hypertension with and without information from a machine learning algorithm on whether a patient will have uncontrolled blood pressure at their next follow up visit through a case-vignette study.

Among adults with uncontrolled blood pressure (BP) at a clinic visit, clinical inertia is common. Clinical inertia is defined as a failure of providers to initiate or intensify treatment (i.e., adding medication or increasing dosage) when guidelines indicate doing so. Prior studies report that clinicians intensify antihypertensive medication treatment in less than 20% of visits where intensification would have been clinically recommended. Thus, patients who have uncontrolled BP may not receive timely therapy to control their BP. To address this issue, the investigators will use a randomized design to test the hypothesis that clinicians will be more likely to intensify the hypertensive regimen and/or assess nonadherence for patients with uncontrolled BP at the current visit when presented with information that a patient is predicted to have uncontrolled BP at the next visit by a machine learning algorithm.

Uncontrolled blood pressure, Antihypertensive medication, Hypertension treatment, Clinical inertia, Machine learning

The investigators will create case vignettes to assess clinician hypertension management behavior, specifically antihypertensive medication intensification among individuals with uncontrolled blood pressure (BP). This arm will not include information from a machine learning algorithm designed to predict uncontrolled BP at a follow up visit.

The investigators will create case vignettes to assess clinician hypertension management behavior, specifically antihypertensive medication intensification among individuals with uncontrolled blood pressure (BP). This arm will include information from a machine learning algorithm designed to predict uncontrolled BP at a follow up visit about whether the algorithm predicts that the patient will have uncontrolled BP at the next visit.

Predicted uncontrolled BP status (yes/no) at follow up visit, derived using a machine learning algorithm

The investigators have created a machine learning algorithm to predict uncontrolled blood pressure (BP) status (yes/no) at a follow up visit among adults with uncontrolled BP at their current visit. The investigators will determine whether adding this information to a vignette describing a patient will increase the likelihood that a clinician will intensify antihypertensive medication treatment.

A clinical vignette of a patient with uncontrolled blood pressure will be presented to clinicians. They will then be asked to assess wether they would intensify antihypertensive medication treatment. This will be assessed using a 5-point Likert scale with scores ranging from 1 ("Strongly disagree") to 5 ("Strongly agree"). Likert scale scores will be compared between clinicians in the control and intervention group.

A second clinical vignette of a patient with uncontrolled blood pressure will be presented to clinicians. They will then be asked to assess wether they would intensify antihypertensive medication treatment. This will be assessed using a 5-point Likert scale with scores ranging from 1 ("Strongly disagree") to 5 ("Strongly agree"). Likert scale scores will be compared between clinicians in the control and intervention group.

A third clinical vignette of a patient with uncontrolled blood pressure will be presented to clinicians. They will then be asked to assess wether they would intensify antihypertensive medication treatment. This will be assessed using a 5-point Likert scale with scores ranging from 1 ("Strongly disagree") to 5 ("Strongly agree"). Likert scale scores will be compared between clinicians in the control and intervention group.

Inclusion Criteria: practicing Temple University Health System primary care clinicians who see patients (i.e., internal medicine, family medicine, attending physicians, nurse practitioners) will be eligible to participate - Exclusion Criteria: -