Models For Comparing Strength And Power Measures Between Computerized Pneumatic Resistance Machines In Older Adults

INTRODUCTION: Maximal strength (1RM) and peak power (PP) are critical markers of physical function and independence in older adults. Among the various modalities to assess these variables, the most prevalent in research and clinical practice are computerized pneumatic resistance machines. Two major corporations, HUR (HUR Inc, Park Ridge, IL, USA) and Keiser (Keiser Corp., Fresno, CA) produce computerized pneumatic resistance machines that provide real-time force, power, and velocity data. PURPOSE: To develop equations to allow 1RM and PP values produced by each device to be compared. METHODS: Thirty-nine older men and women (72.9 ± 6.4y) completed 1RM and PP testing on the HUR and Keiser chest press machines. Anthropometrics including height, weight, and age were included in the initial modeling. Four models were then created: two regressed 1RM from each machine to the other, while the other two regressed PP between machines. Any nonsignificant indicators were excluded from the final model. RESULTS: Stepwise multiple regression modeling produced predictive models for allowing computations of 1RM and PP between machines. For computing HUR results from Keiser results the models were (HUR-1RM = 8.398 + .499[Keiser-1RM (kg)] - 7.173[Gender], where male = 0 and female = 1; r = .919, SEE = 5.221 kg, p < .001) and (HUR-PP = 119.523 + .417[Keiser-PP (W)]; r = .897, SEE = 62.74 W, p < .001). Models allowing conversion of chest press 1RM and PP from HUR to Keiser were (Keiser-1RM = 9.419 + 1.216[HUR-1RM (kg)]; r = .907, SEE = 7.38 kg, p < .001) and (Keiser-PP = 65.483 + 1.453[HUR-PP (W)]; r = .889, SEE = 105.089 W, p < .001). Using a subgroup of 70%, we determined that correlations between predicted and actual values were high (HUR-1RM = .955; HUR-PP = .924; Keiser-1RM = .957; Keiser-PP = .924). Further, coefficients of determination were high for all models (HUR-1RM = .85; HUR-PP = .82; Keiser-1RM = .82; Keiser-PP = .79). Mean absolute percentage errors were determined for each variable (HUR-1RM = 16%; HUR-PP = 32%; Keiser-1RM = 16%; Keiser-PP = 12%). CONCLUSIONS: These models allow clinicians and researchers to compare 1RM and PP results produced on each machine to be compared. However, we suggest developing new models using a larger sample of older men and women order to reduce prediction errors.