Identifying the odds ratio estimated by a two-stage instrumental variable analysis with a logistic regression model

被引:53
作者
Burgess, Stephen [1 ]
机构
[1] Strangeways Res Lab, Dept Publ Hlth & Primary Care, Cambridge CB1 8RN, England
基金
英国惠康基金;
关键词
instrumental variables; noncollapsibility; Mendelian randomization; logistic regression; MENDELIAN RANDOMIZATION; PROPENSITY SCORE; CLINICAL-TRIALS; COLLAPSIBILITY; RISK; ENDOGENEITY; DIAGRAMS; ADJUST; BIAS;
D O I
10.1002/sim.5871
中图分类号
Q [生物科学];
学科分类号
07 ; 0710 ; 09 ;
摘要
An adjustment for an uncorrelated covariate in a logistic regression changes the true value of an odds ratio for a unit increase in a risk factor. Even when there is no variation due to covariates, the odds ratio for a unit increase in a risk factor also depends on the distribution of the risk factor. We can use an instrumental variable to consistently estimate a causal effect in the presence of arbitrary confounding. With a logistic outcome model, we show that the simple ratio or two-stage instrumental variable estimate is consistent for the odds ratio of an increase in the population distribution of the risk factor equal to the change due to a unit increase in the instrument divided by the average change in the risk factor due to the increase in the instrument. This odds ratio is conditional within the strata of the instrumental variable, but marginal across all other covariates, and is averaged across the population distribution of the risk factor. Where the proportion of variance in the risk factor explained by the instrument is small, this is similar to the odds ratio from a RCT without adjustment for any covariates, where the intervention corresponds to the effect of a change in the population distribution of the risk factor. This implies that the ratio or two-stage instrumental variable method is not biased, as has been suggested, but estimates a different quantity to the conditional odds ratio from an adjusted multiple regression, a quantity that has arguably more relevance to an epidemiologist or a policy maker, especially in the context of Mendelian randomization. Copyright (c) 2013 John Wiley & Sons, Ltd.
引用
收藏
页码:4726 / 4747
页数:22
相关论文
共 56 条
[1]  
Angrist JD, 1996, J AM STAT ASSOC, V91, P444, DOI 10.2307/2291629
[2]  
Angrist JD, 2009, MOSTLY HARMLESS ECONOMETRICS: AN EMPIRICISTS COMPANION, P1
[3]   The interpretation of instrumental variables estimators in simultaneous equations models with an application to the demand for fish [J].
Angrist, JD ;
Graddy, K ;
Imbens, GW .
REVIEW OF ECONOMIC STUDIES, 2000, 67 (03) :499-527
[4]  
[Anonymous], 2001, Econometric Analysis of Cross Section and Panel Data
[5]   Usefulness of Mendelian Randomization in Observational Epidemiology [J].
Bochud, Murielle ;
Rousson, Valentin .
INTERNATIONAL JOURNAL OF ENVIRONMENTAL RESEARCH AND PUBLIC HEALTH, 2010, 7 (03) :711-728
[6]   Mendelian randomization analysis of case-control data using structural mean models [J].
Bowden, Jack ;
Vansteelandt, Stijn .
STATISTICS IN MEDICINE, 2011, 30 (06) :678-694
[7]  
Bowden R. J., 1990, INSTRUMENTAL VARIABL
[8]   Use of Mendelian randomisation to assess potential benefit of clinical intervention [J].
Burgess, Stephen ;
Butterworth, Adam ;
Malarstig, Anders ;
Thompson, Simon G. .
BMJ-BRITISH MEDICAL JOURNAL, 2012, 345
[9]   Two-stage instrumental variable methods for estimating the causal odds ratio: Analysis of bias [J].
Cai, Bing ;
Small, Dylan S. ;
Ten Have, Thomas R. .
STATISTICS IN MEDICINE, 2011, 30 (15) :1809-1824
[10]  
Clarke P., 2010, INSTRUMENTAL VARIABL