UBC Theses and Dissertations
Prediction of maximum oxygen uptake in paraplegics and quadraplegics using multiple regression equations Rogers, Allen Robert James
Twenty physically disabled subjects performed a progressive continuous exercise protocol on a wheelchair ergometer, to maximum exertion. Cardiorespiratory responses were monitored by means of direct ECG recording, and HR was reported for the last 30 seconds of each workload (WL). Expired gases were continuously sampled and analyzed for 15 second determinations of respiratory gas exchange variables. The last 30 seconds at each WL was averaged and assumed to be'representative of steady state responses to that WL. In addition, body weights, ages, and maximum breathing capacities were recorded. Two subjects were deleted from the study due to incomplete data. The inability to equate structural and functional characteristics of quadraplegics and paraplegics necessitated the division of the total group into paraplegics (N=13) and quadraplegics (N=5). For each paraplegic subject, three submaximal WL and corresponding cardiorespiratory responses were chosen for multiple regression analysis on maximal oxygen uptake. The three workloads were selected on the basis of HR responses, i.e., those workloads where HR was found to be between 65% and 85% of maximum HR (max HR = 220 - age). The lowest of the three workloads and corresponding cardiorespiratory responses (CRR) for each subject were assigned to the LHR group while the highest workloads and CRR for each were assigned to the HHR group. The remaining WL and CRR for each subject was assigned to the MHR group. Mean HR responses for the LHR, MHR, and HHR groups were: 131, 139,, and 148, respectively. These means corresponded to approximately 70%, 75%, and 80% of maximum heart rate. The squared multiple correlation coefficients (R²) adjusted for both sample size and number of variables contributing to R², were found to be .8761, .9218, and .9094 for LHR, MHR, and HHR, respectively. The respective standard error of estimates were reported to be .1397, .1101, and .1195 or ± CV% equal to 6.5%, 5.2%, and 5.6%. Cross validation was not performed due to the-small sample size. However, the adjusted press prediction gives an indication of how the equation may predict for subjects outside the experimental group. Prediction is performed in turn for each subject with the effects of that subject's data removed from the beta coefficients. The mean absolute errors reported were 22.2%, 11.8%, and 13.2% for the LHR, MHR, and HHR groups, respectively. Multiple regression analysis of the quadraplegic data was restricted to the WL and CRR at the fourth minute of the progressive continuous work- load protocol. The adjusted R² for the prediction equation produced was .9992 with a standard error of estimate of .0058 L/min. The variables contributing to the R² were: Ventilation, V02 ml/kg min., and WL. It was concluded that: 1) Multiple regression analysis appeared to be a suitable method of developing accurate prediction equations for MV02 L/min., in paraplegics. The best equation being: MV02 L/min. = 5.85 + 1.70(VC02 L/min.) - .026(age) - .0015(WL) - .008(HR) - 3.42(RQ). 2) The accuracy of prediction of MV02 L/min. in paraplegics is increased with the increase in the physiological stress (as reflected in % maximum HR). 3) Due to the limited sample size, no conclusions were reached regarding prediction of MV02 L/min. within the quadraplegic population.
Item Citations and Data