Epidemiologic Bias Research Articles

The Evolution of Selection Bias in the Recent Epidemiologic Literature-A Selective Overview.

Selection bias has long been central in methodological discussions across epidemiology and other fields. In epidemiology, the concept of selection bias has been continually evolving over time. In this issue of the Journal, Mathur and Shpitser (Am J Epidemiol. XXXX;XXX(XX):XXXX-XXXX) present simple graphical rules for using a Single World Intervention Graph (SWIG) to assess the presence of selection bias when estimating treatment effects in both the general population and a selected sample. Notably, the authors examine the setting in which the treatment affects selection, an issue not well-addressed in the existing literature on selection bias. To place the work by Mathur and Shpitser in context, we review the evolution of the concept of selection bias in epidemiology, with a primary focus on the developments in the last 20-30 years since the introduction of causal directed acyclic graphs (DAGs) to epidemiologic research.

A Potential Outcomes Approach to Selection Bias.

We propose a novel definition of selection bias in analytic epidemiology using potential outcomes. This definition captures selection bias under both the structural approach (where conditioning on selection into the study opens a noncausal path from exposure to disease in a directed acyclic graph) and the traditional definition (where a given measure of association differs between the study sample and the population eligible for inclusion). This approach is nonparametric, and selection bias under the approach can be analyzed using single-world intervention graphs both under and away from the null hypothesis. It allows the simultaneous analysis of confounding and selection bias, it explicitly links the selection of study participants to the estimation of causal effects using study data, and it can be adapted to handle selection bias in descriptive epidemiology. Through examples, we show that this approach provides a novel perspective on the variety of mechanisms that can generate selection bias and simplifies the analysis of selection bias in matched studies and case-cohort studies.

European’ health care indicators and pancreatic cancer incidence and mortality: A mediation analysis of Eurostat data and Global Burden of Disease Study 2019

AimTo highlight correlations existing between incidence and mortality of pancreatic cancer, and health care indicators in 36 European countries. MethodsThe Global Burden of Disease (GBD) and Eurostat databases were queried between 2004 and 2019. Incidence and mortality were age-standardized. From Eurostat, indicators regarding expenditure, hospital beds, medical technology, health personnel, physicians by medical specialty and unmet needs for medical examination were extracted. Correlations between GBD and Eurostat data were analysed through mediation analysis applying clustering for countries. ResultsIncidence increased by +0.6% per year (p = 0.001) and mortality by +0.3% (p = 0.001), being increasing for most of the European countries considered. Incidence and mortality were strongly positively correlated (p = 0.001). Higher current health expenditure, expenditure in inpatient curative care, the number of available beds, the number of computed tomography scan, magnetic resonance units, practising medical doctors were all related to higher incidence (p < 0.05), whereas the unmet need for medical examinations was related to lower incidence. When the mediator’ effect of incidence was handled, these indicators, together with expenditure on outpatient curative cares, the number of pet scanners and of radiation therapy equipment, were related to lower mortality (p < 0.05). ConclusionsHealth care environment correlates with reported incidence and mortality of pancreatic cancer. This highlights both that ameliorated socio-economic societies suffer from higher incidence but lower mortality, as well as the epidemiological bias originating from countries’ diagnostic ability.

Immortal Time Bias With Time-Varying Exposures in Environmental Epidemiology: A Case Study in Lung Cancer Survival.

Immortal time bias is a well-recognized bias in clinical epidemiology but is rarely discussed in environmental epidemiology. Under the target trial framework, this bias is formally conceptualized as a misalignment between the start of study follow-up (time 0) and treatment assignment. This misalignment can occur when attained duration of follow-up is encoded into treatment assignment using minimums, maximums, or averages. The bias can be exacerbated in the presence of time trends commonly found in environmental exposures. Using lung cancer cases from the California Cancer Registry (2000-2010) linked with estimated concentrations of particulate matter less than or equal to 2.5 μm in aerodynamic diameter (PM2.5), we replicated previous studies that averaged PM2.5 exposure over follow-up in a time-to-event model. We compared this approach with one that ensures alignment between time 0 and treatment assignment, a discrete-time approach. In the former approach, the estimated overall hazard ratio for a 5-μg/m3 increase in PM2.5 was 1.38 (95% confidence interval: 1.36, 1.40). Under the discrete-time approach, the estimated pooled odds ratio was 0.99 (95% confidence interval: 0.98, 1.00). We conclude that the strong estimated effect in the former approach was likely driven by immortal time bias, due to misalignment at time 0. Our findings highlight the importance of appropriately conceptualizing a time-varying environmental exposure under the target trial framework to avoid introducing preventable systematic errors.

Pro-Con Debate: Interdisciplinary Perspectives on Industry-Sponsored Research.

Pro: Nearly all new devices and drugs come from industry that provides two-thirds of the funding for medical research, and a much higher fraction of clinical research. Realistically, without corporate-funded studies, perioperative research would stagnate with little innovation and few new products. Opinions are ubiquitous and normal, but do not constitute epidemiologic bias. Competent clinical research includes many protections against selection and measurement bias, and the publication process provides at least moderate protection against misinterpretation of results. Trial registries largely prevent selective data presentation. Sponsored trials are particularly protected against inappropriate corporate influence because they are usually codesigned with the US Food and Drug Administration, and analyses are based on formal predefined statistical plans, as well as being conducted with rigorous external monitoring. Novel products, which are essential for advances in clinical care, largely come from industry, and industry appropriately funds much of the required research. We should celebrate industry's contribution to improvements in clinical care. Con: While industry funding contributes to research and discovery, examples of industry-funded research demonstrate bias. In the setting of financial pressures and potential conflict of interest, bias can influence the type of study design, hypotheses being tested, rigor and transparency in data analysis, interpretation, as well as reporting of the results. Unlike public granting agencies, industry does not necessarily provide funding based on unbiased peer review following an open call for proposals. The focus on success can influence the choice of a comparator, which might not be ideal among the possible alternatives, the language used in the publication, and even the ability to publish. Unpublished negative trials can result in selected information being withheld from the scientific community and the public. Appropriate safeguards are needed to ensure that research addresses the most important and relevant questions, that results are available even when they do not support the use of a product produced by the funding company, that populations studied reflect the relevant patients, that the most rigorous approaches are applied, that studies have the appropriate power to address the question posed, and that conclusions are presented in an unbiased manner.

Pro-Con Debate: Interdisciplinary Perspectives on Industry-Sponsored Research

ProNearly all new devices and drugs come from industry that provides two-thirds of the funding for medical research, and a much higher fraction of clinical research. Realistically, without corporate-funded studies, perioperative research would stagnate with little innovation and few new products. Opinions are ubiquitous and normal but do not constitute epidemiologic bias. Competent clinical research includes many protections against selection and measurement bias, and the publication process provides at least moderate protection against misinterpretation of results. Trial registries largely prevent selective data presentation. Sponsored trials are particularly protected against inappropriate corporate influence because they are usually codesigned with the US Food and Drug Administration, and analyses are based on formal predefined statistical plans, as well as being conducted with rigorous external monitoring. Novel products, which are essential for advances in clinical care, largely come from industry, and industry appropriately funds much of the required research. We should celebrate industry’s contribution to improvements in clinical care. ConWhile industry funding contributes to research and discovery, examples of industry-funded research demonstrate bias. In the setting of financial pressures and potential conflict of interest, bias can influence the type of study design, hypotheses being tested, rigor and transparency in data analysis, interpretation, as well as reporting of the results. Unlike public granting agencies, industry does not necessarily provide funding based on unbiased peer review following an open call for proposals. The focus on success can influence the choice of a comparator, which might not be ideal among the possible alternatives, the language used in the publication, and even the ability to publish. Unpublished negative trials can result in selected information being withheld from the scientific community and the public. Appropriate safeguards are needed to ensure that research addresses the most important and relevant questions, that results are available even when they do not support the use of a product produced by the funding company, that populations studied reflect the relevant patients, that the most rigorous approaches are applied, that studies have the appropriate power to address the question posed, and that conclusions are presented in an unbiased manner.

Waning of 2-Dose BNT162b2 and mRNA-1273 Vaccine Effectiveness Against Symptomatic SARS-CoV-2 Infection Accounting for Depletion-of-Susceptibles Bias.

Concerns about the duration of protection conferred by coronavirus disease 2019 (COVID-19) vaccines have arisen in postlicensure evaluations. "Depletion of susceptibles," a bias driven by differential accrual of infection among vaccinated and unvaccinated individuals, may obscure vaccine effectiveness (VE) estimates, hindering interpretation. We enrolled California residents who received molecular SARS-CoV-2 tests in a matched, test-negative design, case-control study to estimate VE of mRNA-based COVID-19 vaccines between February 23 and December 5, 2021. We analyzed waning protection following 2 vaccine doses using conditional logistic regression models. Additionally, we used data from a population-based serological study to adjust for "depletion-of-susceptibles" bias and estimated VE for 3 doses, by time since second dose receipt. Pooled VE of BNT162b2 and mRNA-1273 against symptomatic SARS-CoV-2 infection was 91.3% (95% confidence interval (CI): 83.8, 95.4) at 14 days after second-dose receipt and declined to 50.8% (95% CI: 19.7, 69.8) at 7 months. Adjusting for depletion-of-susceptibles bias, we estimated VE of 53.2% (95% CI: 23.6, 71.2) at 7 months after primary mRNA vaccination series. A booster dose of BN162b2 or mRNA-1273 increased VE to 95.0% (95% CI: 82.8, 98.6). These findings confirm that observed waning of protection is not attributable to epidemiologic bias and support ongoing efforts to administer additional vaccine doses to mitigate burden of COVID-19.

Causal Modeling to Mitigate Selection Bias and Unmeasured Confounding in Internet-Based Epidemiology of COVID-19: Model Development and Validation.

BackgroundSelection bias and unmeasured confounding are fundamental problems in epidemiology that threaten study internal and external validity. These phenomena are particularly dangerous in internet-based public health surveillance, where traditional mitigation and adjustment methods are inapplicable, unavailable, or out of date. Recent theoretical advances in causal modeling can mitigate these threats, but these innovations have not been widely deployed in the epidemiological community.ObjectiveThe purpose of our paper is to demonstrate the practical utility of causal modeling to both detect unmeasured confounding and selection bias and guide model selection to minimize bias. We implemented this approach in an applied epidemiological study of the COVID-19 cumulative infection rate in the New York City (NYC) spring 2020 epidemic.MethodsWe collected primary data from Qualtrics surveys of Amazon Mechanical Turk (MTurk) crowd workers residing in New Jersey and New York State across 2 sampling periods: April 11-14 and May 8-11, 2020. The surveys queried the subjects on household health status and demographic characteristics. We constructed a set of possible causal models of household infection and survey selection mechanisms and ranked them by compatibility with the collected survey data. The most compatible causal model was then used to estimate the cumulative infection rate in each survey period.ResultsThere were 527 and 513 responses collected for the 2 periods, respectively. Response demographics were highly skewed toward a younger age in both survey periods. Despite the extremely strong relationship between age and COVID-19 symptoms, we recovered minimally biased estimates of the cumulative infection rate using only primary data and the most compatible causal model, with a relative bias of +3.8% and –1.9% from the reported cumulative infection rate for the first and second survey periods, respectively.ConclusionsWe successfully recovered accurate estimates of the cumulative infection rate from an internet-based crowdsourced sample despite considerable selection bias and unmeasured confounding in the primary data. This implementation demonstrates how simple applications of structural causal modeling can be effectively used to determine falsifiable model conditions, detect selection bias and confounding factors, and minimize estimate bias through model selection in a novel epidemiological context. As the disease and social dynamics of COVID-19 continue to evolve, public health surveillance protocols must continue to adapt; the emergence of Omicron variants and shift to at-home testing as recent challenges. Rigorous and transparent methods to develop, deploy, and diagnosis adapted surveillance protocols will be critical to their success.

Al margen de los cuidados: experiencias de (des)protección sociosanitaria de trabajadoras del sexo en Oporto, Portugal

This article analyzes the social and health challenges linked to sex work and the interventions carried out within this arena, taking into consideration the ways in which prostitution is socially marginalized. Basing our analysis on the experiences and understandings of sex workers practicing in the city of Oporto (Portugal), we attempt to understand the main forms of vulnerability faced by this group, along with the ways in which they experience the relationship with State services and social and health interventions that target them. Qualitative field research was conducted involving a case study of an intervention project. Data collection was carried out from March to June 2019, and included documentary research, participant observation, and semi-structured interviews with project staff and six sex workers. Among the results of our study, we found that sex workers are subject to severe constraints, limiting their use of informal support networks and their access to State social protection and healthcare services. Moreover, although the intervention project based on harm reduction intended to support these women and was indeed valued, it had a strong epidemiological prevention bias favoring an individualistic and assistentialist approach, while failing to address other social vulnerabilities identified by the women themselves.

Avoiding overadjustment bias in social epidemiology through appropriate covariate selection: a primer

Obtaining accurate estimates of the causal effects of socioeconomic position (SEP) on health is important for public health interventions. To do this, researchers must identify and adjust for all potential confounding variables, while avoiding inappropriate adjustment for mediator variables on a causal pathway between the exposure and outcome. Unfortunately, ‘overadjustment bias’ remains a common and under-recognized problem in social epidemiology. This paper offers an introduction on selecting appropriate variables for adjustment when examining effects of SEP on health, with a focus on overadjustment bias. We discuss the challenges of estimating different causal effects including overadjustment bias, provide guidance on overcoming them, and consider specific issues including the timing of variables across the life-course, mutual adjustment for socioeconomic indicators, and conducting systematic reviews. We recommend three key steps to select the most appropriate variables for adjustment. First, researchers should be clear about their research question and causal effect of interest. Second, using expert knowledge and theory, researchers should draw causal diagrams representing their assumptions about the interrelationships between their variables of interest. Third, based on their causal diagram(s) and causal effect(s) of interest, researchers should select the most appropriate set of variables, which maximizes adjustment for confounding while minimizing adjustment for mediators.

Plasma Exosome-Derived microRNAs as Potential Diagnostic and Prognostic Biomarkers in Brazilian Pancreatic Cancer Patients.

Pancreatic cancer represents one of the leading causes of oncological death worldwide. A combination of pancreatic cancer aggressiveness and late diagnosis are key factors leading to a low survival rate and treatment inefficiency, and early diagnosis is pursued as a critical factor for pancreatic cancer. In this context, plasma microRNAs are emerging as promising players due to their non-invasive and practical usage in oncological diagnosis and prognosis. Recent studies have showed some miRNAs associated with pancreatic cancer subtypes, or with stages of the disease. Here we demonstrate plasma exosome-derived microRNA expression in pancreatic cancer patients and healthy individuals from Brazilian patients. Using plasma of 65 pancreatic cancer patients and 78 healthy controls, plasma exosomes were isolated and miRNAs miR-27b, miR-125b-3p, miR-122-5p, miR-21-5p, miR-221-3p, miR-19b, and miR-205-5p were quantified by RT-qPCR. We found that miR-125b-3p, miR-122-5p, and miR-205-5p were statistically overexpressed in the plasma exosomes of pancreatic cancer patients compared to healthy controls. Moreover, miR-205-5p was significantly overexpressed in European descendants, in patients with tumor progression and in those who died from the disease, and diagnostic ability by ROC curve was 0.86. Therefore, we demonstrate that these three microRNAs are potential plasma exosome-derived non-invasive biomarkers for the diagnosis and prognosis of Brazilian pancreatic cancer, demonstrating the importance of different populations and epidemiological bias.

1274The application of simulation to quantifying the influence of bias in reproductive and perinatal epidemiology

Abstract Background Establishing causal effects in reproductive and perinatal epidemiology is challenging due to the many selection and attrition processes from preconception to the postnatal period. Further challenging, is the potential for the misclassification of exposures, outcomes and confounders, contributing to measurement error. The application of simulation enables the illustration and quantification of the magnitude of various types of bias commonly found in observational studies. Methods A systematic search was conducted in PubMed, Medline, Embase, CINAHL and Scopus in August 2020. A gray literature search of Google and Google Scholar, followed by a search of the reference lists of included studies, was undertaken. Results Thirty-nine studies, covering information (n = 14), selection (n = 14), confounding (n = 9), protection (n = 1), and attenuation bias (n = 1) were identified. The methods of simulating data and reporting of results varied, with more recent studies including causal diagrams. Few studies included code for replication. Although there has been an increasing application of simulation in reproductive and perinatal epidemiology since 2015, overall this remains an underexplored area. Conclusions The studies demonstrated effectiveness in the quantification of multiple types of bias using simulation. The limited use implies that further effort is required to increase knowledge of the application of simulation, which will thereby improve causal interpretation in reproductive and perinatal studies. Key messages Practical guidance for researchers is required in the development, analysis and reporting of simulation methods for the quantification of bias.

TDAH dans l’enfance et trouble psychotique à l’âge adulte : quel lien ?

ObjectivesAttention Deficit with/without Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder with frequent comorbid psychiatric disorders. Several studies have underlined the increased risk of developing a psychotic disorder subsequent to a childhood ADHD. The aim of our review is not only to clarify this association and the related physiopathology but also to understand the consequences for therapeutic management. MethodsWe processed a narrative review of available literature based on a research of the PubMed database. Articles related to ADHD and psychotic disorder on a genetical, clinical or biological level were selected by one of the authors. ResultsADHD and psychotic disorders share neonatal, environmental, and genetic risk factors. On a neurobiological level, both disorders are concerned by a dysfunction of the dopaminergic system with an abnormal regulation of dopaminergic neurons’ phasic and tonic activity. Our review aims to explain the « dynamic » model of dopaminergic dysfunctions and propose some guidance for pharmacological treatment of ADHD, with or without psychotic disorder. This model offers a better understanding of why methylphenidate is not associated to an increased risk of psychotic disorder and could act as a protective factor. Association between ADHD and psychotic disorders could be explained by some comorbidities such as substance use disorders which are frequently associated with both conditions and could act as mediator in the genesis of psychotic disorders following ADHD during childhood. Our review also focuses on an epidemiological bias that could be found in some studies such as possible diagnostic errors, as some non-specific clinical signs could be found in both late diagnosed ADHD and in “at risk mental state” of psychosis. ConclusionADHD and psychotic disorders share common risk factors, neurobiological pathways and clinical symptoms. Perspectives for future studies are proposed considering a dimensional aspect of psychiatric disorders using, for example, Research Domain Criteria and exploring the link between the two conditions.

Addressing Sex and Gender Bias in COVID-19 Epidemiology in Bangladesh

Sex and gender are recognized as a vital factor in the infectious disease epidemiology, and disease outcomes, and these two factors also appear to hold for severe acute respiratory syndrome (SARS-CoV-2) infection. Evidence from COVID-19 infection in Bangladesh showed that a variation number of cases and deaths between the sex differences in immune responses to viruses and gender related risk factors among the male and female, but more severe outcomes in aged men of cases and deaths in 31~40 and 61~70 years respectively than female. However, the previous research dataset in the different parts of world evidenced that the men's are vulnerable compare to women's. Similar trends to observe in Bangladesh datasets in COVID-19 epidemiology. Male and female represented 71% and 29% of total reported confirmed COVID-19 cases in Bangladesh, respectively. Moreover, different levels of angiotensin converting enzyme 2 (ACE2) and transmembrane protease serine 2 (TMRSS2) enzymes play important role in COVID-19 infection of men and women. In addition, the effects of testosterone on ACE2 levels and the presence of ACE2 genes on the specific X-chromosome should not be ignored. In fine, this mini-review focuses on sex and gender variations in patients with an infectious disease of COVID-19 epidemiology in Bangladesh. The information of sex differences and age dependence might contribute to proper management and treatment of COVID-19 in Bangladesh.

Educational note: addressing special cases of bias that frequently occur in perinatal epidemiology

The epidemiologic study of pregnancy and birth outcomes may be hindered by several unique and challenging issues. Pregnancy is a time-limited period in which severe cohort attrition takes place between conception and birth and adverse outcomes are complex and multi-factorial. Biases span those familiar to epidemiologists: selection, confounding and information biases. Specific challenges include conditioning on potential intermediates, how to treat race/ethnicity, and influential windows of prolonged, seasonal and potentially time-varying exposures. Researchers studying perinatal outcomes should be cognizant of the potential pitfalls due to these factors and address their implications with respect to formulating questions of interest, choice of an appropriate analysis approach and interpretations of findings given assumptions. In this article, we catalogue some of the more important potential sources of bias in perinatal epidemiology that have more recently gained attention in the literature, provide the epidemiologic context behind each issue and propose practices for dealing with each issue to the extent possible.

Safety of Prescribing Off-Label Drugs for Noncritical Ill Children: A Cross-Sectional Study.

“Off-label” (OL) refers to the use of a drug for indications and/or conditions different from those for which it was licensed.1 However, off-label drug use (OLDU) is not necessarily incorrect. The Food and Drug Administration stated that “Good medical practice and the best interests of the patient require that physicians use legally available drugs ... according to their best knowledge and judgment.”2 Prescribing OL drugs is a common practice. Because of limited pharmacologic research in children, it is more frequent in pediatric care,3 particularly in hospitalized children.4 Adverse drug reactions (ADRs)5 occur with OLDU at frequencies between 23% and 60%, depending on the type of drug and the group of patients considered6; however, information on this topic is limited, particularly in pediatrics, and in many cases the detection of ADRs is based on different definitions and methodologies.7,8 A standardized procedure for searching ADRs can achieve more accurate results.9 There is an urgent need to better understand OLDU, particularly in Latin America, where economic, social, and cultural reasons may guide different drug choices from developed countries and possibly tip the balance toward more OL prescribing in pediatrics. We estimate the prevalence of OLDU and of ADRs in hospitalized children. This was a cross-sectional study that included patients aged <18 years, hospitalized at a pediatric tertiary care hospital, during 2017. We include one whole year to avoid epidemiological bias related to different cause of hospitalization in children across the year (ie, respiratory infection …

Simulated 2017 nationwide sampling at 13,940 major U.S. sewage treatment plants to assess seasonal population bias in wastewater-based epidemiology

Wastewater-based epidemiology (WBE) is an economical technique for monitoring and managing the health and behavior of human populations. Using 2017 nationwide data on geospatial population demographics as a test case, we simulated repeated sampling at all major U.S. wastewater treatment plants (WWTPs; n = 13,940) under constant biomarker loading conditions, to explore the potential sensitivity of WBE for generating skewed data. Simulation of repeated sewage sampling over all four seasons of 2017 yielded a number of expected, inter-dependent phenomena triggered by cooler wintertime temperatures compared to summertime results, including relatively (i) slower in-sewer biomarker decay, (ii) longer distal reach of WBE, (iii) larger effective sewershed monitoring areas, and (iv) an increase in the population represented. Additional important but not necessarily anticipated simulation outcomes included (v) distinct, non-random changes in demographic parameters of monitored subpopulations (e.g., by household income, educational attainment, military service, unemployment, and lack of health insurance), (vi) recurring observation of the latter demographic patterns across various geospatial scales and regions, and (vii) more evenly distributed results in the winter. In contrast, data obtainable by WBE in the summertime were dominated by households residing closest to the WWTP and subpopulations of relatively lesser wealth, educational achievement, healthcare access and employability. The analytical approach presented here should be readily applicable to other regions worldwide and may help to improve the design, robustness and interpretation of future WBE studies.

PNS10 DEFINING CONFOUNDING BY INDICATION: RESOLVING THE GAP BETWEEN AN EPIDEMIOLOGIST AND ECONOMETRICIAN

In Health Economics and Outcomes Research most statistical analyses are embedded in the field of epidemiology or econometrics. However, there is still a gap between the two fields in terms of the used key terminology for statistical concepts. Amongst others, this gap is reflected in the different interpretations of the concept of confounding, and related terms such as selection bias. The problem of confounding appears often in biomedical research. Econometricians often refer to this phenomena as an endogeneity problem, omitted variable bias or unobserved heterogeneity, whereas epidemiologists typically use the term confounding by indication. Furthermore, if we compare the use of the term selection bias within the two disciplines, we can see that selection bias in epidemiology is most commonly used for describing the situation when the study population is not comparable to the source population. For example, if randomized controlled trial participants are healthier, the observed sample is not representative of the source population. In econometrics, however, the term selection bias is mainly used when the treatment uptake is non-random due to self-selection of individuals. This may in turn result in self selection bias when there is a comparison between hospitalized and non-hospitalized individuals without taking into account that hospitalized individuals tend to be less healthy. This poster presentation aims to shed light on the fine differences in terminology for confounding, and its associated terms, within and between each field. We provide examples in order to emphasize the minor differences between apparently similar concepts. We hope that by discerning the difference in approaching methodological issues between the disciplines, a better communication and cooperation between them is stimulated.

Gender differences in anaphylaxis.

Is sexual dimorphism also true in anaphylaxis as described in other allergic diseases? Possible gender differences in the epidemiology, triggers, severity, outcomes of anaphylaxis as well as in the pathogenesis of the disease are discussed. Hormonal status and the X-chromosome-coded factors deeply involved in the regulation of T-cell and B-cell responses may influence the gender difference noticed in allergic diseases, such as asthma and rhinitis. Little is known if sex is also relevant for anaphylaxis, although the description of catamenial anaphylaxis is intriguing. However, epidemiologic bias, lack of reliable animal models for the human disease, differences into diagnostic codes and not harmonized clinical grading unfortunately represent hurdles to obtain meaningful information on this topic. The female sex predisposes to a dysregulation of the immune response as suggested by the increased prevalence of autoimmunity and atopy. In anaphylaxis, pathomechanisms are not fully disclosed, triggers are numerous and IgE-dependent mast cell degranulation only represents a part of the story. Improvement into the definition of the disease including a more careful coding system and better investigations about triggers seem the only way to allow a more precise assessment of the possible different risk for women to develop anaphylaxis.

Immortal Time Bias in Epidemiology

Immortal time occurs when study subjects’ person-time is misclassified. For example, if exposure is assigned over time, but treated as a binary “ever-exposed” variable, subjects in the exposed group are “immortal” prior to their exposure. We describe immortal time and the context in which it introduces bias and describe several approaches to avoid immortal time bias via design or mitigate it through analysis. Several authors have described examples of immortal time bias in clinical epidemiology, pharmacoepidemiology, and perinatal epidemiology. Solutions to immortal time bias include analyses that appropriately account for time-varying exposure, and design solutions that align exposure with the start of follow-up. Immortal time bias is pervasive in epidemiology. It can cause substantial bias. It is, however, easily avoided and can be controlled using appropriate analytic and design strategies.