Asymptotically Unbiased Estimation of Exposure Odds Ratios in Complete Records Logistic Regression

被引:105
作者
Bartlett, Jonathan W. [1 ]
Harel, Ofer [2 ]
Carpenter, James R. [1 ,3 ]
机构
[1] London Sch Hyg & Trop Med, Dept Med Stat, London WC1E 7HT, England
[2] Univ Connecticut, Dept Stat, Coll Liberal Arts & Sci, Storrs, CT 06269 USA
[3] UCL, MRC, Clin Trials Unit, London, England
基金
英国医学研究理事会; 美国国家卫生研究院;
关键词
complete case analysis; logistic regression; missing data; odds ratio; MISSING DATA; MULTIPLE IMPUTATION; EFFICIENCY; MODEL; BIAS;
D O I
10.1093/aje/kwv114
中图分类号
R1 [预防医学、卫生学];
学科分类号
1004 ; 120402 ;
摘要
Missing data are a commonly occurring threat to the validity and efficiency of epidemiologic studies. Perhaps the most common approach to handling missing data is to simply drop those records with 1 or more missing values, in so-called "complete records" or "complete case" analysis. In this paper, we bring together earlier-derived yet perhaps now somewhat neglected results which show that a logistic regression complete records analysis can provide asymptotically unbiased estimates of the association of an exposure of interest with an outcome, adjusted for a number of confounders, under a surprisingly wide range of missing-data assumptions. We give detailed guidance describing how the observed data can be used to judge the plausibility of these assumptions. The results mean that in large epidemiologic studies which are affected by missing data and analyzed by logistic regression, exposure associations may be estimated without bias in a number of settings where researchers might otherwise assume that bias would occur.
引用
收藏
页码:730 / 736
页数:7
相关论文
共 24 条
[1]   Improving upon the efficiency of complete case analysis when covariates are MNAR [J].
Bartlett, Jonathan W. ;
Carpenter, James R. ;
Tilling, Kate ;
Vansteelandt, Stijn .
BIOSTATISTICS, 2014, 15 (04) :719-730
[2]  
Carpenter JamesR., 2013, MULTIPLE IMPUTATION
[3]  
Carpenter JR, 2013, MULTIPLE IMPUTATION, P229
[4]  
Carpenter JR., 2014, HDB MISSING DATA, P435
[5]   Using causal diagrams to guide analysis in missing data problems [J].
Daniel, Rhian M. ;
Kenward, Michael G. ;
Cousens, Simon N. ;
De Stavola, Bianca L. .
STATISTICAL METHODS IN MEDICAL RESEARCH, 2012, 21 (03) :243-256
[6]   Cause-specific mortality in professional flight crew and air traffic control officers: findings from two UK population-based cohorts of over 20,000 subjects [J].
De Stavola, Bianca L. ;
Pizzi, Costanza ;
Clemens, Felicity ;
Evans, Sally Ann ;
Evans, Anthony D. ;
Silva, Isabel dos Santos .
INTERNATIONAL ARCHIVES OF OCCUPATIONAL AND ENVIRONMENTAL HEALTH, 2012, 85 (03) :283-293
[7]   Missing Data A Systematic Review of How They Are Reported and Handled [J].
Eekhout, Iris ;
de Boer, Michiel R. ;
Twisk, Jos W. R. ;
de Vet, Henrica C. W. ;
Heymans, Martijn W. .
EPIDEMIOLOGY, 2012, 23 (05) :729-732
[8]   A COMPARISON OF THE LOGISTIC RISK-FUNCTION AND THE PROPORTIONAL HAZARDS MODEL IN PROSPECTIVE EPIDEMIOLOGIC STUDIES [J].
GREEN, MS ;
SYMONS, MJ .
JOURNAL OF CHRONIC DISEASES, 1983, 36 (10) :715-724
[9]   Multiple imputation: Review of theory, implementation and software [J].
Harel, Ofer ;
Zhou, Xiao-Hua .
STATISTICS IN MEDICINE, 2007, 26 (16) :3057-3077
[10]   A structural approach to selection bias [J].
Hernán, MA ;
Hernández-Díaz, S ;
Robins, JM .
EPIDEMIOLOGY, 2004, 15 (05) :615-625