Imputation of incident events in longitudinal cohort studies.

Department of Biostatistics, School of Public Health, University of Alabama at Birmingham, Ryals Building, Room 327D, 1665 University Boulevard, Birmingham, AL 35294-0022, USA.
American journal of epidemiology (Impact Factor: 4.98). 07/2011; 174(6):718-26. DOI: 10.1093/aje/kwr155
Source: PubMed

ABSTRACT Longitudinal cohort studies normally identify and adjudicate incident events detected during follow-up by retrieving medical records. There are several reasons why the adjudication process may not be successfully completed for a suspected event including the inability to retrieve medical records from hospitals and an insufficient time between the suspected event and data analysis. These "incomplete adjudications" are normally assumed not to be events, an approach which may be associated with loss of precision and introduction of bias. In this article, the authors evaluate the use of multiple imputation methods designed to include incomplete adjudications in analysis. Using data from the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study, 2008-2009, they demonstrate that this approach may increase precision and reduce bias in estimates of the relations between risk factors and incident events.

1 Follower
  • [Show abstract] [Hide abstract]
    ABSTRACT: The most well-known stroke risk score is the Framingham Stroke Risk Score (FSRS), which was developed during the higher stroke risk period of the 1990s and has not been validated for blacks. We assessed the performance of the FSRS among participants in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study to determine whether it is useful in both blacks and whites. Expected annualized stroke rates from the FSRS were compared with observed stroke rates overall and within strata defined by FSRS risk factors (age, sex, systolic blood pressure, use of antihypertensive medications, diabetes mellitus, smoking, atrial fibrillation, left ventricular hypertrophy, and prevalent coronary heart disease). Among 27 748 participants stroke-free at baseline, 715 stroke events occurred over 5.6 years of follow-up. FSRS-estimated incidence rates of stroke were 1.6× higher than observed for black men, 1.9× higher for white men, 1.7× higher for black women, and 1.7× higher for white women. This overestimation was consistent among most subgroups of FSRS factors, although the magnitude of overestimation varied by the risk factor assessed. Although higher FSRS was associated with higher stroke risk, the FSRS overestimated the observed stroke rates in this study, particularly in certain subgroups. This may be because of temporal declines in stroke rates, secular trends in prevention treatments, or differences in populations studied. More accurate estimates of event rates are critical for planning research, including clinical trials, and targeting health-care efforts.
    Stroke 04/2014; 45(6). DOI:10.1161/STROKEAHA.114.004915 · 6.02 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Most implementations of multiple imputation (MI) of missing data are designed for simple rectangular data structures ignoring temporal ordering of data. Therefore, when applying MI to longitudinal data with intermittent patterns of missing data, some alternative strategies must be considered. One approach is to divide data into time blocks and implement MI independently at each block. An alternative approach is to include all time blocks in the same MI model. With increasing numbers of time blocks, this approach is likely to break down because of co-linearity and over-fitting. The new two-fold fully conditional specification (FCS) MI algorithm addresses these issues, by only conditioning on measurements, which are local in time. We describe and report the results of a novel simulation study to critically evaluate the two-fold FCS algorithm and its suitability for imputation of longitudinal electronic health records. After generating a full data set, approximately 70% of selected continuous and categorical variables were made missing completely at random in each of ten time blocks. Subsequently, we applied a simple time-to-event model. We compared efficiency of estimated coefficients from a complete records analysis, MI of data in the baseline time block and the two-fold FCS algorithm. The results show that the two-fold FCS algorithm maximises the use of data available, with the gain relative to baseline MI depending on the strength of correlations within and between variables. Using this approach also increases plausibility of the missing at random assumption by using repeated measures over time of variables whose baseline values may be missing. © 2014 The Authors. Statistics in Medicine published by John Wiley & Sons, Ltd.
    Statistics in Medicine 09/2014; 33(21). DOI:10.1002/sim.6184 · 2.04 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Background and Purpose-Insulin resistance is associated with increased stroke risk, but the effect has not been adequately examined separately in white and black populations. Methods-The association of baseline insulin resistance with risk of cerebral infarction (CI) and intracerebral hemorrhage (ICH) was assessed in 12 366 white and 6782 black participants from the REasons for Geographic And Racial Differences in Stroke (REGARDS) cohort, recruited between 2003 and 2007 and followed for an average of 5.7 years. Insulin resistance was measured with the homeostasis model assessment-insulin resistance. Results-There were 364 incident CI and 41 incident ICH events. The risk for CI increased with the log of insulin resistance in whites (hazards ratio [HR](ln(IR))=1.17; 95% confidence interval [CI], 1.00-1.38) but was largely attenuated by adjustment for stroke risk factors (HRln(IR)=1.05; 95% CI, 0.88-1.26). There was no association in blacks (HRln(IR)=1.01; 95% CI, 0.81-1.25). After adjustment for demographic factors and risk factors, there was a significant difference by race in the association of insulin resistance with risk of ICH (P=0.07), with a decrease in the risk of ICH in whites (HRln(IR)=0.61; 95% CI, 0.35-1.04) but a nonsignificant increase in blacks (HRln(IR)=1.20; 95% CI, 0.60-2.39). Conclusions-These data support the growing evidence that insulin resistance may play a more important role in stroke risk among white than black individuals and suggest a potentially discordant relationship of insulin resistance on CI and ICH among whites.
    Stroke 06/2014; 45(8). DOI:10.1161/STROKEAHA.114.005306 · 6.02 Impact Factor