Abstract: Objective　This study was designed to examine the psychometric properties of the Chinese version of 10-item Connor-Davidson Resilience Scale（CD-RISC-10） in a sample of Chinese patients with cancer.Methods　This study was conducted between December 2015 and July 2016.The psychometric properties of Chinese version of CD-RISC-10 were tested in a randomly selected sample of 520 cancer cases from the First Affiliated Hospital of Guangzhou University of Chinese Medicine，Sun Yat-Sen University Cancer Center and Guangdong Provincial Hospital of Chinese Medicine.The participants were also surveyed by other 2 scales〔Chinese version of Hospital Anxiety and Depression Scale（HADS） and Chinese version of Strategies Used by People to Promote Health（SUPPH）〕 to test their associations with the CD-RISC-10 and get the cut-off point of the CD-RISC-10.424（81.5%） responded to the survey positively，and 50 of them were randomly selected based on a Microsoft Excel-based algorithm and were resurveyed by the Chinese version of CD-RISC-10 two weeks later.The unidimensionality for the Chinese version of CD-RISC-10 was investigated by a two-parameter item response analysis，item analysis was performed using GRM.Furthermore，its convergent validity and retest reliability and cut-off point were analyzed.Results　The Chinese version of CD-RISC-10 retained its original undimensional structure，which could explain 58.06% of the total variance.The item separation index（ISI） was 0.914.The item difficulty was ideally normal，discrimination ability ranged from 1.17 to 1.82 and fit residual value of every item was within the range of -2.5-2.5.There existed no disordered threshold among all the items and the step of thresholds varied from 1.08 to 3.17 logits.The Chinese version of CD-RISC-10 was negatively associated with the Chinese version of HADS and its anxiety and depression subscales but positively associated with the Chinese version of SUPPH and its subscales in terms of scoring（P<0.05）．The retest reliability coefficient，cut-off point，sensitivity and specificity for the Chinese version of CD-RISC-10 was 0.855，21.5，0.735，0.833，respectively.Conclusion　The Chinese version of CD-RISC-10 has a stable undimensional structure with good monotonicity，ideally normal item difficulty，reasonable distribution of step of thresholds，as well as good reliability and external validity，so it is suitable to be adapted in clinical settings for assessing the psychological status of Chinese cancer patients.

Key words: Neoplasms；Resilience, psychological；Item response theory；Reliability；Validity；Cut-off

摘要： 目的　探讨中文版10条目心理韧性量表（CD-RISC-10）在恶性肿瘤患者中的信效度。方法　采用便利抽样法选取2015年12月—2016年7月于广州中医药大学第一附属医院、中山大学肿瘤防治中心、广东省中医院接受治疗的恶性肿瘤患者520例为调查对象，采用一般资料表、中文版CD-RISC-10、医院焦虑抑郁量表（HADS）、癌症患者自我管理效能感量表（SUPPH）对其进行调查。共发放问卷520份，回收有效问卷424份，有效回收率为81.5%。大样本施测2周后，采用EXCEL软件自带的随机程序从424例患者中选取50例采用中文版CD-RISC-10对其进行重测。采用双参数项目反应模型对中文版CD-RISC-10进行单维性检验、等级反应模型（GRM）条目分析，并分析其聚合效度、重测信度及截断值。结果　中文版CD-RISC-10仍保持单维结构，共解释58.06%的变异量，整个量表条目分离指数（ISI）为0.914。10个条目难度分布均匀，区分度为1.17~1.82，残差拟合度为-2.5~2.5，阈值间距为1.08~3.17 logits，不存在阈值倒错的情况。中文版CD-RISC-10得分与HADS总分及其焦虑、抑郁得分呈负相关，与SUPPH总分及其各维度得分呈正相关（P<0.05）。中文版CD-RISC-10的重测信度系数为0.855，截断值为21.5分，灵敏度为0.735，特异度为0.833。结论　中文版CD-RISC-10的单维性、单调性良好，条目的难度、阈值间距分布合理，同时具有较高的信度和外标效度，对于筛查情绪困扰的恶性肿瘤患者具有较高的敏感性，建议临床使用。

关键词: 肿瘤, 韧性, 心理, 项目反应理论, 信度, 效度, 截断值