The Missing Diversity in Human Genetic Studies.

The majority of studies of genetic association with disease have been performed in Europeans. This European bias has important implications for risk prediction of diseases across global populations. In this commentary, we justify the need to study more diverse populations using both empirical examples and theoretical reasoning.

Performance of Topology Tests under Extreme Selection Bias.

Tree tests like the Kishino-Hasegawa (KH) test and chi-square test suffer a selection bias that tests like the Shimodaira-Hasegawa (SH) test and approximately unbiased test were intended to correct. We investigate tree-testing performance in the presence of severe selection bias. The SH test is found to be very conservative and, surprisingly, its uncorrected analog, the KH test has low Type I error even in the presence of extreme selection bias, leading to a recommendation that the SH test be abandoned. A chi-square test is found to usually behave well and but to require correction in extreme cases. We show how topology testing procedures can be used to get support values for splits and compare the likelihood-based support values to the approximate likelihood ratio test (aLRT) support values. We find that the aLRT support values are reasonable even in settings with severe selection bias that they were not designed for. We also show how they can be used to construct tests of topologies and, in doing so, point out a multiple comparisons issue that should be considered when looking at support values for splits.

Viral burden is associated with age, vaccination, and viral variant in a population-representative study of SARS-CoV-2 that accounts for time-since-infection-related sampling bias.

In this study, we evaluated the impact of viral variant, in addition to other variables, on within-host viral burden, by analysing cycle threshold (Ct) values derived from nose and throat swabs, collected as part of the UK COVID-19 Infection Survey. Because viral burden distributions determined from community survey data can be biased due to the impact of variant epidemiology on the time-since-infection of samples, we developed a method to explicitly adjust observed Ct value distributions to account for the expected bias. By analysing the adjusted Ct values using partial least squares regression, we found that among unvaccinated individuals with no known prior exposure, viral burden was 44% lower among Alpha variant infections, compared to those with the predecessor strain, B.1.177. Vaccination reduced viral burden by 67%, and among vaccinated individuals, viral burden was 286% higher among Delta variant, compared to Alpha variant, infections. In addition, viral burden increased by 17% for every 10-year age increment of the infected individual. In summary, within-host viral burden increases with age, is reduced by vaccination, and is influenced by the interplay of vaccination status and viral variant.

Effector-independent Representations Guide Sequential Target Selection Biases in Action.

Previous work shows that automatic attention biases toward recently selected target features transfer across action and perception and even across different effectors such as the eyes and hands on a trial-by-trial basis. Although these findings suggest a common neural representation of selection history across effectors, the extent to which information about recently selected target features is encoded in overlapping versus distinct brain regions is unknown. Using fMRI and a priming of pop-out task where participants selected unpredictable, uniquely colored targets among homogeneous distractors via reach or saccade, we show that color priming is driven by shared, effector-independent underlying representations of recent selection history. Consistent with previous work, we found that the intraparietal sulcus (IPS) was commonly activated on trials where target colors were switched relative to those where the colors were repeated; however, the dorsal anterior insula exhibited effector-specific activation related to color priming. Via multivoxel cross-classification analyses, we further demonstrate that fine-grained patterns of activity in both IPS and the medial temporal lobe encode information about selection history in an effector-independent manner, such that ROI-specific models trained on activity patterns during reach selection could predict whether a color was repeated or switched on the current trial during saccade selection and vice versa. Remarkably, model generalization performance in IPS and medial temporal lobe also tracked individual differences in behavioral priming sensitivity across both types of action. These results represent a first step to clarify the neural substrates of experience-driven selection biases in contexts that require the coordination of multiple actions.

Selection Bias Requires Selection: The Case of Collider Stratification Bias.

In epidemiology, collider stratification bias, the bias resulting from conditioning on a common effect of two causes, is oftentimes considered a type of selection bias, regardless of the conditioning methods employed. In this commentary, we distinguish between two types of collider stratification bias: collider restriction bias due to restricting to one level of a collider (or a descendant of a collider) and collider adjustment bias through inclusion of a collider (or a descendant of a collider) in a regression model. We argue that categorizing collider adjustment bias as a form of selection bias may lead to semantic confusion, as adjustment for a collider in a regression model does not involve selecting a sample for analysis. Instead, we propose that collider adjustment bias can be better viewed as a type of overadjustment bias. We further provide two distinct causal diagram structures to distinguish collider restriction bias and collider adjustment bias. We hope that such a terminological distinction can facilitate easier and clearer communication.

Invited commentary: mixing multiple imputation and bootstrapping for variance estimation.

Multiple imputation (MI) is commonly implemented to mitigate potential selection bias due to missing data. The accompanying article by Nguyen and Stuart (Am J Epidemiol. 2024;193(10):1470-1476) examines the statistical consistency of several ways of integrating MI with propensity scores. As Nguyen and Stuart noted, variance estimation for these different approaches remains to be developed. One common option is the nonparametric bootstrap, which can provide valid inference when closed-form variance estimators are not available. However, there is no consensus on how to implement MI and nonparametric bootstrapping in analyses. To complement Nguyen and Stuart's article on MI and propensity score analyses, we review some currently available approaches on variance estimation with MI and nonparametric bootstrapping.

Attention-deficit hyperactivity disorder symptoms and brain morphology: Addressing potential selection bias with inverse probability weighting.

The goal of this study was to examine what happens to established associations between attention deficit hyperactivity disorder (ADHD) symptoms and cortical surface and thickness regions once we apply inverse probability of censoring weighting (IPCW) to address potential selection bias. Moreover, we illustrate how different factors that predict participation contribute to potential selection bias. Participants were 9- to 11-year-old children from the Generation R study (N = 2707). Cortical area and thickness were measured with magnetic resonance imaging (MRI) and ADHD symptoms with the Child Behavior Checklist. We examined how associations between ADHD symptoms and brain morphology change when we weight our sample back to either follow-up (ages 9-11), baseline (cohort at birth), or eligible (population of Rotterdam at time of recruitment). Weights were derived using IPCW or raking and missing predictors of participation used to estimate weights were imputed. Weighting analyses to baseline and eligible increased beta coefficients for the middle temporal gyrus surface area, as well as fusiform gyrus cortical thickness. Alternatively, the beta coefficient for the rostral anterior cingulate decreased. Removing one group of variables used for estimating weights resulted in the weighted regression coefficient moving closer to the unweighted regression coefficient. In addition, we found considerably different beta coefficients for most surface area regions and all thickness measures when we did not impute missing covariate data. Our findings highlight the importance of using inverse probability weighting (IPW) in the neuroimaging field, especially in the context of mental health-related research. We found that including all variables related to exposure-outcome in the IPW model and combining IPW with multiple imputations can help reduce bias. We encourage future psychiatric neuroimaging studies to define their target population, collect information on eligible but not included participants and use inverse probability of censoring weighting (IPCW) to reduce selection bias.

Towards a Clearer Causal Question Underlying the Association Between Cancer and Dementia.

BACKGROUND: Several observational studies have described an inverse association between cancer diagnosis and subsequent dementia risk. Multiple biologic mechanisms and potential biases have been proposed in attempts to explain this association. One proposed explanation is the opposite expression of Pin1 in cancer and dementia, and we use this explanation and potential drug target to illustrate the required assumptions and potential sources of bias for inferring an effect of Pin1 on dementia risk from analyses measuring cancer diagnosis as a proxy for Pin1 expression. METHODS: We used data from the Rotterdam Study, a population-based cohort. We estimate the association between cancer diagnosis (as a proxy for Pin1) and subsequent dementia diagnosis using two different proxy methods and with confounding and censoring for death addressed with inverse probability weights. We estimate and compare the complements of a weighted Kaplan-Meier survival estimator at 20 years of follow-up. RESULTS: Out of 3634 participants, 899 (25%) were diagnosed with cancer, of whom 53 (6%) had dementia, and 567 (63%) died. Among those without cancer, 15% (411) were diagnosed with dementia, and 667 (24%) died over follow-up. Depending on the confounding and selection bias control, and the way in which cancer was used as a time-varying proxy exposure, the risk ratio for dementia diagnosis ranged from 0.71 (95% confidence interval [CI] = 0.49, 0.95) to 1.1 (95% CI = 0.79, 1.3). CONCLUSION: Being explicit about the underlying mechanism of interest is key to maximizing what we can learn from this cancer-dementia association given available or readily collected data, and to defining, detecting, and preventing potential biases.

Effects of Recall and Selection Biases on Modeling Cancer Risk From Mobile Phone Use: Results From a Case-Control Simulation Study.

BACKGROUND: The largest case-control study (Interphone study) investigating glioma risk related to mobile phone use showed a J-shaped relationship with reduced relative risks for moderate use and a 40% increased relative risk among the 10% heaviest regular mobile phone users, using a categorical risk model based on deciles of lifetime duration of use among ever regular users. METHODS: We conducted Monte Carlo simulations examining whether the reported estimates are compatible with an assumption of no effect of mobile phone use on glioma risk when the various forms of biases present in the Interphone study are accounted for. Four scenarios of sources of error in self-reported mobile phone use were considered, along with selection bias. Input parameters used for simulations were those obtained from Interphone validation studies on reporting accuracy and from using a nonresponse questionnaire. RESULTS: We found that the scenario simultaneously modeling systematic and random reporting errors produced a J-shaped relationship perfectly compatible with the observed relationship from the main Interphone study with a simulated spurious increased relative risk among heaviest users (odds ratio = 1.91) compared with never regular users. The main determinant for producing this J shape was higher reporting error variance in cases compared with controls, as observed in the validation studies. Selection bias contributed to the reduced risks as well. CONCLUSIONS: Some uncertainty remains, but the evidence from the present simulation study shifts the overall assessment to making it less likely that heavy mobile phone use is causally related to an increased glioma risk.

How accurately can we assess zoonotic risk?

Identifying the animal reservoirs from which zoonotic viruses will likely emerge is central to understanding the determinants of disease emergence. Accordingly, there has been an increase in studies attempting zoonotic "risk assessment." Herein, we demonstrate that the virological data on which these analyses are conducted are incomplete, biased, and rapidly changing with ongoing virus discovery. Together, these shortcomings suggest that attempts to assess zoonotic risk using available virological data are likely to be inaccurate and largely only identify those host taxa that have been studied most extensively. We suggest that virus surveillance at the human-animal interface may be more productive.

Bias correction models for electronic health records data in the presence of non-random sampling.

Electronic health records (EHRs) contain rich clinical information for millions of patients and are increasingly used for public health research. However, non-random inclusion of subjects in EHRs can result in selection bias, with factors such as demographics, socioeconomic status, healthcare referral patterns, and underlying health status playing a role. While this issue has been well documented, little work has been done to develop or apply bias-correction methods, often due to the fact that most of these factors are unavailable in EHRs. To address this gap, we propose a series of Heckman type bias correction methods by incorporating social determinants of health selection covariates to model the EHR non-random sampling probability. Through simulations under various settings, we demonstrate the effectiveness of our proposed method in correcting biases in both the association coefficient and the outcome mean. Our method augments the utility of EHRs for public health inferences, as we show by estimating the prevalence of cardiovascular disease and its correlation with risk factors in the New York City network of EHRs.

Steady-state statistical properties and implementation of randomization designs with maximum tolerated imbalance restriction for two-arm equal allocation clinical trials.

In recent decades, several randomization designs have been proposed in the literature as better alternatives to the traditional permuted block design (PBD), providing higher allocation randomness under the same restriction of the maximum tolerated imbalance (MTI). However, PBD remains the most frequently used method for randomizing subjects in clinical trials. This status quo may reflect an inadequate awareness and appreciation of the statistical properties of these randomization designs, and a lack of simple methods for their implementation. This manuscript presents the analytic results of statistical properties for five randomization designs with MTI restriction based on their steady-state probabilities of the treatment imbalance Markov chain and compares them to those of the PBD. A unified framework for randomization sequence generation and real-time on-demand treatment assignment is proposed for the straightforward implementation of randomization algorithms with explicit formulas of conditional allocation probabilities. Topics associated with the evaluation, selection, and implementation of randomization designs are discussed. It is concluded that for two-arm equal allocation trials, several randomization designs offer stronger protection against selection bias than the PBD does, and their implementation is not necessarily more difficult than the implementation of the PBD.

Embedded multilevel regression and poststratification: Model-based inference with incomplete auxiliary information.

Health disparity research often evaluates health outcomes across demographic subgroups. Multilevel regression and poststratification (MRP) is a popular approach for small subgroup estimation as it can stabilize estimates by fitting multilevel models and adjust for selection bias by poststratifying on auxiliary variables, which are population characteristics predictive of the analytic outcome. However, the granularity and quality of the estimates produced by MRP are limited by the availability of the auxiliary variables' joint distribution; data analysts often only have access to the marginal distributions. To overcome this limitation, we embed the estimation of population cell counts needed for poststratification into the MRP workflow: embedded MRP (EMRP). Under EMRP, we generate synthetic populations of the auxiliary variables before implementing MRP. All sources of estimation uncertainty are propagated with a fully Bayesian framework. Through simulation studies, we compare different methods of generating the synthetic populations and demonstrate EMRP's improvements over alternatives on the bias-variance tradeoff to yield valid subpopulation inferences of interest. We apply EMRP to the Longitudinal Survey of Wellbeing and estimate food insecurity prevalence among vulnerable groups in New York City. We find that all EMRP estimators can correct for the bias in classical MRP while maintaining lower standard errors and narrower confidence intervals than directly imputing with the weighted finite population Bayesian bootstrap (WFPBB) and design-based estimates. Performances from the EMRP estimators do not differ substantially from each other, though we would generally recommend using the WFPBB-MRP for its consistently high coverage rates.

A class of computational methods to reduce selection bias when designing Phase 3 clinical trials.

When designing confirmatory Phase 3 studies, one usually evaluates one or more efficacious and safe treatment option(s) based on data from previous studies. However, several retrospective research articles reported the phenomenon of "diminished treatment effect in Phase 3" based on many case studies. Even under basic assumptions, it was shown that the commonly used estimator could substantially overestimate the efficacy of selected group(s). As alternatives, we propose a class of computational methods to reduce estimation bias and mean squared error with a broader scope of multiple treatment groups and flexibility to accommodate summary results by group as input. Based on simulation studies and a real data example, we provide practical implementation guidance for this class of methods under different scenarios. For more complicated problems, our framework can serve as a starting point with additional layers built in. Proposed methods can also be widely applied to other selection problems.

Understanding an impact of patient enrollment pattern on predictability of central (unstratified) randomization in a multi-center clinical trial.

In a multi-center randomized controlled trial (RCT) with competitive recruitment, eligible patients are enrolled sequentially by different study centers and are randomized to treatment groups using the chosen randomization method. Given the stochastic nature of the recruitment process, some centers may enroll more patients than others, and in some instances, a center may enroll multiple patients in a row, for example, on a given day. If the study is open-label, the investigators might be able to make intelligent guesses on upcoming treatment assignments in the randomization sequence, even if the trial is centrally randomized and not stratified by center. In this paper, we use enrollment data inspired by a real multi-center RCT to quantify the susceptibility of two restricted randomization procedures, the permuted block design and the big stick design, to selection bias under the convergence strategy of Blackwell and Hodges (1957) applied at the center level. We provide simulation evidence that the expected proportion of correct guesses may be greater than 50% (i.e., an increased risk of selection bias) and depends on the chosen randomization method and the number of study patients recruited by a given center that takes consecutive positions on the central allocation schedule. We propose some strategies for ensuring stronger encryption of the randomization sequence to mitigate the risk of selection bias.

A unified Bayesian framework for bias adjustment in multiple comparisons from clinical trials.

In clinical trials, multiple comparisons arising from various treatments/doses, subgroups, or endpoints are common. Typically, trial teams focus on the comparison showing the largest observed treatment effect, often involving a specific treatment pair and endpoint within a subgroup. These findings frequently lead to follow-up pivotal studies, many of which do not confirm the initial positive results. Selection bias occurs when the most promising treatment, subgroup, or endpoint is chosen for further development, potentially skewing subsequent investigations. Such bias can be defined as the deviation in the observed treatment effects from the underlying truth. In this article, we propose a general and unified Bayesian framework to address selection bias in clinical trials with multiple comparisons. Our approach does not require a priori specification of a parametric distribution for the prior, offering a more flexible and generalized solution. The proposed method facilitates a more accurate interpretation of clinical trial results by adjusting for such selection bias. Through simulation studies, we compared several methods and demonstrated their superior performance over the normal shrinkage estimator. We recommended the use of Bayesian Model Averaging estimator averaging over Gaussian Mixture Models as the prior distribution based on its performance and flexibility. We applied the method to a multicenter, randomized, double-blind, placebo-controlled study investigating the cardiovascular effects of dulaglutide.

Revising model for end-stage liver disease from calendar-time cross-sections with correction for selection bias.

BACKGROUND: Eurotransplant liver transplant candidates are prioritized by Model for End-stage Liver Disease (MELD), a 90-day waitlist survival risk score based on the INR, creatinine and bilirubin. Several studies revised the original MELD score, UNOS-MELD, with transplant candidate data by modelling 90-day waitlist mortality from waitlist registration, censoring patients at delisting or transplantation. This approach ignores biomarkers reported after registration, and ignores informative censoring by transplantation and delisting. METHODS: We study how MELD revision is affected by revision from calendar-time cross-sections and correction for informative censoring with inverse probability censoring weighting (IPCW). For this, we revised UNOS-MELD on patients with chronic liver cirrhosis on the Eurotransplant waitlist between 2007 and 2019 (n = 13,274) with Cox models with as endpoints 90-day survival (a) from registration and (b) from weekly drawn calendar-time cross-sections. We refer to the revised score from cross-section with IPCW as DynReMELD, and compare DynReMELD to UNOS-MELD and ReMELD, a prior revision of UNOS-MELD for Eurotransplant, in geographical validation. RESULTS: Revising MELD from calendar-time cross-sections leads to significantly different MELD coefficients. IPCW increases estimates of absolute 90-day waitlist mortality risks by approximately 10 percentage points. DynReMELD has improved discrimination over UNOS-MELD (delta c-index: 0.0040, p < 0.001) and ReMELD (delta c-index: 0.0015, p < 0.01), with differences comparable in magnitude to the addition of an extra biomarker to MELD (delta c-index: ± 0.0030). CONCLUSION: Correcting for selection bias by transplantation/delisting does not improve discrimination of revised MELD scores, but substantially increases estimated absolute 90-day mortality risks. Revision from cross-section uses waitlist data more efficiently, and improves discrimination compared to revision of MELD exclusively based on information available at listing.

Waiving the consent requirement to mitigate bias in observational precision medicine research.

Consent bias is a type of selection bias in biomedical research where those consenting to the research differ systematically from those not consenting. It is particularly relevant in precision medicine research because the complexity of these studies prevents certain subgroups from understanding, trusting, and consenting to the research. Because consent bias distorts research findings and causes inequitable distribution of research benefits, scholars propose two types of schemes to reduce consent bias: reforming existing consent models and removing the consent requirement altogether. This study explores the possibility of waiving consent in observational studies using existing data, because they involve fewer risks to participants than clinical trials if privacy safeguards are strengthened. It suggests that data protection mechanisms such as security enhancement and data protection impact assessment should be conducted to protect data privacy of participants in observational studies without consent.

Using the global randomization test as a Mendelian randomization falsification test for the exclusion restriction assumption.

Mendelian randomization may give biased causal estimates if the instrument affects the outcome not solely via the exposure of interest (violating the exclusion restriction assumption). We demonstrate use of a global randomization test as a falsification test for the exclusion restriction assumption. Using simulations, we explored the statistical power of the randomization test to detect an association between a genetic instrument and a covariate set due to (a) selection bias or (b) horizontal pleiotropy, compared to three approaches examining associations with individual covariates: (i) Bonferroni correction for the number of covariates, (ii) correction for the effective number of independent covariates, and (iii) an r2 permutation-based approach. We conducted proof-of-principle analyses in UK Biobank, using CRP as the exposure and coronary heart disease (CHD) as the outcome. In simulations, power of the randomization test was higher than the other approaches for detecting selection bias when the correlation between the covariates was low (r2 < 0.1), and at least as powerful as the other approaches across all simulated horizontal pleiotropy scenarios. In our applied example, we found strong evidence of selection bias using all approaches (e.g., global randomization test p < 0.002). We identified 51 of the 58 CRP genetic variants as horizontally pleiotropic, and estimated effects of CRP on CHD attenuated somewhat to the null when excluding these from the genetic risk score (OR = 0.96 [95% CI: 0.92, 1.00] versus 0.97 [95% CI: 0.90, 1.05] per 1-unit higher log CRP levels). The global randomization test can be a useful addition to the MR researcher's toolkit.

The HOPE cohort: cohort profile and evaluation of selection bias.

The HOPE cohort is a Danish nationwide cohort with ongoing follow-up, holding information on postpartum depression (PPD) symptoms and diagnoses on 170,218 childbirths (142,795 unique mothers). These data have been linked with extensive register data on health and socioeconomic information on the mothers, their partners, parents, and children. This cohort profile aimed to provide an overview of the data collection and content, describe characteristics, and evaluate potential selection bias. PPD screenings, using the Edinburgh Postnatal Depression Scale, were collected from 67 of the 98 Danish municipalities, covering the period January 2015 to December 2021. This data was linked with register data on PPD diagnoses (identified through medication prescriptions and hospital contacts) as well as background information. Cohort characteristics were compared to the source population, defined as all childbirths by women residing in Denmark during the same period (452,207 childbirths). Potential selection bias was evaluated by comparing odds ratios of five well-established associations between the cohort and the source population. The HOPE cohort holds information on 170,218 childbirths (38% of the source population) involving 142,795 unique mothers. The HOPE cohort only differed slightly from the source population on most characteristics examined, but larger differences were observed on specific characteristics with an underrepresentation of the youngest and oldest age groups, women with more than three children or twins/triplets, and women born outside Denmark. Similar associations were identified across the two populations within the five well-established associations. There was no indication of selection bias on the five examined associations, and the HOPE cohort is representative of the source population on important perinatal characteristics.