功能状态评估量表的汉化、改良及在极低出生体重婴儿信效度检验

The localization and improvement of the Functional Status Scale and the reliability and validity in very low birth weight infants

目的:汉化、改良功能状态评分量表(FSS),并以极低出生体重儿(VLBWI)为对象检测其信效度。方法:征得原量表作者同意后,按照相关指南及规范,将原量表内容进行汉化并改良。纳入2018年1月1日至2020年6月30日收治住院的VLBWI,分别于出生后7 d及纠正胎龄34周时进行改良FSS评估,对得分进行描述统计(变异系数法、临界比值法和答案分布分析法)、信度分析[内部一致性信度(克伦巴赫α系数)和评分者间信度(Spearman相关系数)]及效度分析[内容效度(相关系数法、专家评分法)、结构效度(探索性和验证性因子分析法)和已知群效度(受试者操作特征曲线下面积、Pearson相关系数)],最后分析初、复评量表初步反应度。结果:根据纳入和排除标准,先后纳入548名和523名VLBWI进行初评及复评。描述统计中,变异系数法示条目的平均值与中位数接近,最大值和最小值接近或等于两端值,变异系数均大于0.15;临界比值法示初评和复评所有条目|t|>3、P<0.01;答案分布分析法示各条目不同水平答案选择率小于80%。信度分析中,内部一致性检验示初评克伦巴赫α系数为0.803,复评系数为0.708,内部一致性较好;评分者间信度示初评Spearman相关系数0.968,复评Spearman相关系数为0.989(P<0.01)。效度分析中,相关系数法示量表条目相关系数均在0.4以上;专家评分法示初、复评条目内容效度指数均大于0.78,全体一致量表内容效度指数为0.83、平均量表内容效度指数为0.97,Kappa均大于0.74;探索性因子分析示结构效度初评Kaiser-Meyer-Olkin(KMO)值为0.846,复评KMO值为0.843,差异有统计学意义(P<0.01),提取特征值大于1的因子初评和复评各有1个,分别可解释总变异的54.221%和53.403%;验证性因子分析示结构效度初、复评各条目P<0.01,且标准载荷系数值大于0.5;受试者操作特征曲线下面积示已知群效度初评FSS对短期结局具有较好的预测区分能力;复评量表中神志、运动、感知、交流条目与格塞尔发育量表大运动及精细动作评分有较好相关性。初评总得分与复评总得分反应度良好。结论:FSS量表汉化内容可靠、信效度稳定性好,条目简单,易于推广。

Objective:To develop a Chinese version of Functional Status Scale(FSS)and to test its reliability and validity in very low birth weight infants(VLBWIs).Methods:The FSS was translated into Chinese and the content was modified in accordance with relevant guidelines and specifications.The Chinese version of FSS was applied to evaluate VLBWIs admitted from January 2018 to June 2020 at 7th day after birth and 34 weeks of postmenstrual age,respectively.The scores were analyzed by descriptive statistics(coefficient of variation method,critical ratio method and answer distribution analysis method),and the reliability and the validity were analyzed.The internal consistency reliability was analyzed using Cronbach’sαcoefficient,the inter rater reliability was analyzed using Spearman correlation coefficient.The content validity was analyzed using correlation coefficient method and expert scoring method;the structural validity was analyzed using exploratory and confirmatory factor analysis method;the known group validity was analyzed using area under the curve(AUC)value and Pearson correlation coefficient.The preliminary response of the initial and reevaluation scales was calculated.Results:After screening by inclusion and exclusion criteria,548 and 523 VLBWIs were included for initial evaluation and re-evaluation,respectively.Descriptive statistics showed that the mean was close to the median,the maximum and minimum values were close to or equal to the values at both ends,and the coefficient of variation was>0.15.The critical ratio method showed that the|t|value of all items in the initial evaluation and re-evaluation was>3(P<0.01).The answer distribution analysis method showed that the answer selection rate of different levels of each item was<80%.Internal consistency test showed that the general Cronbach’sαwas 0.803 and the re-evaluation Cronbach’sαwas 0.708,with a good internal consistency.According to the inter-rater reliability,the Spearman correlation coefficient was 0.968 in the initial evaluation and 0.989 in the re-evaluation(P<0.01).The correlation coefficient of the items in the scale by the correlation coefficient method was more than 0.4.The item-level content validity index(I-CVI)was greater than 0.78,universal agreement of scale of content validity index was 0.83,the average of scale of content validity index was 0.97 and the Kappa was greater than 0.74.Exploratory factor analysis showed that the initial Kaiser-Meyer-Olkin(KMO)value was 0.846,the re-evaluated KMO value was 0.843(P<0.01).There was one factor with extracted eigenvalue>1,which could explain 54.221%and 53.403%of the total variation respectively,suggesting that there was a common factor in the initial evaluation and re-evaluation scales,which was consistent with the original scale design.Confirmatory factor analysis showed that the items in the initial and re-evaluation were significant(P<0.01),and the value of standard load coefficient was greater than 0.5.The known group validity showed that FSS had a good predictive and discriminative ability for short-term outcomes.The items of mental status,motor function,sensory and communication in the re-evaluation scale had a good correlation with gross motor and fine motor energy areas in Gesell developmental schedule.The Pearson correlation coefficient between initial evaluation and reevaluation was 0.609(P<0.01).Conclusion:The Chinese version FSS scale has good reliability and validity,the included items are simple and easy to be applied in clinical practice.