Explore the vast range of titles available.

Multiple tests arise frequently in epidemiologic research. However, the issue of multiplicity adjustment is surrounded by confusion and controversy, and there is no uniform agreement on whether or when adjustment is warranted. In this paper we compare frequentist and Bayesian frameworks for multiple testing. We argue that the frequentist framework leads to logical difficulties, and is unable to distinguish between relevant and irrelevant multiplicity adjustments. We further argue that these logical difficulties resolve within the Bayesian framework, and that the Bayesian framework makes a clear and coherent distinction between relevant and irrelevant adjustments. We use Directed Acyclic Graphs to illustrate the differences between the two frameworks, and to motivate our arguments.

When studying the association between an exposure and an outcome, it is common to use regression models to adjust for measured confounders. The most common models in epidemiologic research are logistic regression and Cox regression, which estimate conditional (on the confounders) odds ratios and hazard ratios. When the model has been fitted, one can use regression standardization to estimate marginal measures of association. If the measured confounders are sufficient for confounding control, then the marginal association measures can be interpreted as poulation causal effects. In this paper we describe a new R package, stdReg, that carries out regression standardization with generalized linear models (e.g. logistic regression) and Cox regression models. We illustrate the package with several examples, using real data that are publicly available.

The Kaplan–Meier curve is a standard statistical tool that is used in cohort studies to illustrate how survival during follow-up depends on time-fixed covariates that are measured at baseline. For time-varying covariates, an extended Kaplan–Meier curve has been proposed that is constructed by letting subjects move across risk sets as their covariate levels change during follow-up. It has been claimed, but not proven, that, under a particular independence assumption, this extended Kaplan–Meier curve has a causal interpretation as representing a hypothetical cohort whose covariate values remain constant during follow-up. In this note, we show that, in the absence of confounding, this claim is indeed correct. However, we argue that the causal implications of this independence assumptions are highly unrealistic, and that a causal interpretation of the extended Kaplan–Meier curve is therefore typically unwarranted.

Use of ADHD Medication and Criminality Whether pharmacologic treatment for attention deficit–hyperactivity disorder reduces the risk of criminality is not known. In this observational study, patients with ADHD were less likely to be convicted of a crime during periods when they were taking an ADHD medication. About 5% of all children in the Western world fulfill diagnostic criteria for attention deficit–hyperactivity disorder (ADHD), 1 and a large proportion of such children are treated pharmacologically. 2 ADHD has been associated with criminality 3 , 4 and externalizing disorders. 5 Beneficial short-term effects of ADHD medication on symptoms of ADHD and associated conduct problems have been shown in numerous randomized, controlled studies involving children 6 – 8 and adults. 9 – 11 ADHD symptoms are largely persistent from childhood into adulthood, 12 but one prominent feature of ADHD treatment is that the discontinuation of medication is common, 13 , 14 especially in adolescence and early adulthood. 15 The importance of treatment . . .

Measures of causal effects play a central role in epidemiology. A wide range of measures exist, which are designed to give relevant answers to substantive epidemiological research questions. However, due to mathematical convenience and software limitations most studies only report odds ratios for binary outcomes and hazard ratios for time-to-event outcomes. In this paper we show how logistic regression models and Cox proportional hazards regression models can be used to estimate a wide range of causal effect measures, with the R-package stdReg. For illustration we focus on the attributable fraction, the number needed to treat and the relative excess risk due to interaction. We use two publicly available data sets, so that the reader can easily replicate and elaborate on the analyses. The first dataset includes information on 487 births among 188 women, and the second dataset includes information on 2982 women diagnosed with primary breast cancer.

The sibling comparison design is an important epidemiologic tool to control for unmeasured confounding, in studies of the causal effect of an exposure on an outcome. It is routinely argued that within-family associations are automatically controlled for all measured and unmeasured covariates that are shared (constant) within sets of siblings, such as early childhood environment and parental genetic makeup. However, an important lesson from modern causal inference theory is that not all types of covariate control are desirable. In particular, it has been argued that collider control always leads to bias, and that mediator control may or may not lead to bias, depending on the research question. In this article, we use directed acyclic graphs (DAGs) to distinguish between shared confounders, shared mediators and shared colliders, and we examine which of these shared covariates the sibling comparison design really controls for.

A convenient way of dealing with confounding is the sibling comparison design, where the outcome in exposed individuals is compared with the outcome in their unexposed siblings. The standard analysis of sibling comparison designs assumes that the exposure and outcome of an individual do not affect the exposure and outcome of his/her siblings, sometimes referred to as an absence of sibling carryover or contagion effects. Unfortunately, there are many situations where carryover effects are likely to be present. In this article, we explore the consequences of carryover effects for sibling comparison designs. We show, using causal diagrams, when and why carryover effects lead to bias, and we investigate the sign and magnitude of this bias under various scenarios.

The attributable fraction (or attributable risk) is a widely used measure that quantifies the public health impact of an exposure on an outcome. Even though the theory for AF estimation is well developed, there has been a lack of up-to-date software implementations. The aim of this article is to present a new R package for AF estimation with binary exposures. The package AF allows for confounder-adjusted estimation of the AF for the three major study designs: cross-sectional, (possibly matched) case-control and cohort. The article is divided into theoretical sections and applied sections. In the theoretical sections we describe how the confounder-adjusted AF is estimated for each specific study design. These sections serve as a brief but self-consistent tutorial in AF estimation. In the applied sections we use real data examples to illustrate how the AF package is used. All datasets in these examples are publicly available and included in the AF package, so readers can easily replicate all analyses.

There is concern that selective serotonin reuptake inhibitor (SSRI) treatment may increase the risk of suicide attempts or deaths, particularly among children and adolescents. However, debate remains regarding the nature of the relationship. Using nationwide Swedish registers, we identified all individuals aged 6–59 years with an incident SSRI dispensation (N = 538,577) from 2006 to 2013. To account for selection into treatment, we used a within-individual design to compare the risk of suicide attempts or deaths (suicidal behaviour) in time periods before and after SSRI-treatment initiation. Within-individual incidence rate ratios (IRRs) of suicidal behaviour were estimated. The 30 days before SSRI-treatment initiation was associated with the highest risk of suicidal behaviour compared with the 30 days 1 year before SSRI initiation (IRR = 7.35, 95% CI 6.60–8.18). Compared with the 30 days before SSRI initiation, treatment periods after initiation had a reduced risk—the IRR in the 30 days after initiation was 0.62 (95% CI 0.58–0.65). The risk then declined over treatment time. These patterns were similar across age strata, and when stratifying on history of suicide attempts. Initiation with escitalopram was associated with the greatest risk reduction, though CIs for the IRRs of the different SSRI types were overlapping. The results do not suggest that SSRI-treatment increases the risk for suicidal behaviour in either youths or adults; rather, it may reduce the risk. Further research with different study designs and in different populations is warranted.

Sibling comparison studies have the attractive feature of being able to control for unmeasured confounding by factors that are shared within families. However, there is sometimes a concern that these studies may have poor generalizability (external validity) due to the implicit restriction to families that are covariate-discordant, i.e., those families where at least two siblings have different levels of at least one of the covariates (exposure or confounders) under investigation. Even if this selection mechanism has been noted by many authors, previous accounts of the problem tend to be brief. The purpose of this paper is to provide a formal discussion of the implicit restriction to covariate-discordant families in sibling comparison studies. We discuss when and how this restriction may impair the generalizability of the study, and we show that a similar generalizability problem may in fact arise even when all families are covariate-discordant, e.g. even if the exposure is continuous so that all siblings have different exposure levels. We show how this problem can be solved by using a so-called marginal between-within model for estimation of marginal exposure effects. Finally, we illustrate the theoretical conclusions with a simulation study.