Gram Research analysis of a 2026 systematic review found that most studies using U.S. national food survey data examined links between diet and health, with heart disease the most common topic at 18% of publications. Energy, protein, carbohydrates, fat, fiber, and sodium were the nutrients reported most often.
Every few years, the U.S. government collects detailed information about what Americans eat through a national survey called What We Eat in America, part of NHANES. Scientists use this data in thousands of studies. A new review, covering 2,200 published studies from 2013 to 2023, looked at what these studies were about. According to research reviewed by Gram, most of the studies looked for links between diet and health, and heart disease was the most common health topic, appearing in 18% of the publications. Diabetes, weight, bone health, and cancer also came up often. The nutrients reported most frequently were energy, protein, carbohydrates, total fat, fiber, saturated fat, sodium, and several other fats. The authors used artificial intelligence to speed up the work, with human experts checking the results. The review suggests that keeping this nutrient information accurate matters for the health research that depends on it.
Key Statistics
A 2026 systematic review of 2,200 peer-reviewed publications from 2013 to 2023 found that heart disease was the most commonly reported health outcome, appearing in 18% of studies using U.S. national food survey nutrient data.
The same 2026 systematic review found that about 85% of association and mortality studies using U.S. national food survey data reported at least one diet-related health outcome.
A 2026 systematic review of 2,200 studies found that association analyses made up 79% of publications using U.S. national food survey nutrient data from 2013 to 2023.
The Quick Take
- What they studied: The researchers wanted to know what health topics scientists study using a large U.S. government food survey, and which nutrients those studies report most often.
- Who participated: The review did not recruit people. It looked at 2,200 published studies from 2013 to 2023 that used data from the U.S. national food survey (What We Eat in America, part of NHANES).
- Key finding: Most of the studies (about 85% of the association and mortality studies) reported at least one health outcome. Heart disease was the most common topic, appearing in 18% of all publications, followed by weight (10%), diabetes (9%), bone health (4%), and cancer (4%).
- What it means for you: This review does not tell you what to eat. It shows that the food data behind many diet studies is widely used, so keeping it accurate helps the research that shapes nutrition advice may be more reliable. Talk with a doctor or registered dietitian before changing your diet for a health condition.
The Research Details
The researchers built on an earlier 2025 review that counted how often scientists used the U.S. national food survey data. That earlier review did not look closely at what the studies were about, so this team went back to those same publications and examined them in more detail.
To do this, they used an artificial intelligence tool to pull out key details from each paper, such as the health topic, the type of study, and the nutrients reported. Human experts guided the process, checked the tool’s work, and cleaned up the results before analyzing them.
The work followed five steps: writing clear instructions for the AI, extracting and organizing the information, checking that it was accurate, cleaning and analyzing the data, and then describing the patterns in plain language.
A systematic review gathers many studies on one question and looks at them together. Because this review covers 2,200 publications over ten years, it gives a broad picture of how the survey data is used. Using AI to sort through that many papers made the job faster, and the human checks were meant to catch mistakes the AI might make.
The review followed a clear, repeatable process, which is a strength. It also built on an earlier published review, so its starting list of studies was already established. Still, the AI tool and the human checkers may have made small errors when sorting studies, and the review only counts what each paper reported. It does not measure whether the findings in those papers are correct.
What the Results Show
Of the 2,200 publications, about 79% were association studies, which look for links between factors, such as diet and disease. About 30% were surveillance studies, which track how people eat or how health changes over time. About 27% used a dietary index, a score that rates how healthy someone’s eating pattern is. About 11% checked whether people were getting enough of certain nutrients. Some studies fit into more than one group.
Among the 1,753 studies that looked for links or tracked deaths, about 85% (1,496 studies) reported at least one health outcome. This shows that the survey data is used most often to study how eating habits relate to health.
Heart disease was the health topic that appeared most often, in 18% of all publications. Obesity and weight loss came next at 10%, followed by diabetes and insulin resistance at 9%. Bone health and cancer each appeared in 4% of publications.
The nutrients reported most often across health topics and study types were energy (calories), protein, carbohydrates, total fat, dietary fiber, saturated fats, monounsaturated fats, polyunsaturated fats, and sodium. These are the nutrients that most diet studies focus on.
The review found that different types of studies tend to report different nutrients. This suggests that researchers choose nutrients based on the question they are asking. The authors say this pattern can help set priorities for keeping the nutrient data in the survey accurate and complete, which matters for both research and public health guidance.
The 2025 scoping review by Kogan and colleagues showed how often this survey data was used in science. This newer review adds detail by showing which health topics and study designs the data supports. Together, the two reviews show that the survey is a major resource for nutrition research.
The review relied on an AI tool, which may have missed details or misread some papers, even with human checks. It only looked at publications that used this specific U.S. survey, so it does not describe nutrition research in other countries or using other data sources. It also counted what studies reported, not whether their results were strong or correct. Finally, the review does not state how many people were in the original survey studies, so the true number of participants is unclear.
The Bottom Line
Research shows that the U.S. national food survey is one of the most widely used sources for nutrition and health studies, with heart disease as the most common topic. This review does not test any diet or food, so it does not offer new eating advice. Its main value is showing why accurate nutrient data matters. Confidence in its findings about research patterns is moderate to high, since it is based on 2,200 publications. Confidence that these patterns show what foods are healthiest is low, because the review does not evaluate health results.
Nutrition researchers, public health planners, dietitians, students, and people who follow health news will find this useful. Readers who want personal diet advice should look to guidance from doctors or registered dietitians, not this review alone.
This review describes past research, so it does not produce changes in your health on its own. Any improvements in diet and health would come from long-term changes in eating habits, which often take months to show measurable effects, such as changes in blood pressure, blood sugar, or weight.
Frequently Asked Questions
What health topics do scientists study using the U.S. national food survey?
Heart disease is the most common topic, appearing in 18% of studies reviewed from 2013 to 2023. Obesity and weight loss appeared in 10%, diabetes in 9%, and bone health and cancer in 4% each. Most studies were designed to look for links between diet and health.
Which nutrients do researchers measure most often in national food survey studies?
Researchers most often report energy (calories), protein, carbohydrates, total fat, dietary fiber, saturated fats, and sodium. These nutrients appeared across many health topics and study types in a 2026 review of 2,200 publications.
How many studies used the What We Eat in America survey data?
A 2026 systematic review identified 2,200 peer-reviewed publications from 2013 to 2023 that reported nutrient data from the What We Eat in America survey, part of NHANES. Association studies made up about 79% of them.
Does this research prove which foods are healthiest?
No. The review described how the survey data is used in published research but did not test any foods or diets. It shows research patterns, not which foods improve health. Talk with a doctor or registered dietitian about personal nutrition decisions.
Want to Apply This Research?
- Track the nutrients this review highlights most often: calories, protein, carbohydrates, total fat, fiber, saturated fat, and sodium. Log them daily and review weekly totals, such as average daily fiber in grams or average daily sodium in milligrams.
- Use the app to set one small, measurable goal, such as adding one serving of high-fiber food per day or checking sodium on packaged food labels. Review your progress each week and adjust the goal as needed.
- Review your nutrient trends monthly over several months rather than day to day. Pair your food logs with regular health checks, such as blood pressure and blood sugar readings, and share the trends with your healthcare provider.
This article summarizes a systematic review of published studies. It describes how scientists have used U.S. national food survey data and does not test any diet, food, or treatment. The information is for general education only and is not medical advice. Consult a doctor or registered dietitian before making changes to your diet, especially if you have a health condition or take medication.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.