Applications of multiple imputation in medical studies: from AIDS to NHANES

Abstract

Rubin's multiple imputation is a three-step method for handling complex missing data, or more generally, incomplete-data problems, which arise frequently in medical studies. At the first step, m (> 1) completed-data sets are created by imputing the unobserved data m times using m independent draws from an imputation model, which is constructed to reasonably approximate the true distributional relationship between the unobserved data and the available information, and thus reduce potentially very serious nonresponse bias due to systematic difference between the observed data and the unobserved ones. At the second step, m complete-data analyses are performed by treating each completed-data set as a real complete-data set, and thus standard complete-data procedures and software can be utilized directly. At the third step, the results from the m complete-data analyses are combined in a simple, appropriate way to obtain the so-called repeated-imputation inference, which properly takes into account the uncertainty in the imputed values. This paper reviews three applications of Rubin's method that are directly relevant for medical studies. The first is about estimating the reporting delay in acquired immune deficiency syndrome (AIDS) surveillance systems for the purpose of estimating survival time after AIDS diagnosis. The second focuses on the issue of missing data and noncompliance in randomized experiments, where a school choice experiment is used as an illustration. The third looks at handling nonresponse in United States National Health and Nutrition Examination Surveys (NHANES). The emphasis of our review is on the building of imputation models (i.e. the first step), which is the most fundamental aspect of the method.

The effect of preceptor role effectiveness on newly licensed registered nurses' perceived psychological empowerment and professional autonomy

“…Furthermore, Raymond and Roberts (1987) estimated that a variable should be retained with 40% or less missing data. Removing cases with missing data from data analysis procedures can result in reduced sample size, compromised statistical power, and inaccurate parameter estimates (Barnard & Meng, 1999;Patrician, 2002;Tabachnick & Fidell, 2001). Due to the small percentage of missing data, the decision was made to retain all cases and use individual mean imputation for missing data points (Shrive, Stuart, Quan, & Ghali, 2006).…”

Relation of Adult-Onset Asthma to Coronary Heart Disease and Stroke

“…Missing covariate values, which occurred at <3% for any covariate, were imputed using multiple imputation methods. 18 Crude incidence density rates of CHD and stroke were calculated among men and women for non-asthmatics, child-onset asthmatics, and adult-onset asthmatics. Crude and multivariate hazard ratios comparing each asthma subtype to non-asthmatics were computed using Cox proportional hazards models.…”

Twenty-five years of the AIDS epidemic in Brazil: principal epidemiological findings, 1980-2005

“…The application of more sophisticated statistical techniques such as: the imputation of missing data 71 ; the utilization of discriminating analysis for the purpose of defining contrasting profiles of AIDS cases, and, afterwards, reclassification of missing data 11 ; and the utilization of Baysean estimators in procedures using back-calculation for incubation periods that violate the presupposition of stationarity 72 , to give examples of some of these alternatives.…”