中医临床队列研究中缺失数据的分析处理
|
摘要
中医诊疗以个体化辨证论治和复杂干预为特点,接近真实世界研究的队列设计更能充分显示中医诊疗优势,但由于队列研究存在样本量大、研究周期长、开放度高、允许偏倚等问题,直接或间接破坏了数据完整性。传统以删失、均值替代为主的数据处理方法容易造成样本浪费,降低了统计效能,而以多重填补和最大似然思想为基础的数据缺失处理方法较多又难以选择,且方法间存在误用的问题。若能按照数据缺失判定、方法选用及拟合、拟合效果评价三个步骤,并根据缺失数据的属性仔细辨别方法间的区别和使用条件,构建中医临床队列研究的缺失数据拟合处理路径,将有效提高数据质量,使拟合后数据接近患者治疗的真实世界,更确切和真实地反映中医药临床疗效。
Traditional Chinese medicine( TCM) diagnosis and treatment is characterized by individualized differentiation of symptoms and signs as well as complex intervention. The cohort study close to the real world research can fully highlight the advantages of TCM diagnosis and treatment. However,data integrity is directly or indirectly damaged due to the large sample size,long research cycle,high openness and allowable bias of cohort research. Traditional data processing methods based on deletion and mean substitution are prone to waste samples and reduce statistical efficiency,while data missing processing methods based on multi-filling and maximum likelihood are more and difficult to choose,and there are problems of misuse among methods. If the missing data fitting processing path can be constructed according to the three steps of data missing determination,method selection,fitting and fitting effect evaluation,and according to the attributes of missing data carefully distinguishing the differences between methods and using conditions,the data quality will be effectively improved,and the data after fitting will be close to patient treatment,which can more accurately and truly reflects the clinical efficacy of Chinese medicine.
Traditional Chinese medicine( TCM) diagnosis and treatment is characterized by individualized differentiation of symptoms and signs as well as complex intervention. The cohort study close to the real world research can fully highlight the advantages of TCM diagnosis and treatment. However,data integrity is directly or indirectly damaged due to the large sample size,long research cycle,high openness and allowable bias of cohort research. Traditional data processing methods based on deletion and mean substitution are prone to waste samples and reduce statistical efficiency,while data missing processing methods based on multi-filling and maximum likelihood are more and difficult to choose,and there are problems of misuse among methods. If the missing data fitting processing path can be constructed according to the three steps of data missing determination,method selection,fitting and fitting effect evaluation,and according to the attributes of missing data carefully distinguishing the differences between methods and using conditions,the data quality will be effectively improved,and the data after fitting will be close to patient treatment,which can more accurately and truly reflects the clinical efficacy of Chinese medicine.
引文
[1]刘保延.有关辨证论治临床评价若干问题的思考[J].中医杂志,2007,48(1):12-14.
[2]盛凤,蒋健,郑鑫,等.中医辨证论治及个体化治疗的临床研究方法[J].中华中医药杂志,2011,26(1):115-118.
[3]江丽杰,谢琪,刘保延.效果比较研究概述[J].世界中医药,2013,8(6):695-700.
[4]MARKO NF,WEIL RJ. An introduction to comparative effectiveness research[J]. Neuros,2012,70(2):425-434.
[5]CRUTZEN R,VIECHTBAUER W,KOTZ D,et al. No differential attrition was found in randomized controlled trials published in general medical journals:a meta-analysis[J]. J Clin Epidemiol,2013,66(9):948-954.
[6]FLYER P,HIRMAN J. Missing data in confirmatory clinical trials[J]. J Biophar Stat,2009,19(6):969-979.
[7]刘建平.队列研究的设计、实施及方法学问题[J].中西医结合学报,2008,6(4):331-336.
[8]ELM EV,ALTMAN DG,EGGER M,et al. The strengthening the reporting of observational studies in epidemiology,(STROBE)statement:guidelines for reporting observational studies[J]. Int J Surg,2014,18(6):1453-1457.
[9]赵俊康,王彤,荣惠英,等.不同缺失机制并存时偏倚校正的模拟研究[J].中国卫生统计,2014,31(4):570-574,579.
[10]NETTEN AP,DEKKER FW,RIEFFE C,et al. Missing data in the field of otorhinolaryngology and head&neck surgery:need for improvement[J]. Ear Hear,2017,38(1):1-6.
[11]KARAHALIOS A,BAGLIETTO L,CARLIN JB,et al.A review of the reporting and handling of missing data in cohort studies with repeated assessment of exposure measures[J]. BMC Med Res Methodol,2012,12:96. doi:10. 1186/1471-2288-12-96.
[12]The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Statistical principals for clinical trials[EB/OL].(1998-02-05)[2016-03-10]. http://www. ICH. org/filesdmin/Public_web_site/ICH. Products/Guidelines/Efficcy/E9/Step4/E9_Guideline.
[13]LITTLE RJA,RUBIN DB. Statistical Analysis with Missing Data,Second Edition[M]. New York:Wiley,2002:200-220.
[14]KENWARD MG,MOLENBERGHS G. Last observation carried forward:a crystal ball?[J]. J Biophar Stat,2009,19(5):872-888.
[15]邹莉玲,吴娟丽,李觉.多重填补法在任意缺失随访资料中的应用[J].中国卫生统计,2015,32(2):221-223.
[16]SCHAFER J,GRAHAM J. Missing data:our view of the state of the art[J]. Psychol Methods,2002,7(2):147-177.
[17]LITTLE JA,RUBIN DB,et al. Statistical analysis with missing data[J]. Technometrics,1987,30(4):455-455.
[18]KARAHALIOS A,BAGLIETTO L,LEE KJ,et al. The impact of missing data on analyses of atime-dependent exposure in a longitudinal cohort:a simulation study[J].Emerg Themes Epidemiol,2013, 10(1):6. doi:10. 1186/1742-7622-10-6.
[19]JELICIC H,PHELPS E,LERNER RM. Use of missing data methods in longitudinal studies:The persistence of bad practices in developmental psychology[J]. Dev Psychol,2009,45(4):1195-1199.
[20]王梅.中医药疗效评价队列研究方法学探索[D].沈阳:辽宁中医药大学,2016.
[21]彭倩倩.遗传流行病学纵向分析中的若干统计问题:数据分析与降维[D].上海:复旦大学,2012.
[22]梁公文,王丹蕾,王兵,等.医学研究可持续发展支撑平台的建设与运行机制研究[J].中国医院管理,2013,33(2):66-67.
[2]盛凤,蒋健,郑鑫,等.中医辨证论治及个体化治疗的临床研究方法[J].中华中医药杂志,2011,26(1):115-118.
[3]江丽杰,谢琪,刘保延.效果比较研究概述[J].世界中医药,2013,8(6):695-700.
[4]MARKO NF,WEIL RJ. An introduction to comparative effectiveness research[J]. Neuros,2012,70(2):425-434.
[5]CRUTZEN R,VIECHTBAUER W,KOTZ D,et al. No differential attrition was found in randomized controlled trials published in general medical journals:a meta-analysis[J]. J Clin Epidemiol,2013,66(9):948-954.
[6]FLYER P,HIRMAN J. Missing data in confirmatory clinical trials[J]. J Biophar Stat,2009,19(6):969-979.
[7]刘建平.队列研究的设计、实施及方法学问题[J].中西医结合学报,2008,6(4):331-336.
[8]ELM EV,ALTMAN DG,EGGER M,et al. The strengthening the reporting of observational studies in epidemiology,(STROBE)statement:guidelines for reporting observational studies[J]. Int J Surg,2014,18(6):1453-1457.
[9]赵俊康,王彤,荣惠英,等.不同缺失机制并存时偏倚校正的模拟研究[J].中国卫生统计,2014,31(4):570-574,579.
[10]NETTEN AP,DEKKER FW,RIEFFE C,et al. Missing data in the field of otorhinolaryngology and head&neck surgery:need for improvement[J]. Ear Hear,2017,38(1):1-6.
[11]KARAHALIOS A,BAGLIETTO L,CARLIN JB,et al.A review of the reporting and handling of missing data in cohort studies with repeated assessment of exposure measures[J]. BMC Med Res Methodol,2012,12:96. doi:10. 1186/1471-2288-12-96.
[12]The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use. Statistical principals for clinical trials[EB/OL].(1998-02-05)[2016-03-10]. http://www. ICH. org/filesdmin/Public_web_site/ICH. Products/Guidelines/Efficcy/E9/Step4/E9_Guideline.
[13]LITTLE RJA,RUBIN DB. Statistical Analysis with Missing Data,Second Edition[M]. New York:Wiley,2002:200-220.
[14]KENWARD MG,MOLENBERGHS G. Last observation carried forward:a crystal ball?[J]. J Biophar Stat,2009,19(5):872-888.
[15]邹莉玲,吴娟丽,李觉.多重填补法在任意缺失随访资料中的应用[J].中国卫生统计,2015,32(2):221-223.
[16]SCHAFER J,GRAHAM J. Missing data:our view of the state of the art[J]. Psychol Methods,2002,7(2):147-177.
[17]LITTLE JA,RUBIN DB,et al. Statistical analysis with missing data[J]. Technometrics,1987,30(4):455-455.
[18]KARAHALIOS A,BAGLIETTO L,LEE KJ,et al. The impact of missing data on analyses of atime-dependent exposure in a longitudinal cohort:a simulation study[J].Emerg Themes Epidemiol,2013, 10(1):6. doi:10. 1186/1742-7622-10-6.
[19]JELICIC H,PHELPS E,LERNER RM. Use of missing data methods in longitudinal studies:The persistence of bad practices in developmental psychology[J]. Dev Psychol,2009,45(4):1195-1199.
[20]王梅.中医药疗效评价队列研究方法学探索[D].沈阳:辽宁中医药大学,2016.
[21]彭倩倩.遗传流行病学纵向分析中的若干统计问题:数据分析与降维[D].上海:复旦大学,2012.
[22]梁公文,王丹蕾,王兵,等.医学研究可持续发展支撑平台的建设与运行机制研究[J].中国医院管理,2013,33(2):66-67.