精神筛查量表（SSMD）的初步研制

发布时间：2018-03-27 10:12

本文选题：精神障碍　切入点：筛查量表　出处：《福建医科大学》2012年硕士论文

【摘要】：目的：精神障碍流行病学结果取决于筛查量表的功能及诊断工具。专业的精神科或心身科门诊，以及非专业的心理咨询工作室乃至人才测评领域，均迫切需要简短有效的筛查量表，而国内外精神障碍的筛查问卷为数不多，如GHQ-12或SRQ-24也只能对轴I的部分病种有鉴别力，特别是智力因子很难用问卷法检出。为此，本研究试图改变测量学策略，编制新的比GHQ-12更广谱的精神障碍筛查量表。方法：经过3个试用版的临床病人初试，新编33条SSDM初始问卷，加上GHQ12条目，在心身科门诊采集1554例各种随机病人及208例正常人，共有效样本1355例。多元统计分析了SSMD及GHQ-12的因子构成、条目分析后确定划界分，比较两套量表的信度、ROC曲线等。结果： 1.因子分析发现SSMD有20个独立的因子，解释总方差的80.323%，其中18个因子有症状扫描价值，包括了智能和人格障碍两个重要因子。GHQ-12只有3个因子，解释总方差67.680%，且多数条目不具有临床价值。SSMD与GHQ-12量表在抑郁因子上有少部分共同因子负荷。 2.条目分析结果显示SSMD只有2个条目不具有临床价值，再减去K条目1条，，保留30个条目，对不同病种的正答率高于GHQ-12。探索了筛查量表全新的条目设计技术，SSMD条目设计还可以再优化。 3. SSMD的Cronbach's Alpha系数为0.910，与GHQ-12的系数0.917相当。SSMD与GHQ-12总分的Spearman相关等级相关系数为0.816。 4.开创了临床法与统计学结合的划界方法：找出正常人与病人的分布图交叉点，剔除正常人中的潜在病人，ROC分析找出灵敏度特异度最合理的分界值。 5.ROC分析：SSMD以6分为初步划界分，其SSMD的灵敏度为94.6%，特异度为89.1%，假阴性率为5.4%，假阳性率为10.9%；GHQ-12以4为划界分的灵敏度为87.7%，特异度为79.8%，假阴性率为12.3%，假阳性率为20.2%。SSMD的灵敏度与特异度均高于GHQ-12，而假阴性率与假阳性率均低于GHQ-12。漏诊率明显低于GHQ-12。结论： 1. SSMD以最少的题目获得目前最多的诊断学因子； 2. SSMD的内部一致性信度与GHQ-12相当，SSMD的灵敏度，特异度比GHQ-12高；其假阴性率与假阳性率均低于GHQ-12。总体效度明显优于GHQ-12。达到最高的筛查量表的心理测验学技术水平。 3. SSMD划界分为6分；GHQ-12为4分。 4.开创了全新的临床法与测量学相结合的正常-异常划界分方法。 5.探索了筛查量表条目设计的新技术；SSMD的条目表述和因子构成还可以进一步完善改进。 6.我们心理卫生门诊应用说明，SSMD的症状扫描范围超过高年资的精神专科医生的个人经验，也优于多数临床病人的自主表达能力，且与病情及疗效直接相关。
[Abstract]:Objective: the epidemiological results of mental disorders depend on the function and diagnostic tools of the screening scale, specialized psychiatric or psychosomatic outpatient clinics, non-professional psychological counseling workshops and even the field of talent evaluation. There is an urgent need for short and effective screening scales, and there are few screening questionnaires for mental disorders at home and abroad. For example, GHQ-12 or SRQ-24 can only identify some of the diseases of axis I, especially the intelligence factor is difficult to detect by questionnaire. This study attempts to change measurement strategies and develop a new screening scale for mental disorders that is broader than GHQ-12. Methods: after the first trial of 3 trial versions of clinical patients, 33 new SSDM initial questionnaires and GHQ12 items, 1554 random patients and 208 normal subjects were collected in psychosomatic outpatient clinic. The factor composition of SSMD and GHQ-12 was analyzed by multivariate statistical analysis, the demarcation score was determined after item analysis, and the reliability of the two sets of scales was compared. Results:. 1. Factor analysis showed that SSMD had 20 independent factors, which accounted for 80.32323 of the total variance, 18 of which had symptomatic scanning value, including two important factors of intelligence and personality disorder. GHQ-12 had only 3 factors. Explain total variance 67.680 and most items have no clinical value. SSMD and GHQ-12 scale have a few common factor loads on depression factors. 2. Item analysis showed that only 2 items in SSMD had no clinical value. The positive response rate for different diseases was higher than that for GHQ-12. A new item design technique for screening scale was explored and SSMD item design could be optimized. 3. The Cronbach's Alpha coefficient of SSMD is 0.910, which is equivalent to that of GHQ-12 (0.917). The correlation coefficient of Spearman grade between GHQ-12 and SSMD is 0.816. 4. The method of combining clinical method and statistics was established: to find out the crossing point of distribution map between normal people and patients, and to eliminate the potential patients in normal people to find out the most reasonable boundary value of sensitivity specificity by ROC analysis. According to the 5.ROC analysis, the first division was divided into six categories. The sensitivity of SSMD is 94.6, the specificity is 89.1, the false negative rate is 5.4, the false positive rate is 10.9GHQ-12, the sensitivity is 87.7, the specificity is 79.8, the false negative rate is 12.3, the false positive rate is higher than GHQ-12, and the false negative rate is higher than GHQ-12. The false positive rate and false positive rate were lower than GHQ-12.The rate of missed diagnosis was significantly lower than that of GHQ-12. Conclusion:. 1. SSMD obtained the most diagnostic factors with the least number of questions; 2.The reliability of internal consistency of SSMD is equal to that of GHQ-12. The sensitivity and specificity of SSMD are higher than that of GHQ-12, and the false negative rate and false positive rate are both lower than GHQ-12.The overall validity is obviously better than that of GHQ-12.At the same time, the sensitivity and specificity of SSMD is higher than that of GHQ-12. 3. SSMD demarcation is divided into 6 minutes and GHQ-12 is 4 points. 4. A new normal-abnormal demarcation method combining clinical method and measurement is developed. 5. The new technique of item design of screening scale is explored. The expression and factor composition of SSMD items can be further improved. 6. Our mental health outpatient application showed that SSMD's symptom scan range exceeded that of the senior psychiatrist's personal experience, and was superior to the self-expression ability of most clinical patients, and was directly related to the condition and curative effect.
【学位授予单位】：福建医科大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R749

【参考文献】