Objective To explore the associations of body mass index ( BMI) with common factors. Methods A-nalysis of the impact level and trend of common factors on BMI used Quntile regression modeling. Results There were statistically significant correlations between BMI and age, gender, waist-to-hip ratio, physical activity, diet index, and life event index. As BMI increased, the effect of age on BMI displayed a weak yet steady decline( e. g. , the coefficient between age and BMI decreased from -0. 044 at the 10th percentile to -0. 029 at the 90th per-centile);while that of gender, substantial and continued increase ( the coefficient between gender and BMI in-creased from 0. 329 at the 10th percentile to 1. 546 at the 90th percentile). Waist-to-hip ratio showed increasing positive relations through all BMI quantiles and the coefficient was 18. 024 at the 10 th percentile and 31. 716 for the 90th percentile. Physical activity was negatively associated with BMI for most quantiles (quantile>0. 4). The effect of diet behavior index remained relatively stable for most BMI quantiles yet began to increase rapidly starting from the 80th percentile of BMI. The effect of life event index on BMI started and ended relatively high and reached its bottom in the middle part of BMI quantiles. Conclusion The influence of common factors on different level of BMI presented statistical difference.%目的 探讨常见因素对留守村民体质指数( BMI)的影响. 方法 采用分位数回归等方法分析常见因素对BMI影响程度及趋势. 结果 随着BMI水平的增加,年龄对BMI的影响呈持续微弱下降趋势,在BMI 10%和90%分位数水平时回归系数(95%CI)分别为-0. 044( -0. 063,-0. 026)和-0. 029 ( -0. 050,-0. 009). 当BMI大于30%分位数后,性别对BMI的影响逐渐明显. 腰臀比与BMI呈正相关性,且相关性随着 BMI 的增加而增加. 在 BMI 的10%和90%分位数水平时回归系数( 95% CI )分别为 18. 024 (15. 454, 20. 594)、31. 716(28. 798,34. 633). 当 BMI 水平大于40%分位数时,体力活动量与BMI呈负相关性. 饮食指数与生活事件指数在BMI整个分位数范围内均呈正相关性,且饮食指数在BMI 80%分位数之前相关性较平稳(回归系数约为0. 830),之后急剧增加(回归系数最高达1. 267);而生活事件则呈现中间低两头高的趋势(回归系数最高为1. 164,最低为0. 714). 文化程度、吸烟史、饮酒史在BMI的整个分位数范围内均无相关性. 结论 常见因素对不同分位数水平BMI的影响有明显差异.
展开▼
