大学生心理障碍影响因素的Logistic回归分析

背景:随着现代社会的发展,大学生心理障碍在国内己引起民心理健康教育界高度关注。目的:探讨大学二年级学生心理障碍的影响因素。设计:整群抽样,1∶2配比观察,病例对照研究。单位:武汉工业学院医学院。对象:于2001-06在武汉工业学院通过整群抽样选择1200名2年级学生,年龄19～20岁,经SCL-90测试(包含躯体化、强迫、人际敏感、抑郁、焦虑、敌对、恐怖、偏执、精神病性等9种因子,采用分级评分法,级数越高,症状越严重)分析筛出145名神经症中因子分≥3项的学生(包括精神分裂样症状、抑郁症、焦虑症)为心理障碍病例组。采用病例与健康对照按1∶2配比,分别选与病例同班级,同专业、年龄相近(相差≤2岁),经SCL-90各项指标均正常290名学生作配比组,两组受试对象中,男生173名,女生262名,平均年龄19.2岁。方法:①对两组受试对象进行心理卫生调查及艾森克人格问卷(成人版)进行人格特征调查。运用华中科技大学同济医学院少儿教研室EPQ统计软件处理,获得内外倾、神经质、精神质、掩饰性等4个分值及量表剖析图和内、外倾-神经质坐标图的象限位置确定人格倾向。②以心理障碍为应变量,各有关因素(对心理障碍有影响因素为父母理解自己的程度,与家庭成员的关系,对学习的要求,人格特征,精神质分值偏高,内、外倾分值偏高、偏低,神经质分值偏高、偏低)为自变量,应用单因素1∶2配比Logistic回归分析。③在单因素1∶2配比Logistic回归分析的基础上,将P<0.10的因素放入多因素1∶2配比Logistic回归模型中,采取逐步回归的方法筛选变量。主要观察指标:①心理障碍组与正常组学生SCL-90得分比较。②心理障碍相关因素单因素及多因素分析结果。结果:纳入435名学生全部进入结果分析。①心理障碍学生组学生躯体化、强迫、人际关系、抑郁、焦虑、敌意、恐怖、偏执、精神病性各因子均分高于正常组(t=13.39,29.64,27.09,26.04,19.52,18.13,14.15,18.61,22.41,P<0.01)。②经Logistic回归分析,单因素分析选出9项因子,其中父母理解自己的程度、与家庭成员关系、对学习的要求、人格特征、E>61.5、神经质<38.5等6项因素为保护因素,其他3项(精神质分值>61.5、内、外倾分值<38.5、神经质分值>61.5)为危险因素;多因素分析结果显示与家庭成员关系、对学习的要求、人格特征3项因素为保护因素,神经质分值>61.5为危险因素。结论:大学二年级学生心理障碍与学生同家庭成员关系,父母理解的程度,对学习的要求程度以及人格特征和个性特质有关。

Logistic regression analysis of the influencing factors of psychological problems faced by college students

BACKGROUND: With the rapid development of modern society, psychological problems in college students have been paid high attention by educatienal organizations of psychological health in China. OBJECTIVE: To probe into the influencing factors of psychological problems in college students. DESIGH: Cluster sampling 1∶2 paired observation, case-control study. SETTING: Medical College of Wuhan Polytechnic University. PARTICTPANTS: 1 200 sophomores with the age ranged 19-20 years were selected from Wuhan Polytecnic University by cluster sampling at June 2001, and 145 students with the scores of more than 3 factors greater than 3 points were screened by Symptom Checklist 90 (SCL-90) (including 9 factors of somatization, compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobia, paranoid and schizophrenia. The higher the grade was, the severer the symptom was) to be the psychological problem group. Subjects were selected to be the normal controls at the ratio of 1∶2 with those in the psychological problem group, whose classes, specialties and ages (the difference was less than 2 years) were similar to those of the enrolled patients. 290 students normal in the indexes of SCL-90 were taken as the matched group. Of all subjects in both experiment groups, there were 173 males and 262 females with an average age of 19.2 years. METHODS: ① The subjects in the two experiment groups were investigated of the psychological health by SCL-90 as well as the personality characteristics and Eysenck Personality Questionnaire (EPQ) (For Adult). The EPQ statistic software designed by the Institute for Juvenile and Children, Tongji Medical College of Huazhong University of Science and Technology was adopted to acquire the E score (introversion, extroversion), N score (nervousness), P score (psychoticism), L score (conceal) and quantity form quadrant of E-N coordinate picture, so as to define the tendencies of characters. ② The psychological problem was considered to be the dependent variable, while all relevant factors (The influencing factors of psychological problem included the understanding from parents, relationship with family members, desire for learning, personality characteristics, over-high P score, over-high/low E score, over-high/low N score) were taken as the independent variable. Single-factor Logistic regression analysis was conducted at 1:2. ③ Based on 1:2 single-factor Logistic regression analysis, the variances were selected from factors of P ＜ 0.01 together with the Logistic regression models of multifactor at 1:2 by stepwise regression method.MAIN OUTCOME MEASURES: ① Comparison in the SCL-90 score between the psychological problem group and the normai control group. ② Single-factor and multivariate analysis of the relevant factors of psychological problems. RESULTS: A total of 435 enrolled students were involved in the analysis of results. ① The average score of somatization, compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobia, paranoid and schizophrenia in students of the psychological problem group was higher than that in the normal control group (t=-13.39, 29.64, 27.09, 26.04, 19.52, 18.13, 14.15, 18.61, 22.41, P＜ 0.01). ② Nine factors were selected in the single factor analysis after the Logistic regression analysis, including 6 protective factors of understanding from parents, relationship with family members, desire for learning, personality characteristics, E ＞ 61.5, N ＜ 38.5, while he other 3 factors (P ＞ 61.5, E ＜ 38.5, N ＞ 61.5) were risk factors. The multivariate analysis showed that factors of relationship with family members, desire for learning as well as personality characteristics were protective factors, while it was a risk factor when N ＞ 61.5.CONCLUSION: The psychological problems in college students relate to the relationship with family members,understanding from the parents, desire for learning as well as the character trait and personality characteristics.