目的：(1)在冠心病人和正常人群中应用D型人格量表(DS14)中文版，检验其信、效度及其在这两个样本中的适用性。(2) D型人格在冠心病中的作用研究：包括D型人格者心理社会因素和临床资料的分析、D型人格与冠脉病变程度和冠心病危险因素聚集程度的关系、D型人格-冠心病不良预后的结构方程模型的研究，进一步探讨D型人格影响冠心病预后的作用机制和途径。 方法：(1) 187名冠心病病人和453名中老年正人完成了一般情况调查问卷、DS14、情绪强度量表、情绪表达性量表、90项症状清单(SCL-90)、Beck抑郁量表(BDI)以及A型人格量表(TAS-C)的评定；对DS14的信度检验采用Cronbach a系数、条目间平均相关系数和重测信度；效度检验包括结构效度和效标效度。 (2)对187名冠心病病人继续进行有关临床情况的测查，测查的内容包括：冠心病家族史、体重指数、心率、血压、血糖、血清总胆固醇、血甘油三酯、高密度脂蛋白、低密度脂蛋白、血浆同型半胱氨酸、C-反应蛋白以及冠状动脉造影术。采用t检验或X²检验比较冠心病人D型人格与非D型人格的心理测评结果和临床资料，采用多重回归分析检验D型人格及其与心理学变量和冠心病传统危险因素的交互作用对危险因素聚集性的预测性，采用结构方程模型研究D型人格影响冠心病预后的作用机制和途径。 结果：(1)信、效度检验：在正常组和冠心病组，DS14及其分量表消极情感(NA)和社交抑制(SI)的Cronbach a系数的范围在0.68～ 0.89，条目间平均相关系数的范围在0.24～0.52，NA和SI的重测信度系数分别为0.79和0.82，均在可接受范围或良好。验证性因素分析显示，DS14的两因素结构在两个样本中其X²/df<5；CFI、IFI、GFI均大于0.90；RMSEA均小于等于0.08；DS14总量表与分量表间以及两个分量表间相关系数的范围在0.58～0.95(p＜0.01)。NA分量表与消极情绪强度呈正相关(γ=0.22～0.29，p<0.01)，与积极情绪强度呈负相关或无显著相关(γ=-0.12～-0.19)，SI分量表与情绪表达强度呈负相关(γ=-0.26～-0.28，p<0.01)。DS14的NA和SI均与SCL-90的躯体化因子、抑郁症状因子、焦虑症状因子、恐怖症状因子和人际敏感因子呈正相关(γ=0.18～0.46，p<0.05～0.01)，D型人格的NA与SCL-90的敌意症状因子呈正相关(γ=0.23～0.38，p<0.05～0.01)，而SI则与SCL-90的敌意无显著性相关(γ=-0.03～0.15，p>0.05)。NA和SI均与BDI的三个因子呈显著正相关(γ=0.25～0.49，p<0.01)。 NA和SI分量表均与TAS-C的竞争和敌意分量表无显著相关性(p>0.05)；在冠心病组，SI与TAS-C的时间紧迫感分量表呈正相关(γ=0.18，p<0.05)；在正常组， NA与TAS-C的时间紧迫感分量表呈正相关(γ=0.18，p<0.05)。 (2) D型人格在冠心病中的作用研究：D型人格中有冠心病家族史者显著高于非D型人格(X²=6.86，p<0.05)，吸烟量、空腹血糖和C反应蛋白均显著高于非D型人格(p<0.05)。D型人格的SI显示出预测危险因素的聚集性的趋势，但未达统计学显著性(β=0.20，p=0.06)； 抑郁、焦虑、愤怒、情绪反应强度和情绪表达强度对冠心病危险因素聚集性的预测无显著性意义(p>0.05)；除了D型人格的SI与情绪表达强度的交互作用可减少对冠心病危险因素聚集性的预测外(β=-1.47， p<0.05)，NA和SI与负性情绪及情绪强度等心理因素的交互作用对危险因素聚集性的预测无显著性意义(p>0.05)。SI与血清总胆固醇的交互作用可增加对危险因素聚集性的预测(β=1.81，p<0.05)，而SI与吸烟的交互作用可减少对危险因素聚集性的预测(β=-1.03，p<0.05)。 D型人格的SI对冠脉病变程度的预测具有显著性意义(β=2.08，p< 0.05)。在D型人格-冠心病不良预后的结构方程模型假设中，基准模型和三个嵌套模型的比较显示，模型-3的拟合指标最佳(X²/df =1.682、RMSEA=0.071、CFI=0.909、IFI=0.912、GFI=0.908)，并且路径最简捷，所有路经系数的估计值均具有统计学显著性意义(p< 0.01)；所以，接受模型-3为最佳模型，既D型人格是通过负性情绪和生物学危险因素两个中介变量的间接作用影响冠心病预后。 结论：(1)D型人格量表(DS14)中文版在中老年正常人群和冠心病人两个样本中，均显示了良好的信、效度，适用于我国中老年人群和冠心病人的D型人格的评估。(2) D型人格的SI与总胆固醇的交互作用可增加对危险因素聚集性的预测性，而与吸烟的交互作用可减少这个预测力；D型人格的SI可预测冠脉病变的严重程度。(3) D型人格可能是通过中介变量影响冠心病预后，两条中介路径分别为负性情绪和生物学危险因素，负性情绪最终也是通过生物学机制间接影响冠心病预后。D型人格对负性情绪的影响效应大于其对生物学危险因素的影响效应。 关键词 D型人格，冠心病不良预后，信度，效度，结构方程模型

 Objective: To examine the validity and reliability of the Type D personality Scale (DS14) in both a coronary heart disease (CHD) sample and a normative sample. To explore the possible mechanisms of the adverse prognosis of CHD induced by Type D personality, and the following studies were included: the clinical and epidemiological characteristics of Type D personality in patients with CHD; the interactions between Type D and, ① psychological factors, ② classical risk factors for predicting the rate of clustering of CHD risk factors; the hypothesis of Type D personality—adverse prognosis in CHD patients using structural equation modeling (SEM). Methods: The study included 187 patients with CHD and 453 normative subjects, and they all filled out the following scales: Demographics questionnaire, DS14, Affect intensity scale (AIS), Emotional expressivity scale (EES), Symptom checklist-90 (SCL-90), Beck Depression Inventory (BDI), and Type A personality scales (TAS-C). CHD patients were administered clinical examination, including family history of CHD, body mass index (BMI), heart rate, blood pressure, serum total cholesterol (TC), triglyceride, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, fasting glucose, plasma homocysteine, C-reactive protein (CRP), and selective coronary angiography. The reliability analysis of DS14 included the Cronbach's alpha coefficients, the mean inter-item correlations, and the test-retest reliability. The validity analysis of DS14 included construct validity (confirmatory factor analyses are performed for two samples) and concurrent validity. The hierarchical multiple regression analyses were performed to examine the predictability of the DS14 for the rate of clustering of CHD risk factors. The analyses of SEM were used to investigate the possible mechanisms of the adverse prognosis induced by Type D personality in CHD patients. Results: (1) The Cronbach's alpha coefficients of the two subscales of DS14 ranged from 0.68～0.89, the mean inter-item correlations ranged from 0.28～0.52, the test-retest coefficients was 0.79 for NA and 0.82 for SI in the normative sample. A confirmatory factor analysis provided a good fit for a two-factor model in two samples (with X²/df＜5; CFI、IFL、 GFI;≥0.90; RMSEA≤0.08). The correlation coefficients of the two subscales (NA and SI) of DS14 with the total score of DS14 ranged from 0.58～0.95 (p＜ 0.01). NA was significantly correlated with the negative affect intensity of AIS (γ =0.22～0.29, p＜ 0.01), whereas it was not significantly correlated with the positive affect intensity of AIS (p ＞0.01). SI were significantly correlated with the emotional expressivity of EES (γ = -0.26～-0.28, p＜ 0.01). Both NA and SI were significantly correlated with the factors of SCL-90 as followings: somatization, depression, anxiety, phobic anxiety, and interpersonal relation (γ=0.18～0.46, p＜ 0.05～0.01); NA were positively correlated with the hostility factor of SCL-90 (γ =0.23～0.38,p＜ 0.05～0.01) whereas SI was not significantly correlated with it (p＞0.05). Both NA and SI were positively correlated with the three factors of BDI ( γ =0.25～0.49, p＜ 0.01). None of the correlation between DS14 (NA and SI) and the Competition and Hostility subscale of the TAS-C showed statistically significance for both samples excepting the moderate correlations between Time Hurry subscale of the TAS-C with SI in CHD sample (γ =0.18, p＜ 0.05), and with NA in normative sample (γ =0.18,p＜ 0.05). (2)The percentage of the family history of CHD, smoking, fasting glucose, and CRP in patients with Type D personality was higher than the ones in patients without Type D personality (p＜ 0.05 ). None of the two-way interactions between DS14 (NA and SI, respectively) and depression, anxiety, anger, affect intensity, and emotional expressivity, showed significant predictive effects for increased rate of clustering of CHD risk factors. The interactions between SI and TC predicted increased the rate of clustering of CHD risk factors (β=1.81, p ＜0.05); while the interactions between SI and smoking predicted decreased the rate of clustering of CHD risk factors ( β = -1.03, p ＜0.05). Elevated levels of SI predicted increased severity of coronary artery lesions in CHD patients (β=0.28, p＜0.05); A comparison among the baseline model and the three competitive models using SEM showed that the model-3 was the optimal model, with better fit indexes than the others (X²/df = 1.682、RMSEA=0.071、CFI=0.909、IFI=0.912、GFI=0.908), and with the most parsimonious path (p＜0.01). Conclusion: (1) The Chinese version of the DS14 demonstrated satisfactory reliability and validity in both CHD sample and normative sample, and can be considered as an appropriate tool for assessing Type D personality in Chinese sample. (2) In CHD patients, the interaction between SI and TC exhibited increased predictive power for the rate of CHD risks clustering while the interaction between SI and smoking exhibited the decreased predictive power; the elevated levels of SI predicted increased severity of coronary artery lesions. (3) Analysis using SEM showed that Type D personality predicted adverse prognosis in CHD patients via the two mediating pathways: one is the negative affects, and the other is the physiological risk factors. Furthermore, the negative affects exert indirect influence on adverse prognosis through the physiological mechanism. Key words Type D personality, adverse prognosis, validity and reliability, structural equation modeling