The Efficacy of a Non-Exercise Estimated Cardiorespiratory Fitness Algorithm Without Physical Activity to Stratify Health Risk and Response to Treatment

Thumbnail Image
Abrego, Carlos G.
Cardiorespiratory fitness , Clinical Exercise Science
Non-exercise estimated cardiorespiratory fitness (eCRF) provides feasible estimates of CRF. The widespread use of electronic health records (EHRs) is rapidly emerging, allowing for the retrospective assessment of CRF. However, a majority of eCRF algorithms are derived using physical activity (PA) data, a main modifiable determinant of CRF that is not routinely documented in EHRs. For this reason, several eCRF algorithms have been derived without the inclusion of a PA data. However, these algorithms tend to lack accuracy due in part to being derived from relatively small samples. As a result, Sloan and colleagues derived an eCRF algorithm without a PA variable using a large sample and variables common to EHRs. Yet to be determined is whether eCRF derived using the Sloan algorithm can accurately: 1) predict mCRF 2) follow the exercise-induced change in mCRF. Six hundred and one adults with overweight or obesity participated in lifestyle-based, randomized control trials. Of these, 287 participants with complete pre- and post-intervention CRF values were included to determine the algorithm’s ability to follow exercise-induced change in mCRF determined using indirect calorimetry. No significant difference between eCRF and mCRF was observed at baseline. However, the observed error between eCRF and mCRF was large. Additionally, the estimates derived using the Sloan algorithm correctly classified 42 – 62% of participants into their respective mCRF quartiles. Change in eCRF was significantly different to the observed changes in mCRF, as the exercise-induced changes in mCRF were substantially greater than the change scores predicted by the Sloan algorithm. eCRF derived using the Sloan algorithm was not different from mCRF values. However, the observed error suggests that estimates derived using the Sloan algorithm lack the ability to accurately stratify CRF associated health risk on an individual basis. Estimates of CRF derived using the Sloan algorithm were not capable of predicting exercise-induced change in mCRF.
External DOI