BACKGROUND
The American Heart Association Predicting Risk of Cardiovascular Disease EVENT (PREVENT) equations were developed from observational research cohorts and electronic health record data and provide sex-specific risk estimates for cardiovascular disease (CVD), atherosclerotic CVD (ASCVD), and heart failure (HF). External validation in large contemporary samples across multiple health systems in the United States is needed.
METHODS
We assembled a national electronic health record-based cohort of US adults with individual-level patient data pooled from a collective of 30 health systems (Truveta) to externally validate the outcome-specific 10-year PREVENT equations (PREVENT-CVD, PREVENT-ASCVD, and PREVENT-HF). We included patients aged 30 to 79 years without a history of prior CVD and with an ambulatory encounter in the electronic health record between 2013 and 2018. The outcomes were defined as total CVD (composite of ASCVD and HF), ASCVD, and HF through December 2024 using diagnosis codes. Model performance of the outcome-specific PREVENT base equations was assessed with the Harrell C statistic and calibration slope, stratified by sex.
RESULTS
Of the 680 864 adults included, the mean (SD) age was 55 (13) years, and 56% were female. Over a mean (SD) follow-up of 6.8 (2.3) years, there were 29 535 incident CVD events, 19 280 incident ASCVD events, and 16 824 incident HF events. The median (interquartile range) 10-year predicted risk of PREVENT-CVD among women was 3.6% (1.3%-8.8%), and among men was 5.8% (2.5%-11.7%). The C statistic (95% CI) was 0.788 (0.786-0.790), and the calibration slope (95% CI) was 0.98 (0.95-1.01) for PREVENT-CVD. PREVENT-ASCVD and PREVENT-HF demonstrated similar C statistics (0.774 [0.771-0.777] and 0.824 [0.820-0.828]) and calibration slopes (1.07 [1.04-1.10] and 1.01 [0.97-1.04]) for prediction of the 10-year risk of ASCVD and HF, respectively.
CONCLUSIONS
The PREVENT equations accurately and precisely estimate the 10-year risk of CVD, ASCVD, and HF in a large sample of US adults. These findings support the generalizability of the PREVENT equations to inform guideline-recommended risk assessment and preventive efforts.
Sadiya S. Khan, Y. Sang, Xiaoning Huang et al.· Circulation. Population heal...· 0 citations
Background Dietary assessment is the cornerstone of clinical management and research studies evaluating diet and health. Traditional methods such as food diaries and 24-hour recalls can be burdensome, prone to recall bias, and difficult to adhere to. Image-based dietary assessment using vision-language models (VLMs) offers a potential solution. Objective Our goal was to benchmark state-of-the-art VLMs for automated food recognition, weight estimation, and calorie estimation using Google’s Nutrition5k dataset. Methods We evaluated 3,229 food images using ten approaches: proprietary VLMs (Gemini 2.0 Flash, 2.5 Flash, 3.0 Flash, and 3.1 Flash Lite; GPT- 4o, GPT-4o-mini, and GPT-5 Mini; and Claude Haiku 4.5), an open-source VLM (Qwen2-VL-7B), and a commercial food recognition API (FatSecret). We assessed calorie and weight estimation using Lin’s Concordance Correlation Coefficient (CCC) and component detection using Jaccard similarity. Results Gemini 3.0 Flash achieved the best calorie estimation (CCC 0.767, MAE 80.7 kcal), while Gemini 3.1 Flash Lite offered very comparable accuracy (CCC 0.754) with the highest ingredient recognition (Jaccard 0.655) at the lowest cost among top-performing models ($0.59/1K images). Among earlier-generation models, Gemini 2.0 Flash remained competitive (CCC 0.742, Jaccard 0.621) at a fraction of the cost ($0.10/1K images). A human validation study in which four annotators reviewed 440 images revealed systematic omissions in the original Nutrition5k labels. After correction, the extrapolated ingredient-overlap score for Gemini 2.0 Flash increased from 0.62 to an estimated 0.82, suggesting that raw Jaccard scores substantially underestimate true model performance. Conclusions Current VLMs can perform automated dietary assessment with reasonable accuracy from single overhead photographs. Our results inform model selection for dietary assessment applications and highlight remaining challenges in calorie estimation and component detection for complex, multi-item meals.
Sam Sterling, Lauren T. Berube, Andrea J. Glenn et al.· bioRxiv· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
