A framework for estimating cardiorespiratory fitness from diverse exercise tests
Gonzales, T. I.; Wareham, N. J.; Brage, S.
Show abstract
Comparisons of cardiorespiratory fitness across population studies are hindered by a lack of standardisation in exercise testing and reliance on estimation methods that assume steady-state exercise. We developed a generalised framework that standardises maximal oxygen consumption [Formula] estimation by harmonising dynamic heart rate modelling with individualised calibration of exercise energetics. We constructed and evaluated the framework using submaximal and maximal tests (treadmill walking and running, cycle ergometer, stepping, overground walking) from 911 adults. Agreement between framework estimates and directly measured [Formula] was strong (non-significant bias, Pearsons r=0.80). We then applied both framework and canonical methods to estimate [Formula] in 11,307 UK Fenland Study adults and examined associations with cardiometabolic risk. Framework-estimated [Formula] was more strongly associated with blood glucose, fasting insulin, and high-density lipoprotein cholesterol (combined risk factor model AIC=125,743) compared to canonical estimates (AIC=129,227). The framework standardises fitness estimation across diverse tests, yielding a more reliable and comparable biomarker of cardiometabolic health.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Estimating body fat distribution – a driver of cardiometabolic health – from silhouette images 91%
- Digital health technologies and machine learning augment patient reported outcomes to remotely characterise rheumatoid arthritis 90%
- Continuous-Time and Dynamic Suicide Attempt Risk Prediction with Neural Ordinary Differential Equations 90%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.