Missing phenotype data can be a major hurdle to mapping quantitative trait loci (QTL). Though in many cases experiments may be designed to minimize the occurrence of missing data, it is often unavoidable in practice; thus, statistical methods to account for missing data are needed. In this paper we describe an approach for conjoining multiple imputation and QTL mapping. Methods are applied to map genes associated with increased breathing effort in mice after lung inflammation due to allergen challenge in developing lines of the Collaborative Cross, a new mouse genetics resource. Missing data poses a particular challenge in this study because the desired phenotype summary to be mapped is a function of incompletely observed dose-response curves. Comparison of the multiple imputation approach to two naive approaches for handling missing data suggest that these simpler methods may yield poor results: ignoring missing data through a complete case analysis may lead to incorrect conclusions, while using a last observation carried forward procedure, which does not account for uncertainty in the imputed values, may lead to anti-conservative inference. The proposed approach is widely applicable to other studies with missing phenotype data.
multiple imputation; missing data; quantitative trait loci
We focus on estimation of the causal effect of treatment on the functional status of individuals at a fixed point in time t* after they have experienced a catastrophic event, from observational data with the following features: (1) treatment is imposed shortly after the event and is non-randomized, (2) individuals who survive to t* are scheduled to be interviewed, (3) there is interview non-response, (4) individuals who die prior to t* are missing information on pre-event confounders, (5) medical records are abstracted on all individuals to obtain information on post-event, pre-treatment confounding factors. To address the issue of survivor bias, we seek to estimate the survivor average causal effect (SACE), the effect of treatment on functional status among the cohort of individuals who would survive to t* regardless of whether or not assigned to treatment. To estimate this effect from observational data, we need to impose untestable assumptions, which depend on the collection of all confounding factors. Since pre-event information is missing on those who die prior to t*, it is unlikely that these data are missing at random (MAR). We introduce a sensitivity analysis methodology to evaluate the robustness of SACE inferences to deviations from the MAR assumption. We apply our methodology to the evaluation of the effect of trauma center care on vitality outcomes using data from the National Study on Costs and Outcomes of Trauma Care.
We consider the problem of comparing cumulative incidence functions of non-mortality events in the presence of informative coarsening and the competing risk of death. We extend frequentist-based hypothesis tests previously developed for non-informative coarsening and propose a novel Bayesian method based on comparing a posterior parameter transformation to its expected distribution under the null hypothesis of equal cumulative incidence functions. Both methods use estimates derived by extending previously published estimation procedures to accommodate censoring by death. The data structure and analysis goal are exemplified by the AIDS Link to the Intravenous Experience (ALIVE) study, where researchers are interested in comparing incidence of human immunodeficiency virus seroconversion by risk behavior categories. Coarsening in the forms of interval and right censoring and censoring by death in ALIVE are thought to be informative, thus we perform a sensitivity analysis by incorporating elicited expert information about the relationship between seroconversion and censoring into the model.
Bayesian Analysis; Frequentist Analysis; Hypothesis Test; Interval Censoring; Markov Chain Monte Carlo; Sensitivity Analysis
In this paper, the authors use the rubric of “coarsened data,” of which missing and censored data are special cases, to motivate the elicitation and use of expert information for performing sensitivity analyses of censored event-time data. Elicited information is important because observed data are insufficient to estimate how study participants with coarsened data compare with participants with uncoarsened data, and misspecifying this comparison may produce biased analysis results. In the presence of coarsening, performing a sensitivity analysis over a range of plausible assumptions is the best one can do. Here the authors illustrate an approach for eliciting expert information for use in sensitivity analyses to compare cumulative incidence functions of censored nonmortality outcomes. An example of such data is the AIDS Link to Intravenous Experience (ALIVE) Study, where the authors aim to estimate and compare cumulative incidence functions for human immunodeficiency virus between risk factor categories. The interval and right-censoring and censoring due to death found in the ALIVE data (1988–1998) are thought to be informative; thus, a sensitivity analysis is performed using information elicited from 2 ALIVE scientists and an expert in acquired immunodeficiency syndrome epidemiology about the relation between seroconversion and censoring.
Bayesian analysis; frequentist approach; HIV; hypothesis test; incidence; interval censoring; sensitivity analysis
Using validation sets for outcomes can greatly improve the estimation of vaccine efficacy (VE) in the field (Halloran and Longini, 2001; Halloran and others, 2003). Most statistical methods for using validation sets rely on the assumption that outcomes on those with no cultures are missing at random (MAR). However, often the validation sets will not be chosen at random. For example, confirmational cultures are often done on people with influenza-like illness as part of routine influenza surveillance. VE estimates based on such non-MAR validation sets could be biased. Here we propose frequentist and Bayesian approaches for estimating VE in the presence of validation bias. Our work builds on the ideas of Rotnitzky and others (1998, 2001), Scharfstein and others (1999, 2003), and Robins and others (2000). Our methods require expert opinion about the nature of the validation selection bias. In a re-analysis of an influenza vaccine study, we found, using the beliefs of a flu expert, that within any plausible range of selection bias the VE estimate based on the validation sets is much higher than the point estimate using just the non-specific case definition. Our approach is generally applicable to studies with missing binary outcomes with categorical covariates.
Bayesian; Expert opinion; Identifiability; Influenza; Missing data; Selection model; Vaccine efficacy
In randomized studies with missing outcomes, non-identifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying asssumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment comparisons to missing data assumptions, it may be dissatisfying to some researchers/decision makers because a single summary is not provided. In this paper, we present a fully Bayesian methodology that allows the investigator to draw a ‘single’ conclusion by formally incorporating prior beliefs about non-identifiable, yet interpretable, selection bias parameters. Our Bayesian model provides robustness to prior specification of the distributional form of the continuous outcomes.
Dirichlet process prior; Identifiability; MCHC; Non-parametric Bayes; Selection model; Sensitivity analysis
The etiology of bacterial vaginosis is unknown, and there are no long-term therapies for preventing this frequently recurring condition. Vaginal douching has been reported to be associated with bacterial vaginosis in observational studies. However, this association may be due to confounding by indication—that is, confounding by women douching in response to vaginal symptoms associated with bacterial vaginosis. The authors used marginal structural modeling to estimate the causal effect of douching on bacterial vaginosis risk while controlling for this confounding effect. In 1999–2002, nonpregnant women (n = 3,620) were recruited into a prospective study when they visited one of 12 public health clinics in Birmingham, Alabama, for routine care. Participants were assessed quarterly for 1 year. Bacterial vaginosis was based on a Nugent's Gram stain score of 7 or higher. Thirty-two percent of participants douched in every study interval, and 43.0% never douched. Of the 12,349 study visits, 40.2% were classified as involving bacterial vaginosis. The relative risk for regular douching as compared with no douching was 1.21 (95% confidence interval: 1.08, 1.38). These findings indicate that douching confers increased risk of disruption of vaginal flora. In the absence of a large randomized trial, these findings provide the best evidence to date for a risk of bacterial vaginosis associated with douching.
confounding factors (epidemiology); epidemiologic methods; longitudinal studies; models, structural; vaginal douching; vaginosis, bacterial
To evaluate the risk for bacterial vaginosis (BV) in a douching cessation trial.
Thirty-nine reproductive-age women who reported use of douche products were enrolled into a 20-week study consisting of a 4-week douching observation (phase I) followed by 12-weeks of douching cessation (phase II). In phase III, participants then chose to resume douching or continue cessation for the remaining 4 weeks. Self-collected vaginal samples were obtained twice-weekly in the first 16 weeks and one sample was collected during week 20 (1,107 samples total). BV was diagnosed by Nugent score ≥7. Conditional logistic regression was used to evaluate douching cessation on the risk of BV.
The adjusted odds ratio (aOR) for BV in the douching cessation phase as compared to the douching observation phase was 0.76 (95% CI:0.33–1.76). Among women who reported their primary reason for douching was to cleanse after menstruation, BV was significantly reduced in douching cessation (aOR:0.23; 95% CI:0.12–0.44).
Vaginal douching cessation may reduce the risk for BV in a subset of women.
vaginal douching; bacterial vaginosis; intravaginal cleansing; intravaginal washing
The etiology of bacterial vaginosis (BV) is unknown and there are no long-term therapies for preventing this frequently recurring condition. Vaginal douching has been reported to be associated with BV in observational studies. However, this association may be due to confounding by indication-- women douching in response to vaginal symptoms associated with BV. Marginal structural modeling was used to estimate the causal effect of douching on risk for BV controlling for this confounding effect. From 1999-2002, non-pregnant women (n=3620) were recruited into a prospective study when presenting for routine care at 12 public health clinics in Birmingham, Alabama. Participants were assessed quarterly for one year. BV was based on Nugent’s Gram stain score ≥ 7. Thirty-two percent of participants douched in every interval and 43.0% never douched. Of the 12,349 visits, 40.2% were classified as having BV. The relative risk for regular douching practice compared with no douching practice was 1.21 (95% CI: 1.08, 1.38). Our findings indicate that douching confers an increased risk for disruption of vaginal flora. In the absence of a large randomized trial, these findings provide the best evidence to date for the risk of douching on BV.
Vaginosis, Bacterial; Vaginal Douching; epidemiologic methods; Longitudinal Studies; Models, Structural; Confounding Factor; Epidemiologic; time-dependent confounding