Using Heart Rate to Predict Energy Expenditure: A Validity Generalization Study

Abstract

Physical activity (PA) plays a critical role in preventing chronic diseases. Heart rate has been frequently used to predict PA energy expenditure (PAEE). While a number of prediction equations has been developed, their generalizability has not been confirmed. PURPOSE: To examine the validity generalization of heart rate prediction equations for PAEE. METHODS: Key words such as “energy metabolism,” “energy expenditure,” “heart rate,” “heart rate determination,” “prediction,” and “equation,” were searched in the scientific databases including PubMed, Web of Science, Google Scholar etc. Eligibility criteria was restricted to studies that predict PAEE using heart rate. The validity generalization model (Pearlman, Schmidt, & Hunter, 1980) was utilized for the analysis. Four components including sample size, observed validity coefficients, test reliability coefficients, and criterion reliability coefficients were summarized and examined from each study. When test and criterion reliability coefficients were not reported, the information was derived from the literature. The percentages of variance accounted for “by artifacts” were computed. RESULT: 98 validity studies were screened and 27 studies (M±SD: Sample size = 45±13; Validity coefficients = .70±.13; Test reliability coefficients = .83±.03; & Criterion reliability coefficients = .95±.02) were analyzed to determine the degree of validity generalization using heart rate to predict PAEE. The percentage of variance accounted for “by the artifacts” was only at 65%. Thus, the needed “75% decision rule” was no met. The estimated “90% credibility value” for the true validities was at .55, and the estimated average true validity was at .69. CONCLUSION: The validity to use heart rate predicting PAEE did not appear to be generalizable and these heart rate prediction equations should be used with caution, especially when it is used for another population.