线性回归主要用于分析连续资料结局(如血压值、血糖值等)的危险因素分析,但在慢性病学研究中,可能关注更多的是二分类资料的结局,如是否发生高血压、是否肥胖等,对于此类资料,采用线性回归并不适合,因为线性回归的结局理论上可以取任意值,而二分类资料只能取2个值。二分类资料结局的危险因素分析通常采用Logistic回归。慢性病学研究中,Logistic回归用途非常广泛,除了可以探索疾病危险因素及定量描述危险因素的 Linear regression is primarily used to analyze risk factors for continuous data outcomes (eg, blood pressure, blood glucose, etc.), but in chronic disease studies more attention may be given to the outcome of dichotomous data such as whether hypertension occurs or is obese For such data, the use of linear regression is not suitable, because the outcome of linear regression theoretically can take any value, and the binary data can only take two values. Logistic regression is commonly used to analyze risk factors for outcome in dichotomous data sets. In chronic disease research, Logistic regression is very versatile, in addition to exploring disease risk factors and quantifying risk factors