Formulating grip strength and key pinch strength prediction models for Taiwanese: a comparison between stepwise regression and artificial neural networks

Abstract

Forceful exertions of the hand have been recognized as a risk factor for the development of work-related musculoskeletal disorders (MSDs) of the upper extremity. Gripping and pinching are frequently used hand strength in various occupational activities and in clinical evaluation of the hand. Therefore, formulating grip and pinch prediction models with easily obtainable personal parameters will help facilitate the design and evaluation of workplace environments or facilitate the hand impairment or progression assessments. This study developed maximum voluntary contraction (MVC) grip and key pinch strength prediction models using regression method and artificial neural networks (ANN). Thirty-three right-handed Taiwanese (19 males and 14 females) voluntarily participated in this study. Grip regression model (adjusted R 2 = 0.897) was formulated based on gender, hand length, weight and thumb length. Key pinch regression model (adjusted R 2 = 0.845) was developed using gender, fingertip to root digit 5 (little finger length), weight and maximum hand breadth. The ANN models predict grip and key pinch strength more accurately than regression models with smaller root mean square errors (RMSEs). However, no significant differences were found between values predicted by both kinds of equations.