The Prediction Model of Individual Differences in Health Promotion by Strength Training

Objective: To establish predictive models of strength training by analyzing individual differences and reproducibility of training effects, and to provide reference for formulating personalized and precise strength training program. Methods: Physical inactive adults completed 12 weeks of low-intensity training(LIST12, n=193), then the participants with low training effects conducted a 36-week washout period(DP, n=53). After that, 29 of them completed another 8 weeks(HIST8) and 12 weeks of highintensity training(HIST12). Questionnaire surveys, body composition, muscle thickness and muscle strength tests were conducted before and after each intervention. Individual effect size of Cohen's d effect size(ES) was calculated according to the strength index.Δ1 RM squat/bench press with an ES<0.6 was defined as inefficient, and ES≥0.6 was moderate or above. A strength training effect prediction model was constructed through multiple linear stepwise regression. Results: 1) After the LIST12 intervention, the average weight of 1 RM squat(Δ1 RM squat/%) was increased by 46.19%(ranges from-20.00% to 196.67%, P<0.01, ES=1.15), the inefficient rate was 19.17%. The average weight of 1 RM bench press(Δ1 RM bench press/%) was increased by 36.25%(ranges from-31.25% to 176.92%, P<0.01, ES=0.72), and the inefficient rate was 39.90%; 2) the training effect was washed out by DP, Δ1 RM squats was reduced by-32.03%(ranges from-392.16% to 0, P<0.01, ES=1.07), and Δ1 RM bench presses was reduced by-19.87%(ranges from-251.57% to 0, P<0.01, ES=0.55); 3) after the HIST8 intervention, the Δ1 RM squat was increased by an average of 19.80%(ranges from 3.57% to 43.10%, P<0.01, ES=0.66), the inefficient rate was 37.93%; and the Δ1 RM bench press was increased 22.84%(ranges from-11.76% to 60.00) %, P<0.01, ES=0.68), 44.83% of them were inefficient. After 12 weeks of HIST intervention, Δ1 RM squats was increased by 33.89%(ranges from 3.03% to 71.43%, P<0.01, ES=1.13), the inefficient participants was 17.24%; and Δ1 RM bench press was increased by 31.32%(ranges from-4.41% to 72.73%, P<0.01, ES=0.93), the inefficient participants was 34.48%. HIST reproduced the individual differences in the effectiveness of the LIST intervention. Only56%(1 RM squat) and 41%(1 RM bench press) of participants achieved moderate or above training effects after the LIST12, HIST8 and HIST12 interventions. Other participants were inefficient after one or two programs; 4) the predictive factors for the individual response of Δ1 RM squat were initial value of 1 RM squat weight(β=-0.765), intervention plan(β=-0.214), trunk muscle content(β=0.278), body fat content(β =-0.106), the total explanation to Δ1 RM squat was 40.9%; the factors for the individual response ofΔ1 RM bench press were initial value of 1 RM bench press weight(β=-0.849), upper limb muscle content(β=0.425), rectus femoris and intermediate femoris thickness(β=-0.189), which explains 42.5% of Δ1 RM bench press. Conclusions: The individual difference is observed in strength training, and the inefficiency in strength training can be reduced by modifying the training program. The initial value of 1 RM squat/bench press weight, intervention plan, trunk muscle content, total body fat content, upper limb muscle content, rectus femoris and intermedius femoris thickness can predict around 40% of individual differences in strength training.