Researchers used advanced computer programs to analyze health and diet information from over 12,000 American adults to figure out who might be at risk for heart disease. They found that certain nutrients—especially vitamin B12—along with age, cholesterol levels, and waist size were the strongest clues that someone might develop heart problems. This study shows that what you eat and basic health measurements can help doctors predict heart disease risk earlier, potentially allowing people to make changes before serious problems develop.

The Quick Take

  • What they studied: Can computer programs predict who will develop heart disease by looking at what people eat and their basic health measurements?
  • Who participated: Over 12,000 American adults aged 18 and older who participated in a national health survey between 2017 and 2023
  • Key finding: Computer models correctly predicted heart disease risk about 82% of the time, with age being the strongest predictor, followed by vitamin B12 levels, cholesterol, inflammation markers, and waist size
  • What it means for you: Getting enough vitamin B12 and maintaining healthy cholesterol and waist measurements may help reduce heart disease risk, though these findings need to be confirmed with more research before making major diet changes

The Research Details

Researchers looked at health and diet information collected from thousands of Americans as part of a national health survey. They used advanced computer programs (called machine learning models) to find patterns in the data that could predict who might develop heart disease. The computer programs learned from the data by looking at 41 different factors—including what people ate, their weight, blood pressure, and blood test results—to identify which factors were most important for predicting heart disease risk.

To make sure the computer programs were accurate, the researchers tested five different types of computer models and compared how well each one worked. They also used special techniques to explain why the computer programs made their predictions, so doctors could understand which factors were most important rather than just getting a yes-or-no answer from a black box.

The researchers had to handle an important problem: most people in the study didn’t have heart disease, so the computer programs had to be specially adjusted to learn effectively from this imbalanced data.

This approach is important because it helps identify which health and diet factors are truly most important for heart disease risk, rather than relying on guesses or old assumptions. By using transparent computer programs that explain their reasoning, doctors can understand and trust the predictions better than with older methods.

The study used a large, nationally representative sample of Americans, which makes the findings more likely to apply to the general population. The researchers tested multiple computer models and used established techniques to ensure accuracy. However, the study is based on survey data where people report what they eat, which can be less accurate than directly measuring diet. The findings show what factors are associated with heart disease risk but don’t prove that changing these factors will prevent heart disease.

What the Results Show

The computer programs successfully predicted heart disease risk with about 82% accuracy. One model (XGBoost) was slightly better at catching people who actually had heart disease, while another model (Random Forest) was slightly better at overall accuracy.

Age emerged as the single strongest predictor of heart disease risk—older people were at higher risk. After age, vitamin B12 levels were surprisingly important, followed by total cholesterol, C-reactive protein (a blood marker of inflammation), and waist circumference.

People who had heart disease in the study had noticeably lower levels of several nutrients compared to those without heart disease, including vitamin B12, niacin (vitamin B3), and other important nutrients. They also tended to have higher blood pressure, larger waist sizes, and higher inflammation markers.

The computer programs identified 30 key factors out of the 41 that were originally examined, meaning some factors weren’t as important for prediction as others.

Beyond the main predictors, the study found that several other nutrients and health measurements contributed to heart disease risk prediction. The fact that people with heart disease had consistently lower nutrient intake across multiple vitamins and minerals suggests that overall nutritional status may play a role in heart health. The study also confirmed that established risk factors like blood pressure, cholesterol, and body size remain important, alongside these dietary factors.

This study builds on decades of research showing that age, cholesterol, blood pressure, and weight are important for heart disease risk. The new finding is that vitamin B12 and other specific nutrients appear to be stronger predictors than previously thought. However, most previous research has focused on these traditional risk factors rather than detailed nutrient analysis, so this represents a newer perspective on the importance of diet quality.

The study has several important limitations. First, it’s based on survey data where people report what they eat, which can be inaccurate—people may forget what they ate or underestimate portions. Second, the study shows which factors are associated with heart disease but doesn’t prove that changing these factors will prevent heart disease; it’s possible that other unmeasured factors explain the connection. Third, the study only looked at one point in time for most participants, so we don’t know if these factors predict future heart disease development. Finally, the computer programs were tested on the same data they learned from, which can make them appear more accurate than they would be with completely new patients.

The Bottom Line

Based on this research, it’s reasonable to ensure adequate vitamin B12 intake (found in meat, fish, eggs, and fortified cereals), maintain healthy cholesterol levels, keep blood pressure in a healthy range, and maintain a healthy waist circumference. However, these are general health recommendations that should be discussed with your doctor, especially if you have a family history of heart disease. The evidence suggests these factors are important, but more research is needed to confirm that changing them will actually prevent heart disease.

Anyone concerned about heart disease risk should pay attention to these findings, especially people over 40, those with a family history of heart disease, or those with high cholesterol or blood pressure. People with existing heart disease should definitely discuss these findings with their cardiologist. However, these findings are most relevant for prevention in people who don’t yet have heart disease.

Changes in diet and lifestyle typically take several weeks to months to show effects on blood markers like cholesterol and inflammation. Meaningful changes in heart disease risk would likely take months to years of consistent healthy habits. Don’t expect overnight results, but consistent attention to these factors over time may reduce risk.

Want to Apply This Research?

  • Track weekly vitamin B12 intake (servings of meat, fish, eggs, or fortified cereals), monthly cholesterol and blood pressure readings if available, and waist circumference measurements monthly to monitor progress on the key predictors identified in this study
  • Add one vitamin B12-rich food to your diet daily (such as an egg at breakfast, a serving of fish at lunch, or fortified cereal), and measure your waist circumference monthly to track changes in this important risk factor
  • Create a simple dashboard showing your vitamin B12 intake frequency, waist circumference trend over time, and any available blood pressure or cholesterol readings. Set monthly goals to maintain or improve these key factors, and share results with your doctor during annual checkups to discuss whether your efforts are working

This study identifies factors associated with heart disease risk but does not prove that changing these factors will prevent heart disease. The findings are based on survey data and should not replace personalized medical advice from your doctor. If you have concerns about heart disease risk, have a family history of heart disease, or are considering making significant dietary changes, please consult with your healthcare provider or a registered dietitian before making changes based on this research. This information is for educational purposes only and is not medical advice.