Explore the vast range of titles available.

Background and Aims Considering the opposite outcome—for example, survival instead of death—may affect conclusions about which subpopulation benefits more from a treatment or suffers more from an exposure. Methods For case studies on death following COVID‐19 and bankruptcy following melanoma, we compute and interpret the relative risk, odds ratio, and risk difference for different age groups. Since there is no established effect measure or outcome for either study, we redo these analyses for survival and solvency. Results In a case study on COVID‐19 that ignores confounding, the relative risk of death suggested that 40–49‐year‐old Mexicans with COVID‐19 suffered more from their unprepared healthcare system, using Italy's system as a baseline, than their 60–69‐year‐old counterparts. The relative risk of survival and the risk difference suggested the opposite conclusion. A similar phenomenon occurred in a case study on bankruptcy following melanoma treatment. Conclusion To increase transparency around this paradox, researchers reporting one outcome should note if considering the opposite outcome would yield different conclusions. When possible, researchers should also report or estimate underlying risks alongside effect measures.

In clinical trials, the relative risk or risk ratio (RR) is a mainstay of reporting of the effect magnitude for an intervention. The RR is the ratio of the probability of an outcome in an intervention group to its probability in a control group. Thus, the RR provides a measure of change in the likelihood of an event linked to a given intervention. This measure has been widely used because it is today considered a measure with “portability” across varying outcome prevalence, especially when the outcome is rare. It turns out, however, that there is a much more important problem with this ratio, and this paper aims to demonstrate this problem. We used mathematical derivation to determine if the RR is a measure of effect magnitude alone (i.e., a larger absolute value always indicating a stronger effect) or not. We also used the same derivation to determine its relationship to the prevalence of an outcome. We confirm the derivation results with a follow-up analysis of 140,620 trials scraped from the Cochrane. We demonstrate that the RR varies for reasons other than the magnitude of the effect because it is a ratio of two posterior probabilities, both of which are dependent on baseline prevalence of an outcome. In addition, we demonstrate that the RR shifts toward its null value with increasing outcome prevalence. The shift toward the null happens regardless of the strength of the association between intervention and outcome. The odds ratio (OR), the other commonly used ratio, measures solely the effect magnitude and has no relationship to the prevalence of an outcome in a study nor does it overestimate the RR as is commonly thought. The results demonstrate the need to (1) end the primary use of the RR in clinical trials and meta-analyses as its direct interpretation is not meaningful, (2) replace the RR by the OR, and (3) only use the postintervention risk recalculated from the OR for any expected level of baseline risk in absolute terms for purposes of interpretation such as the number needed to treat. These results will have far-reaching implications such as reducing misleading results from clinical trials and meta-analyses and ushering in a new era in the reporting of such trials or meta-analyses in practice.

The standardized mean difference (SMD) can be calculated from different mean differences (MDs) and standard deviations (SDs). This study aims to investigate how clinical trials calculated, reported and interpreted the SMD, and to examine the variation between different SMDs. We searched the PubMed for randomized controlled trials of general medicine and psychiatry that estimated SMDs. We explored how the SMD was computed and interpreted. We calculated SMDs based on different MDs and SDs, and the variation in these SMD estimates for each study. We included 161 articles. Various MDs and SDs were used to calculate SMDs, yet 69.0% studies failed to provide sufficient details. Variations in SMD estimates using different MDs and SDs in one study could be substantial (median of the absolute differences was 0.3, interquartile range IQR 0.17 to 0.53). However, 68.3% studies interpreted the SMD based on the same reference, Cohen's rule of thumb. The largest variations were observed in studies with small sample sizes and large reported effects. SMDs using different MDs and SDs could vary considerably, but the report was often insufficient and the interpretation was oversimplified. To avoid selective reporting bias and misinterpretation, prespecifying and reporting the method and interpreting the result from multiple perspectives are desirable.

In a recent paper we suggest that the relative risk (RR) be replaced with the odds ratio (OR) as the effect measure of choice in clinical epidemiology. In response, Chu, and colleagues raise several points that argue for the status quo. In this paper, we respond to their response. We use the same examples given by Chu and colleagues to recompute estimates of effect and demonstrate the problem with the RR. We reaffirm the following findings: a) the OR and RR measure different things and their numerical difference is only important if misinterpreted b) this potential misinterpretation is a trivial issue compared to the lack of portability of the RR c) the same examples reaffirm non-portability of the RR and demonstrate how misleading the results might be in contrast to the OR, which is independent of the baseline risk d) the concept of non-collapsibility for the OR should be expected in the presence of a non-confounding risk factor, and is not a bias e) the log link in regression models that generate RRs as well as the use of RRs in meta-analysis is shown to be problematic using the same examples. The OR should replace the RR in clinical research and meta-analyses though there should be conversion of the end product into ratios or differences of risk, solely, for interpretation. To this end we provide a Stata module (logittorisk) for this purpose.

This article is to establish recommendations for conducting quantitative synthesis, or meta-analysis, using study-level data in comparative effectiveness reviews (CERs) for the Evidence-based Practice Center (EPC) program of the Agency for Healthcare Research and Quality. We focused on recurrent issues in the EPC program and the recommendations were developed using group discussion and consensus based on current knowledge in the literature. We first discussed considerations for deciding whether to combine studies, followed by discussions on indirect comparison and incorporation of indirect evidence. Then, we described our recommendations on choosing effect measures and statistical models, giving special attention to combining studies with rare events; and on testing and exploring heterogeneity. Finally, we briefly presented recommendations on combining studies of mixed design and on sensitivity analysis. Quantitative synthesis should be conducted in a transparent and consistent way. Inclusion of multiple alternative interventions in CERs increases the complexity of quantitative synthesis, whereas the basic issues in quantitative synthesis remain crucial considerations in quantitative synthesis for a CER. We will cover more issues in future versions and update and improve recommendations with the accumulation of new research to advance the goal for transparency and consistency.

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.

Climate change has been widely conceptualized as a threat multiplier of pre-existing disadvantage, yet quantitative research has lacked a structured framework for empirically testing how climate-related exposures amplify disadvantage–outcome relationships. We present a theoretical framework and diagram adapted from epidemiologic thinking that outline key analytical pathways for examining climate change as a threat multiplier. By translating this concept into epidemiologic terms, this framework enables systematic testing of threat multiplication, moving beyond approaches that focus solely on heterogeneity in the direct effects of climate exposures on an outcome. In doing so, it provides a more precise and operationalizable approach for advancing quantitative climate justice research, including more targeted and equity-oriented intervention and policy strategies.

Background Despite many efforts, preterm birth (PTB) is poorly understood and remains a major public health problem in the United States. Toxicological work suggests gestational parent (GP) diet may modify the effect of ambient pollutants on birth outcomes. We assessed risk of PTB in humans in relation to fine particulate matter (PM 2.5 ), ozone (O 3 ), and nitrogen dioxide (NO 2 ) and variation by diet. Methods 684 GP-singleton infant pairs in the Newborn Epigenetics Study prospective birth cohort were attributed ambient air pollutant exposures for each trimester based on residence. Total energy intake, percent of energy intake from saturated fat, and percent of energy intake from total fat were dichotomized at the 75th percentile. >We used log binomial regressions to estimate risk ratios (RR (95%CI)) for PTB by pollutant interquartile ranges, adjusting for GP age, pre-pregnancy body mass index, GP race/ethnicity, GP education, season of conception, household income, and each diet factor. We assessed departure from additivity using interaction contrast ratios (ICRs). We addressed missing covariate data with multiple imputation. Results Point estimates suggest that O 3 may be inversely associated with PTB when exposure occurs in trimester 2 (min RR: 0.77, 95% CI: 0.39, 1.49), but may be harmful when exposure occurs in trimester 3 (max RR: 1.51, 95% CI: 0.62, 3.64). Additionally, PM 2.5 may be inversely associated with PTB when considered with total fat and saturated fat in trimester 2. Imprecise ICRs suggest departure from additivity (evidence of modification) with some pollutant-diet combinations. Conclusions While confidence intervals are wide, we observed potential modification of pollutant associations by dietary factors. It is imperative that large cohorts collect the required data to examine this topic, as more power is necessary to investigate the nuances suggested by this work.

We are often confronted in public health by associations that vary by population or subpopulation. Much effort has focused on the statistical and biological interpretation of such effect measure modification (EMM) because of the importance to public health. However, EMM remains difficult to conceptualize because it apparently violates everyday understanding of causes as usually acting consistently, making it difficult to predict when EMM may occur and raises questions about how to determine the external validity of interventions without extensive retesting by population or subpopulation. Methods exposition. We propose that EMM can be thought of as mediation of an intervention (or exposure) on outcome by mechanism(s) whose relevance differs between population groups, which can be illustrated in causal diagrams, that is, directed acyclic graphs and selection diagrams, meaning external validity can be formally considered as mediation according to “selection variables.” EMM can then be represented graphically and its consequences predicted. This new conceptualization of EMM transforms EMM from a concept that violates everyday understanding of causes into an insight generating means of thinking about interventions (or exposures) in terms of their mediating mechanism(s) and corresponding population- or subpopulation-specific attributes to help target interventions effectively.

Background Studies included in a meta-analysis are often heterogeneous. The traditional random-effects models assume their true effects to follow a normal distribution, while it is unclear if this critical assumption is practical. Violations of this between-study normality assumption could lead to problematic meta-analytical conclusions. We aimed to empirically examine if this assumption is valid in published meta-analyses. Methods In this cross-sectional study, we collected meta-analyses available in the Cochrane Library with at least 10 studies and with between-study variance estimates > 0. For each extracted meta-analysis, we performed the Shapiro–Wilk (SW) test to quantitatively assess the between-study normality assumption. For binary outcomes, we assessed between-study normality for odds ratios (ORs), relative risks (RRs), and risk differences (RDs). Subgroup analyses based on sample sizes and event rates were used to rule out the potential confounders. In addition, we obtained the quantile–quantile (Q–Q) plot of study-specific standardized residuals for visually assessing between-study normality. Results Based on 4234 eligible meta-analyses with binary outcomes and 3433 with non-binary outcomes, the proportion of meta-analyses that had statistically significant non-normality varied from 15.1 to 26.2%. RDs and non-binary outcomes led to more frequent non-normality issues than ORs and RRs. For binary outcomes, the between-study non-normality was more frequently found in meta-analyses with larger sample sizes and event rates away from 0 and 100%. The agreements of assessing the normality between two independent researchers based on Q–Q plots were fair or moderate. Conclusions The between-study normality assumption is commonly violated in Cochrane meta-analyses. This assumption should be routinely assessed when performing a meta-analysis. When it may not hold, alternative meta-analysis methods that do not make this assumption should be considered.