They remain reliable in populations that look nothing like the ones they were originally built on.
For decades, medicine has wrestled with a quiet but consequential question: can a tool built in one corner of the world be trusted in another? A large international study, drawing on data from 6.4 million people across dozens of cohorts and clinical trials, now offers a measured answer — two of the most widely used cardiovascular risk algorithms, Prevent and Score2, hold their predictive power across borders, populations, and health systems. In a field where the stakes are heart attacks and strokes, that consistency is not a minor technical footnote; it is the foundation upon which global preventive medicine can be built.
- Until now, two major heart risk tools had been validated mainly in the regions that created them, leaving clinicians elsewhere uncertain whether the predictions could be trusted for their patients.
- A collaboration spanning 44 observational studies and 18 clinical trials pooled data on 6.4 million people, generating nearly 300,000 documented cardiovascular events — enough real-world evidence to put the question to a serious test.
- Researchers found that both Prevent and Score2 maintained solid, consistent performance across populations that differed markedly in geography, ethnicity, and clinical context, directly challenging the assumption that these tools were regionally bound.
- The study's mixed methodology — combining long-term observational follow-up with randomized trial data — strengthened the conclusion by showing the algorithms held up not just in controlled research settings but in the complexity of actual clinical practice.
- Clinicians in countries where these tools have not been standard can now apply them with greater confidence, potentially improving decisions about who needs preventive treatment and who does not across the full spectrum of global health systems.
A question that has quietly troubled global cardiology — whether risk-prediction tools developed in one region can be trusted elsewhere — has now received a rigorous answer. A study published in Nature Medicine, involving the Biomedical Research Institute of Lleida among other institutions, pooled data from 44 observational studies and 18 randomized clinical trials to validate two algorithms: Prevent, developed by the American Heart Association, and Score2, embedded in European clinical guidelines. Together, the datasets covered 6.4 million people across different continents and health systems.
Both tools serve the same essential purpose: estimating an individual's likelihood of suffering a heart attack or stroke in the coming years, and helping doctors decide whether preventive treatment is warranted. The problem was that Prevent had been validated primarily in American populations and Score2 in European ones, leaving an obvious gap whenever either was applied beyond its home territory.
Over an average follow-up of 5.1 years, researchers documented 293,737 cardiovascular events under Prevent's criteria and 258,086 under Score2's — a volume of real-world outcomes large enough to test whether the algorithms' predictions matched reality across genuinely diverse settings. They did. José Manuel Valdivielso, who leads the Translational Vascular and Renal Research group in Lleida, highlighted that the tools retained their predictive power even in populations that looked nothing like those they were originally built on.
The practical implications reach beyond academic validation. Doctors in regions where these algorithms have not been extensively used can now apply them with confidence, identifying high-risk patients who might otherwise be missed and avoiding unnecessary treatment for those at low risk. The study does not merely confirm that two tools work — it expands the geography of evidence-based preventive medicine.
Two widely used tools for predicting heart disease risk have now been tested on a scale that settles a long-standing question: do they work the same way everywhere, or only in the places where they were invented?
The answer, according to a study published in Nature Medicine, is that they work. Researchers from 44 observational studies and 18 randomized clinical trials pooled data on 6.4 million people across different continents and health systems to validate two algorithms called Prevent and Score2. The work involved the Biomedical Research Institute of Lleida in Spain, among other institutions. What they found was that both tools maintained solid, consistent performance across populations that differed significantly from one another—a result that matters because until now, these algorithms had been tested mainly in the regions where they originated.
Prevent, developed by the American Heart Association, and Score2, embedded in European clinical guidelines, do the same essential job: they estimate an individual's likelihood of having a heart attack or stroke in the coming years, and they help doctors decide whether someone needs preventive treatment. They're not new tools. But they had lived largely in their home territories. Prevent was validated primarily in American populations; Score2 in European ones. The gap in the research was obvious: what happens when you use an American algorithm on a Japanese patient, or a European one on someone from Brazil? Does the prediction still hold up?
This study answers that question by examining what actually happened to millions of people over time. During an average follow-up period of 5.1 years, researchers recorded 293,737 cardiovascular events using Prevent's definition of the disease, and 258,086 using Score2's definition. The sheer volume of events—nearly 300,000 across the combined dataset—gave the researchers enough real-world data to see whether the algorithms' predictions matched reality across different settings.
José Manuel Valdivielso, who leads the Translational Vascular and Renal Research group at the Lleida institute, emphasized that the study's real contribution was demonstrating that these tools don't lose their predictive power when you move them across borders. They remain reliable in populations that look nothing like the ones they were originally built on. That consistency matters for global health. It means a doctor in a country where Score2 hasn't been extensively tested can still use it with confidence. It means the American algorithm can travel.
The validation was rigorous in its scope. The researchers didn't just look at one type of patient or one health system. They examined data from diverse geographic regions and clinical contexts, including both observational studies—where researchers simply followed people over time—and randomized trials, where patients were assigned to different treatments. This mix of study types strengthened the conclusion. The algorithms worked not just in ideal research conditions, but in the messier reality of actual clinical practice.
For preventive medicine, the implications are practical. Doctors worldwide now have stronger evidence that these risk-prediction tools can guide their decisions about who needs treatment and who doesn't, regardless of where those patients live. The algorithms can help prevent unnecessary treatment in low-risk people and identify high-risk people who might otherwise be missed. As these tools become more widely adopted in regions where they haven't been the standard, this validation gives clinicians the confidence to use them.
Notable Quotes
The tools maintain solid and consistent performance in populations very different from one another— José Manuel Valdivielso, Biomedical Research Institute of Lleida
The Hearth Conversation Another angle on the story
Why does it matter that these algorithms were only tested in their home regions before?
Because an algorithm built on American hearts might miss something about European hearts, or vice versa. You don't know if the patterns hold up when the population changes—different genetics, different diets, different healthcare systems. Until you test it, you're guessing.
So this study is basically saying: we tested them everywhere, and they still work?
Exactly. Nearly 300,000 real heart attacks and strokes across 6.4 million people. That's not a guess anymore. That's evidence.
What would have happened if they didn't work globally?
Then doctors outside America or Europe would have had to either develop their own algorithms or use these ones knowing they might be unreliable. This validation means they can confidently use the same tool.
Does this mean the algorithms are perfect?
No. It means they're consistently good across different populations. There's always room for improvement, but now we know the tool doesn't break when you move it.
Who benefits most from this?
Patients in countries where these algorithms haven't been the standard. A doctor in Brazil or Japan can now use Prevent or Score2 knowing it's been tested on millions of people like their patients, not just on Americans or Europeans.
What happens next?
Wider adoption, probably. And maybe more research into whether you can improve these algorithms further, now that you understand how they perform globally.