The purpose of this study was to develop a personalized step test and a valid regression model that used non-exercise data and data collected during the step test to estimate VO 2 max in males and females 18 to 30 years of age. All participants (N= 80) successfully completed a step test with the starting step rate and step height being determined by the self-reported perceived functional ability (PFA) score and participant's height, respectively. All participants completed a maximal graded exercise test (GXT) to measure VO 2 max. Multiple linear regression analysis yielded the following equation (R = 0.90, SEE = 3.43 mL≅kg-1 ≅min-1): 45.938 + 9.253(G)-0.140(KG) + 0.670(PFA) + 0.429(FSR)-0.149(45sRHR) to predict VO 2 max (mL≅kg-1 ≅min-1) where: G is gender (0=female;1=male), KG is body mass in kg, PFA is the sum of the two PFA questions, FSR is the final step rate (step-ups/min), and 45sRHR is the recovery heart rate 45 seconds following the conclusion of the step test. Each independent variable was significant (p < 0.05) in predicting VO 2 max and the resulting regression equation accounted for roughly 83% (R 2 =0.8281) of the shared variance of measured VO 2 max. Based on the standardized β-weights, gender (0.606) explained the largest proportion of variance in VO 2 max values followed by PFA (0.315), body mass (-0.256), FSR (-0.248), and the 45sRHR (-0.238). The cross validation statistics (R PRESS = 0.88, SEE PRESS = 3.57 mL≅kg-1 ≅min-1) show minimal shrinkage in the accuracy of the regression model. This study presents a relatively accurate model to predict VO 2 max from a submaximal step test that is convenient, easy to administer, and individualized.

The purpose of this study was to develop a multiple linear regression model to predict treadmill VO 2max scores using both exercise and non-exercise data. One hundred five college-aged participants (53 male, 52 female) successfully completed a submaximal cycle ergometer test and a maximal graded exercise test on a motorized treadmill. The submaximal cycle protocol required participants to achieve a steady-state heart rate equal to at least 70% of age-predicted maximum heart rate (220-age), while the maximal treadmill graded exercise test required participants to exercise to volitional fatigue. Relevant submaximal cycle ergometer test data included a mean (±SD) ending steady-state heart rate and ending workrate equal to 164.2 ± 13.0 bpm and 115.3 ± 27.0 watts, respectively. Relevant non-exercise data included a mean (±SD) body mass (kg), perceived functional ability score, and physical activity rating score of 74.2 ± 15.1, 15.7 ± 4.3, and 4.7 ± 2.1, respectively. Multiple linear regression was used to generate the following prediction of (R = .91, standard error of estimates (SEE) = 3.36 ml·kg −1 ·min −1 ): VO 2max = 54.513 + 9.752 (gender, 1 = male, 0 = female) -.297 (body mass, kg) + .739 (perceived functional ability, 2-26) + .077 (work rate, watts) -.072 (steady-state heart rate). Each predictor variable was statistically significant (p < .05) with beta weights for gender, body mass, perceived functional ability, exercise workrate, and steady-state heart rate equal to . 594, -.544, .388, .305, and -.116, respectively. The predicted residual sums of squares (PRESS) statistics reflected minimal shrinkage (R PRESS = .90, SEE PRESS = 3.56 ml·kg −1 ·min −1 ) for the multiple linear regression model. In summary, the submaximal cycle ergometer protocol and accompanying prediction model yield relatively accurate VO 2max estimates in healthy college-aged participants using both exercise and non-exercise data.