A standardized medication algorithm cut strokes and heart attacks across California's largest public health system

University of California San Francisco

Seventy-two strokes that did not happen. Forty-eight heart attacks that never arrived. Thirty-eight people who did not die. These are not projections from a drug trial or estimates from a simulation. They are the calculated outcomes of a two-year program that did something deceptively simple: it gave doctors across six hospitals the same set of instructions for treating high blood pressure.

The 5.5 percentage point shift that saved lives

The University of California operates one of the largest public academic health systems in the United States, with more than nine million outpatient visits annually across six medical centers. In 2023, UC Health rolled out a tool called the UC Way Hypertension Medication Algorithm - a stepwise guide for increasing medication types and doses, built into the system's electronic health records. The idea was not new science. It was consistent execution of existing science.

Over two years ending in mid-2025, blood pressure control among roughly 90,000 patients rose from 68.5% to nearly 74%. That 5.5 percentage point improvement translated to approximately 4,860 additional people with controlled blood pressure - and the estimated prevention of 72 strokes, 48 heart attacks, and 38 deaths. The results were published in BMJ Open Quality.

Lead author Sandeep P. Kishore, MD, PhD, an internist and associate professor of medicine at UCSF, puts it in human terms: these are real Californians who did not end up in an emergency room, did not develop a disability, did not lose time with their families.

Why standardization matters when the science is settled

Hypertension is not a mystery. We know what causes it. We know how to treat it. The pharmacology is mature, the medications are inexpensive, and the evidence base is deep. Nearly 120 million American adults have high blood pressure, and it remains a leading cause of death, contributing to heart disease, heart failure, stroke, kidney disease, and pregnancy complications.

Yet only about half of people with hypertension have it under control. The problem is not a lack of effective drugs. It is a lack of consistent application - what doctors prescribe, when they escalate doses, and how they adjust for individual patient factors like age, race, and kidney function.

Before the UC Way algorithm, treatment approaches varied across UC's six medical centers. Different physicians used different protocols, escalated medications on different timelines, and prioritized different drug classes. The algorithm standardized those decisions while still allowing clinicians to adjust for individual factors and special populations, such as elderly patients.

The tool was developed by multidisciplinary teams including cardiologists, internists, primary care physicians, nurses, pharmacists, and data scientists. The group began meeting in 2020, spent three years building a comprehensive strategy that emphasized medication affordability and reduced treatment variation, and implemented the system in 2023.

The racial disparity the algorithm narrowed but did not close

High blood pressure hits some populations harder than others. By age 55, 75% of Black adults develop hypertension, compared with 54% of white men and 40% of white women. The UC Way tool improved hypertension control among Black patients from 63.4% to 67.3% - a meaningful gain, but one that still left a significant gap compared with overall control rates.

The authors are candid about this limitation. While the algorithm improved outcomes across racial groups, it did not eliminate disparities. The factors driving higher hypertension rates and lower control rates among Black and Hispanic populations extend well beyond prescribing patterns. They include social determinants of health - income, access to healthy food, neighborhood stress, healthcare access - that a medication algorithm alone cannot address.

Kishore and colleagues note that more targeted interventions are needed to close the remaining gap. Standardizing medication is necessary but not sufficient when the underlying inequities run deeper than the prescription pad.

Lifestyle recommendations alongside the algorithm

The UC Way program does not rely on medication alone. Study experts also recommend a set of evidence-based lifestyle practices: quitting smoking, limiting alcohol to two drinks per day for men and one for women, keeping sodium below one teaspoon per day, exercising at least 150 minutes per week, maintaining a healthy weight (BMI under 25), eating a balanced diet, and using at-home blood pressure monitors regularly.

These recommendations are not new. They appear in virtually every hypertension guideline. But packaging them alongside a structured medication protocol - and integrating both into the electronic health record - creates a more complete system than either approach alone.

What one health system's experience cannot prove

The study has limitations worth acknowledging. It was conducted within a single health system - large and diverse, but still one system with its own institutional culture, patient demographics, and infrastructure. Whether the same approach would produce similar results in community health centers, rural hospitals, or private practice networks remains an open question.

The study also lacked a randomized control group. All six UC medical centers adopted the algorithm simultaneously, so there is no untreated comparison group within the system. The improvement could theoretically reflect broader trends in hypertension management, changes in patient populations, or other concurrent initiatives. The authors address this by noting the timing and magnitude of the change, but the design does not allow for definitive causal claims.

Kishore also discloses advisory fees from Redesign Health and grants and equity in CareX.AI, a health technology company. No external funding for the study was disclosed.

From blood pressure to diabetes - and beyond

UC Health is already adapting the approach for diabetes management, and Kishore suggests it could be tailored to other chronic conditions in the future. The underlying principle is transferable: when effective treatments exist but are inconsistently applied, a standardized algorithm integrated into electronic records can close the gap between what we know and what we do.

The challenge, as Kishore puts it, is not the science. We know how to control blood pressure. What UC Health's experience demonstrates is that with the right infrastructure and institutional commitment, large and complex health systems can translate that knowledge into fewer strokes, fewer heart attacks, and fewer deaths. The algorithm is not elegant. It is not novel. But for 4,860 patients in California, it worked.