Adding objectivity to Submaximal Exercise testing by Assessment of heart Rate ReCovery—a Healthy volunteer study (SEARCH)

Abstract

For each modelling method, after building the predicted model, the root mean square error (RMSE) from Leave One Out Cross Validation is calculated and after comparing, the model with the lowest RMSE is selected as the best prediction model. The multiple regression model included gender, PW/BW, and PW/BMI. MELD (measured at the time of CPEX test) score and Na (measured at the time of LT) can be used to predict LOH after LT. PW/BW and PW/BMI are both involved in the prediction model as independent variables and perform better than AT according to RMSE (18.34 vs 18.46, 18.47). PWI may be a better predictor of LOH than the currently used AT. Although promising, this finding needs to be validated in more and different LT centres.