Suspected survivor bias in case–control studies: stratify on survival time and use a negative control

ObjectivesSelection bias in case–control studies occurs when control selection is inappropriate. However, selection bias due to improper case sampling is less well recognized. We describe how to recognize survivor bias (i.e., selection on exposed cases) and illustrate this with an example study.Study Design and SettingA case–control study was used to analyze the effect of statins on major bleedings during treatment with vitamin K antagonists. A total of 110 patients who experienced such bleedings were included 18–1,018 days after the bleeding complication and matched to 220 controls.ResultsA protective association of major bleeding for exposure to statins (odds ratio [OR]: 0.56; 95% confidence interval: 0.29–1.08) was found, which did not become stronger after adjustment for confounding factors. These observations lead us to suspect survivor bias. To identify this bias, results were stratified on time between bleeding event and inclusion, and repeated for a negative control (an exposure not related to survival): blood group non-O. The ORs for exposure to statins increased gradually to 1.37 with shorter time between outcome and inclusion, whereas ORs for the negative control remained constant, confirming our hypothesis.ConclusionWe recommend the presented method to check for overoptimistic results, that is, survivor bias in case–control studies.     

Phenotype validation in electronic health records based genetic association studies

The linkage between electronic health records (EHRs) and genotype data makes it plausible to study the genetic susceptibility of a wide range of disease phenotypes. Despite that EHR‐derived phenotype data are subjected to misclassification, it has been shown useful for discovering susceptible genes, particularly in the setting of phenome‐wide association studies (PheWAS). It is essential to characterize discovered associations using gold standard phenotype data by chart review. In this work, we propose a genotype stratified case‐control sampling strategy to select subjects for phenotype validation. We develop a closed‐form maximum‐likelihood estimator for the odds ratio parameters and a score statistic for testing genetic association using the combined validated and error‐prone EHR‐derived phenotype data, and assess the extent of power improvement provided by this approach. Compared with case‐control sampling based only on EHR‐derived phenotype data, our genotype stratified strategy maintains nominal type I error rates, and result in higher power for detecting associations. It also corrects the bias in the odds ratio parameter estimates, and reduces the corresponding variance especially when the minor allele frequency is small.     

Self-selection and bias in a large prospective pregnancy cohort in Norway

Self‐selection in epidemiological studies may introduce selection bias and influence the validity of study results. To evaluate potential bias due to self‐selection in a large prospective pregnancy cohort in Norway, the authors studied differences in prevalence estimates and association measures between study participants and all women giving birth in Norway. Women who agreed to participate in the Norwegian Mother and Child Cohort Study (43.5% of invited; n = 73 579) were compared with all women giving birth in Norway (n = 398 849) using data from the population‐based Medical Birth Registry of Norway in 2000–2006. Bias in the prevalence of 23 exposure and outcome variables was measured as the ratio of relative frequencies, whereas bias in exposure‐outcome associations of eight relationships was measured as the ratio of odds ratios. Statistically significant relative differences in prevalence estimates between the cohort participants and the total population were found for all variables, except for maternal epilepsy, chronic hypertension and pre‐eclampsia. There was a strong under‐representation of the youngest women (<25 years), those living alone, mothers with more than two previous births and with previous stillbirths (relative deviation 30–45%). In addition, smokers, women with stillbirths and neonatal death were markedly under‐represented in the cohort (relative deviation 22–43%), while multivitamin and folic acid supplement users were over‐represented (relative deviation 31–43%). Despite this, no statistically relative differences in association measures were found between participants and the total population regarding the eight exposure‐outcome associations. Using data from the Medical Birth Registry of Norway, this study suggests that prevalence estimates of exposures and outcomes, but not estimates of exposure‐outcome associations are biased due to self‐selection in the Norwegian Mother and Child Cohort Study.     

Future cases as present controls to adjust for exposure trend bias in case-only studies

Self-matched case-only studies (such as the case-crossover or self-controlled case-series method) control by design for time-invariant confounders (measured or unmeasured), but they do not control for confounders that vary with time. A bidirectional case-crossover design can be used to adjust for exposure-time trends. In pharmacoepidemiology, however, illness often influences future use of medications, making a bidirectional design problematic. Suissa's case-time-control design combines a case-crossover and case-control design, and adjusts for exposure-trend bias in the cases' self-controlled odds ratio by dividing that ratio by the corresponding self-controlled odds ratio in a concurrent matched control group. However, if not well matched, the control group may reintroduce selection bias. We propose a "case-case-time-control" that involves crossover analyses in cases and future-case controls. This person-time sampling strategy improves matching by restricting controls to future cases. We evaluate the proposed study design through simulations and analysis of a theoretically null relationship using Veterans Administration (VA) data. Simulation studies show that the case-case-time-control can adjust for exposure trends while controlling for time-invariant confounders. Use of an inappropriate control group left case-time-control analyses biased by exposure-time trends. When analyzing the relationship between vitamin exposure and stroke, using data on 3192 patients in the VA system, a case-crossover odds ratio of 1.5 (95% confidence interval = 1.3-1.7) was reduced to 1.1 (0.9-1.3) when divided by the concurrent exposure trend odds ratio (1.4) in matched future cases. This applied example demonstrates how our approach can adjust for exposure trends observed across time axes.     

Correcting “winner's curse” in odds ratios from genomewide association findings for major complex human diseases

Genome‐wide association studies (GWAS) provide an important approach for identifying common genetic variants that predispose to human disease. However, odds ratio (OR) estimates for the reported findings from GWAS discovery data are typically affected by a bias away from the null sometimes referred to the “winner's curse”. Also standard confidence intervals (CIs) may have far from the desired coverage rates. We applied a bias reduction method to GWAS findings from several major complex human diseases, including breast cancer, colorectal cancer, lung cancer, prostate cancer, type I diabetes, and type II diabetes. We found the simple bias correction procedure allows one to estimate bias‐adjusted ORs that have substantial consistency with ORs from subsequent replication studies, and that corresponding selection‐adjusted CIs appear to help quantify the uncertainty of the findings. Selection‐adjusted ORs and CIs can provide a reliable summary of GWAS data, and can help to choose single nucleotide polymorphisms for subsequent validation studies. Genet. Epidemiol. 34:78–91, 2010. © 2009 Wiley‐Liss, Inc.     

Semiparametric odds ratio model for case–control and matched case–control designs

We propose a semiparametric odds ratio model that extends Umbach and Weinberg's approach to exploiting gene–environment association model for efficiency gains in case–control designs to both discrete and continuous data. We directly model the gene–environment association in the control population to avoid estimating the intercept in the disease risk model, which is inherently difficult because of the scarcity of information on the parameter with the sampling designs. We propose a novel permutation‐based approach to eliminate the high‐dimensional nuisance parameters in the matched case–control design. The proposed approach reduces to the conditional logistic regression when the model for the gene–environment association is unrestricted. Simulation studies demonstrate good performance of the proposed approach. We apply the proposed approach to a study of gene–environment interaction on coronary artery disease. Copyright © 2013 John Wiley & Sons, Ltd.     

Gene–environment interactions in case–control studies with silent disease

Genome‐wide association studies (GWAS) often measure gene–environment interactions (G × E). We consider the problem of accurately estimating a G × E in a case–control GWAS when a subset of the controls have silent, or undiagnosed, disease and the frequency of the silent disease varies by the environmental variable. We show that using case–control status without accounting for misdiagnosis can lead to biased estimates of the G × E. We further propose a pseudolikelihood approach to remove the bias and accurately estimate how the relationship between the genetic variant and the true disease status varies by the environmental variable. We demonstrate our method in extensive simulations and apply our method to a GWAS of prostate cancer.     

A Bayesian approach to strengthen inference for case‐control studies with multiple error‐prone exposure assessments

In case‐control studies, exposure assessments are almost always error‐prone. In the absence of a gold standard, two or more assessment approaches are often used to classify people with respect to exposure. Each imperfect assessment tool may lead to misclassification of exposure assignment; the exposure misclassification may be differential with respect to case status or not; and, the errors in exposure classification under the different approaches may be independent (conditional upon the true exposure status) or not. Although methods have been proposed to study diagnostic accuracy in the absence of a gold standard, these methods are infrequently used in case‐control studies to correct exposure misclassification that is simultaneously differential and dependent. In this paper, we proposed a Bayesian method to estimate the measurement‐error corrected exposure‐disease association, accounting for both differential and dependent misclassification. The performance of the proposed method is investigated using simulations, which show that the proposed approach works well, as well as an application to a case‐control study assessing the association between asbestos exposure and mesothelioma. Copyright © 2013 John Wiley & Sons, Ltd.     

A strategy to reduce healthy worker effect in a cross-sectional study of asthma and metalworking fluids

This report describes the reanalysis of a cross-sectional study of asthma in a large cohort of autoworkers with exposure to metalworking fluids (MWF). There is strong evidence from case reports, clinical studies, and medical surveillance data that exposure to MWF can cause asthma, yet no association was found in the original analysis. The central hypothesis of the reanalysis was that the absence of an association between asthma and MWF exposure was the result of bias caused by the self-selection of asthmatics out of exposed jobs. We addressed the potential job transfer bias by redefining exposure and disease status at the time of asthma onset, rather than at the time of the health survey. This permitted us to treat the cross-sectional study as if it were a historical cohort study, despite the fact that the population was a biased sample of the full cohort. This approach resulted in a significantly elevated incidence rate ratio of 3.2 (95% CI: 1.2–8.3) for synthetic MWF estimated in a Cox proportional hazards model. Although the cross-sectional design makes it impossible to document or control for differential selection out of the workforce, the approach described here provides a strategy for reducing the healthy-worker effect due to job transfer bias in cross-sectional studies. Am. J. Ind. Med. 31:671–677, 1997. © 1997 Wiley-Liss, Inc.     

Inference for case-control studies when exposure status is both informatively missing and misclassified

In case–control studies, it is common for a categorical exposure variable to be misclassified. It is also common for exposure status to be informatively missing for some individuals, in that the probability of missingness may be related to exposure. Procedures for addressing the bias due to misclassification via validation data have been extensively studied, and related methods have been proposed for dealing with informative missingness based on supplemental sampling of some of those with missing data. In this paper, we introduce study designs and analytic procedures for dealing with both problems simultaneously in a 2×2 analysis. Results based on convergence in probability illustrate that the combined effects of missingness and misclassification, even when the latter is non‐differential, can lead to naïve exposure odds ratio estimates that are inflated or on the wrong side of the null. The motivating example comes from a case–control study of the association between low birth weight and the diagnosis of breast cancer later in life, where self‐reported birth weight for some women is supplemented by accurate information from birth certificates. Copyright © 2006 John Wiley & Sons, Ltd.     

Using full‐cohort data in nested case–control and case–cohort studies by multiple imputation

In many large prospective cohorts, expensive exposure measurements cannot be obtained for all individuals. Exposure–disease association studies are therefore often based on nested case–control or case–cohort studies in which complete information is obtained only for sampled individuals. However, in the full cohort, there may be a large amount of information on cheaply available covariates and possibly a surrogate of the main exposure(s), which typically goes unused. We view the nested case–control or case–cohort study plus the remainder of the cohort as a full‐cohort study with missing data. Hence, we propose using multiple imputation (MI) to utilise information in the full cohort when data from the sub‐studies are analysed. We use the fully observed data to fit the imputation models. We consider using approximate imputation models and also using rejection sampling to draw imputed values from the true distribution of the missing values given the observed data. Simulation studies show that using MI to utilise full‐cohort information in the analysis of nested case–control and case–cohort studies can result in important gains in efficiency, particularly when a surrogate of the main exposure is available in the full cohort. In simulations, this method outperforms counter‐matching in nested case–control studies and a weighted analysis for case–cohort studies, both of which use some full‐cohort information. Approximate imputation models perform well except when there are interactions or non‐linear terms in the outcome model, where imputation using rejection sampling works well. Copyright © 2013 John Wiley & Sons, Ltd.     

Quantifying and correcting for the winner's curse in genetic association studies

Genetic association studies are a powerful tool to detect genetic variants that predispose to human disease. Once an associated variant is identified, investigators are also interested in estimating the effect of the identified variant on disease risk. Estimates of the genetic effect based on new association findings tend to be upwardly biased due to a phenomenon known as the “winner's curse.” Overestimation of genetic effect size in initial studies may cause follow‐up studies to be underpowered and so to fail. In this paper, we quantify the impact of the winner's curse on the allele frequency difference and odds ratio estimators for one‐ and two‐stage case‐control association studies. We then propose an ascertainment‐corrected maximum likelihood method to reduce the bias of these estimators. We show that overestimation of the genetic effect by the uncorrected estimator decreases as the power of the association study increases and that the ascertainment‐corrected method reduces absolute bias and mean square error unless power to detect association is high. Genet. Epidemiol. 33:453–462, 2009. © 2009 Wiley‐Liss, Inc.     

Effects of early cat or dog ownership on sensitisation and asthma in a high‐risk cohort without disease‐related modification of exposure

Almqvist C, Garden F, Kemp AS, Li Q, Crisafulli D, Tovey ER, Xuan W, Marks GB for the CAPS investigators. Effects of early cat or dog ownership on sensitisation and asthma in a high‐risk cohort without disease‐related modification of exposure. Paediatric and Perinatal Epidemiology 2010; 24: 171–178. Variation in the observed association between pet ownership and allergic disease may be attributable to selection bias and confounding. The aim of this study was to suggest a method to assess disease‐related modification of exposure and second to examine how cat acquisition or dog ownership in early life affects atopy and asthma at 5 years. Information on sociodemographic factors and cat and dog ownership was collected longitudinally in an initially cat‐free Australian birth cohort based on children with a family history of asthma. At age 5 years, 516 children were assessed for wheezing, and 488 for sensitisation. Data showed that by age 5 years, 82 children had acquired a cat. Early manifestations of allergic disease did not foreshadow a reduced rate of subsequent acquisition of a cat. Independent risk factors for acquiring a cat were exposure to tobacco smoke at home odds ratio (OR) 1.92 [95% confidence interval (CI) 1.13, 3.26], maternal education ≤12 years OR 1.95 [1.08, 3.51] and dog ownership OR 2.23 [1.23, 4.05]. Cat or dog exposure in the first 5 years was associated with a decreased risk of any allergen sensitisation, OR 0.50 [0.28, 0.88] but no association with wheeze OR 0.96 [0.57, 1.61]. This risk was not affected by age at which the cat was acquired or whether the pet was kept in‐ or outdoors. In conclusion, cat or dog ownership reduced the risk of subsequent atopy in this high‐risk birth cohort. This cannot be explained by disease‐related modification of exposure. Public health recommendations on the effect of cat and dog ownership should be based on birth cohort studies where possible selection bias has been taken into account.     

Severity of bias of a simple estimator of the causal odds ratio in Mendelian randomization studies

Mendelian randomization studies estimate causal effects using genetic variants as instruments. Instrumental variable methods are straightforward for linear models, but epidemiologists often use odds ratios to quantify effects. Also, odds ratios are often the quantities reported in meta‐analyses. Many applications of Mendelian randomization dichotomize genotype and estimate the population causal log odds ratio for unit increase in exposure by dividing the genotype‐disease log odds ratio by the difference in mean exposure between genotypes. This ‘Wald‐type’ estimator is biased even in large samples, but whether the magnitude of bias is of practical importance is unclear. We study the large‐sample bias of this estimator in a simple model with a continuous normally distributed exposure, a single unobserved confounder that is not an effect modifier, and interpretable parameters. We focus on parameter values that reflect scenarios in which we apply Mendelian randomization, including realistic values for the degree of confounding and strength of the causal effect. We evaluate this estimator and the causal odds ratio using numerical integration and obtain approximate analytic expressions to check results and gain insight. A small simulation study examines finite sample bias and mild violations of the normality assumption. For our simple data‐generating model, we find that the Wald estimator is asymptotically biased with a bias of around 10% in fairly typical Mendelian randomization scenarios but which can be larger in more extreme situations. Recently developed methods such as structural mean models require fewer untestable assumptions and we recommend their use when the individual‐level data they require are available. The Wald‐type estimator may retain a role as an approximate method for meta‐analysis based on summary data. Copyright © 2012 John Wiley & Sons, Ltd.     

Estimation of bias in nongenetic observational studies using "mendelian triangulation"

PURPOSE: Phenotype-disease odds ratios calculated from the effect of a genotype on its phenotype and on disease risk ("Mendelian triangulation") can be used as a standard to assess bias on the corresponding odds ratio from nongenetic studies. Statistical tests are commonly used to compare these odds ratios. We propose a method to estimate the magnitude of the bias and judge the validity of the phenotype-disease association. METHODS: For four published examples, we obtained 10,000 random values from distributions of the odds ratios from both genetic and nongenetic studies. A range of values compatible with an unbiased odds ratio was then calculated from the empirical distribution of the differences between both odds ratios. RESULTS: We show that estimating a range of likely values for an unbiased odds ratio is useful to judge the effect of the phenotype and identify cases for which information from genetic studies adds little to the evaluation of the phenotype-disease association. Conversely, statistical tests could be misleading. CONCLUSIONS: Estimating a range of values for an unbiased odds ratio is more informative and appropriate than statistical tests when using the Mendelian triangulation approach for assessment of bias in phenotype-disease association studies.     

Network‐based regularization for matched case‐control analysis of high‐dimensional DNA methylation data

The matched case‐control designs are commonly used to control for potential confounding factors in genetic epidemiology studies especially epigenetic studies with DNA methylation. Compared with unmatched case‐control studies with high‐dimensional genomic or epigenetic data, there have been few variable selection methods for matched sets. In an earlier paper, we proposed the penalized logistic regression model for the analysis of unmatched DNA methylation data using a network‐based penalty. However, for popularly applied matched designs in epigenetic studies that compare DNA methylation between tumor and adjacent non‐tumor tissues or between pre‐treatment and post‐treatment conditions, applying ordinary logistic regression ignoring matching is known to bring serious bias in estimation. In this paper, we developed a penalized conditional logistic model using the network‐based penalty that encourages a grouping effect of (1) linked Cytosine‐phosphate‐Guanine (CpG) sites within a gene or (2) linked genes within a genetic pathway for analysis of matched DNA methylation data. In our simulation studies, we demonstrated the superiority of using conditional logistic model over unconditional logistic model in high‐dimensional variable selection problems for matched case‐control data. We further investigated the benefits of utilizing biological group or graph information for matched case‐control data. We applied the proposed method to a genome‐wide DNA methylation study on hepatocellular carcinoma (HCC) where we investigated the DNA methylation levels of tumor and adjacent non‐tumor tissues from HCC patients by using the Illumina Infinium HumanMethylation27 Beadchip. Several new CpG sites and genes known to be related to HCC were identified but were missed by the standard method in the original paper. Copyright © 2012 John Wiley & Sons, Ltd.     

A polytomous conditional likelihood approach for combining matched and unmatched case–control studies

In genetic association studies it is becoming increasingly imperative to have large sample sizes to identify and replicate genetic effects. To achieve these sample sizes, many research initiatives are encouraging the collaboration and combination of several existing matched and unmatched case–control studies. Thus, it is becoming more common to compare multiple sets of controls with the same case group or multiple case groups to validate or confirm a positive or negative finding. Usually, a naive approach of fitting separate models for each case–control comparison is used to make inference about disease–exposure association. But, this approach does not make use of all the observed data and hence could lead to inconsistent results. The problem is compounded when a common case group is used in each case–control comparison. An alternative to fitting separate models is to use a polytomous logistic model but, this model does not combine matched and unmatched case–control data. Thus, we propose a polytomous logistic regression approach based on a latent group indicator and a conditional likelihood to do a combined analysis of matched and unmatched case–control data. We use simulation studies to evaluate the performance of the proposed method and a case–control study of multiple myeloma and Inter‐Leukin‐6 as an example. Our results indicate that the proposed method leads to a more efficient homogeneity test and a pooled estimate with smaller standard error. Copyright © 2010 John Wiley & Sons, Ltd.     

A Bayesian approach to genetic association studies with family‐based designs

For genome‐wide association studies with family‐based designs, we propose a Bayesian approach. We show that standard transmission disequilibrium test and family‐based association test statistics can naturally be implemented in a Bayesian framework, allowing flexible specification of the likelihood and prior odds. We construct a Bayes factor conditional on the offspring phenotype and parental genotype data and then use the data we conditioned on to inform the prior odds for each marker. In the construction of the prior odds, the evidence for association for each single marker is obtained at the population‐level by estimating its genetic effect size by fitting the conditional mean model. Since such genetic effect size estimates are statistically independent of the effect size estimation within the families, the actual data set can inform the construction of the prior odds without any statistical penalty. In contrast to Bayesian approaches that have recently been proposed for genome‐wide association studies, our approach does not require assumptions about the genetic effect size; this makes the proposed method entirely data‐driven. The power of the approach was assessed through simulation. We then applied the approach to a genome‐wide association scan to search for associations between single nucleotide polymorphisms and body mass index in the Childhood Asthma Management Program data. Genet. Epidemiol. 34:569–574, 2010. © 2010 Wiley‐Liss, Inc.     

Bias correction to secondary trait analysis with case–control design

In genetic association studies with densely typed genetic markers, it is often of substantial interest to examine not only the primary phenotype but also the secondary traits for their association with the genetic markers. For more efficient sample ascertainment of the primary phenotype, a case–control design or its variants, such as the extreme‐value sampling design for a quantitative trait, are often adopted. The secondary trait analysis without correcting for the sample ascertainment may yield a biased association estimator. We propose a new method aiming at correcting the potential bias due to the inadequate adjustment of the sample ascertainment. The method yields explicit correction formulas that can be used to both screen the genetic markers and rapidly evaluate the sensitivity of the results to the assumed baseline case‐prevalence rate in the population. Simulation studies demonstrate good performance of the proposed approach in comparison with the more computationally intensive approaches, such as the compensator approaches and the maximum prospective likelihood approach. We illustrate the application of the approach by analysis of the genetic association of prostate specific antigen in a case–control study of prostate cancer in the African American population. Copyright © 2012 John Wiley & Sons, Ltd.     

Paraquat exposure and Parkinson’s disease: A systematic review and meta-analysis

Abstract

To reconcile and unify available results regarding paraquat exposure and Parkinson’s disease (PD), we conducted a systematic review and meta-analysis to provide a quantitative estimate of the risk of PD associated with paraquat exposure. Six scientific databases including PubMed, Cochrane libraries, EMBASE, Scopus, ISI Web of Knowledge, and TOXLINE were systematically searched. The overall odds ratios (ORs) with corresponding 95% CIs were calculated using a random-effects model. Of 7,309 articles identified, 13 case control studies with 3,231 patients and 4,901 controls were included into our analysis. Whereas, one prospective cohort studies was included into our systematic review. A subsequent meta-analysis showed an association between PD and paraquat exposure (odds ratio?=?1.64 (95% CI: 1.27–2.13; I²?=?24.8%). There is a statistically significant association between paraquat exposure and PD. Thus, future studies regarding paraquat and Parkinson’s disease are warranted.