M Wang1, A B Leger, G A Dumas. 1. Department of Mechanical and Materials Engineering, Mclaughlin Hall, Queen's University, Kingston, ON, Canada K7L 3N6.
Abstract
OBJECTIVES: The purpose of this study was to develop a regression equation to predict back extensor maximal voluntary contraction (back strength) for females based on several anthropometric and strength measurements using a multiple regression technique. BACKGROUND:Back strength is an important parameter in low back pain studies. However, the measurement of back strength is problematic in certain populations such as low back pain patients and pregnant women. METHODS:Back strength was measured as both moment at L4/L5 and force. Ten anthropometric or strength measurements were chosen to develop the prediction equation. The data used for developing the models were from eighty non-pregnant female subjects, age 18-42 and with no history of back pain in the past year. Backwards stepwise analysis was performed in order to choose the best fit predictors. The predictive ability of each of the models was checked using the cross-validation technique on 20 other subjects. FINDINGS: Two prediction models were developed for moment and force, respectively. The models explained 46.9% and 48.2% of the variance in back strength. No multicollinearity problem was found. The validation study showed that the observed back strength was highly correlated with the predicted back strength. INTERPRETATION:Mass, height, trunk length, grip strength and quadriceps strength are the best predictors of back strength in this study. The models developed in this study can be used for both general female low back pain patients and the pregnancy population.
RCT Entities:
OBJECTIVES: The purpose of this study was to develop a regression equation to predict back extensor maximal voluntary contraction (back strength) for females based on several anthropometric and strength measurements using a multiple regression technique. BACKGROUND: Back strength is an important parameter in low back pain studies. However, the measurement of back strength is problematic in certain populations such as low back painpatients and pregnant women. METHODS: Back strength was measured as both moment at L4/L5 and force. Ten anthropometric or strength measurements were chosen to develop the prediction equation. The data used for developing the models were from eighty non-pregnant female subjects, age 18-42 and with no history of back pain in the past year. Backwards stepwise analysis was performed in order to choose the best fit predictors. The predictive ability of each of the models was checked using the cross-validation technique on 20 other subjects. FINDINGS: Two prediction models were developed for moment and force, respectively. The models explained 46.9% and 48.2% of the variance in back strength. No multicollinearity problem was found. The validation study showed that the observed back strength was highly correlated with the predicted back strength. INTERPRETATION: Mass, height, trunk length, grip strength and quadriceps strength are the best predictors of back strength in this study. The models developed in this study can be used for both general female low back painpatients and the pregnancy population.
Authors: Markus Gerber; Serge Ayekoé; Bassirou Bonfoh; Jean T Coulibaly; Dao Daouda; Bomey Clément Gba; Benal Kouassi; Sylvain G Traoré; Rosa du Randt; Siphesihle Nqweniso; Cheryl Walter; Marceline F Finda; Elihaika G Minja; Getrud J Mollel; Honorati Masanja; Fredros O Okumu; Johanna Beckmann; Stefanie Gall; Christin Lang; Kurt Z Long; Ivan Müller; Nicole Probst-Hensch; Uwe Pühse; Peter Steinmann; Juerg Utzinger Journal: BMJ Open Date: 2022-06-06 Impact factor: 3.006