E-value: a superior alternative to P-value and its adjustments in DNA methylation studies.

DNA methylation plays a crucial role in transcriptional regulation. Reduced representation bisulfite sequencing (RRBS) is a technique of increasing use for analyzing genome-wide methylation profiles. Many computational tools such as Metilene, MethylKit, BiSeq and DMRfinder have been developed to use RRBS data for the detection of the differentially methylated regions (DMRs) potentially involved in epigenetic regulations of gene expression. For DMR detection tools, as for countless other medical applications, P-values and their adjustments are among the most standard reporting statistics used to assess the statistical significance of biological findings. However, P-values are coming under increasing criticism relating to their questionable accuracy and relatively high levels of false positive or negative indications. Here, we propose a method to calculate E-values, as likelihood ratios falling into the null hypothesis over the entire parameter space, for DMR detection in RRBS data. We also provide the R package 'metevalue' as a user-friendly interface to implement E-value calculations into various DMR detection tools. To evaluate the performance of E-values, we generated various RRBS benchmarking datasets using our simulator 'RRBSsim' with eight samples in each experimental group. Our comprehensive benchmarking analyses showed that using E-values not only significantly improved accuracy, area under ROC curve and power, over that of P-values or adjusted P-values, but also reduced false discovery rates and type I errors. In applications using real RRBS data of CRL rats and a clinical trial on low-salt diet, the use of E-values detected biologically more relevant DMRs and also improved the negative association between DNA methylation and gene expression.

A controlled effects approach to assessing immune correlates of protection.

An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.

Assessment of the E-value in the presence of bias amplification: a simulation study.

BACKGROUND: The E-value, a measure that has received recent attention in the comparative effectiveness literature, reports the minimum strength of association between an unmeasured confounder and the treatment and outcome that would explain away the estimated treatment effect. This study contributes to the literature on the applications and interpretations of E-values by examining how the E-value is impacted by data with varying levels of association of unobserved covariates with the treatment and outcome measure when covariate adjustment is applied. We calculate the E-value after using regression and propensity score methods (PSMs) to adjust for differences in observed covariates. Propensity score methods are a common observational research method used to balance observed covariates between treatment groups. In practice, researchers may assume propensity score methods that balance treatment groups across observed characteristics will extend to balance of unobserved characteristics. However, that assumption is not testable and has been shown to not hold in realistic data settings. We assess the E-value when covariate adjustment affects the imbalance in unobserved covariates. METHODS: Our study uses Monte Carlo simulations to evaluate the impact of unobserved confounders on the treatment effect estimates and to evaluate the performance of the E-Value sensitivity test with the application of regression and propensity score methods under varying levels of unobserved confounding. Specifically, we compare observed and unobserved confounder balance, odds ratios of treatment vs. control, and E-Value sensitivity test statistics from generalized linear model (GLM) regression models, inverse-probability weighted models, and propensity score matching models, over correlations of increasing strength between observed and unobserved confounders. RESULTS: We confirm previous findings that propensity score methods - matching or weighting - may increase the imbalance in unobserved confounders. The magnitude of the effect depends on the strength of correlation between the confounder, treatment, and outcomes. We find that E-values calculated after applying propensity score methods tend to be larger when unobserved confounders result in more biased treatment effect estimates. CONCLUSIONS: The E-Value may misrepresent the size of the unobserved effect needed to change the magnitude of the association between treatment and outcome when propensity score methods are used. Thus, caution is warranted when interpreting the E-Value in the context of propensity score methods.

Trends in Apgar scores and umbilical artery pH: a population-based cohort study on 10,696,831 live births in Germany, 2008-2022.

Low Apgar scores and low umbilical arterial (UA) blood pH are considered indicators of adverse perinatal events. This study investigated trends of these perinatal health indicators in Germany. Perinatal data on 10,696,831 in-hospital live births from 2008 to 2022 were obtained from quality assurance institutes. Joinpoint regression analysis was used to quantify trends of low Apgar score and UA pH. Additional analyses stratified by mode of delivery were performed on term singletons with cephalic presentation. Robustness against unmeasured confounding was analyzed using the E-value sensitivity analysis. The overall rates of 5-min Apgar scores < 7 and UA pH < 7.10 in liveborn infants were 1.17% and 1.98%, respectively. For low Apgar scores, joinpoint analysis revealed an increase from 2008 to 2011 (annual percent change (APC) 5.19; 95% CI 3.66-9.00) followed by a slower increase from 2011 to 2019 (APC 2.56; 95% CI 2.00-3.03) and a stabilization from 2019 onwards (APC - 0.64; 95% CI - 3.60 to 0.62). The rate of UA blood pH < 7.10 increased significantly between 2011 and 2017 (APC 5.90; 95% CI 5.15-7.42). For term singletons in cephalic presentation, the risk amplification of low Apgar scores was highest after instrumental delivery (risk ratio 1.623, 95% CI 1.509-1.745), whereas those born spontaneous had the highest increase in pH < 7.10 (risk ratio 1.648, 95% CI 1.615-1.682).          Conclusion: Rates of low 5-min Apgar scores and UA pH in liveborn infants increased from 2008 to 2022 in Germany. What is Known: • Low Apgar scores at 5 min after birth and umbilical arterial blood pH are associated with adverse perinatal outcomes. • Prospective collection of Apgar scores and arterial blood pH data allows for nationwide quality assurance. What is New: • The rates of liveborn infants with 5-min Apgar scores < 7 rose from 0.97 to 1.30% and that of umbilical arterial blood pH < 7.10 from 1.55 to 2.30% between 2008-2010 and 2020-2022. • In spontaneously born term singletons in cephalic presentation, the rate of metabolic acidosis with pH < 7.10 and BE < -5 mmol/L in umbilical arterial blood roughly doubled between the periods 2008-2010 and 2020-2022.

A controlled effects approach to assessing immune correlates of protection.

An immune correlate of risk (CoR) is an immunologic biomarker in vaccine recipients associated with an infectious disease clinical endpoint. An immune correlate of protection (CoP) is a CoR that can be used to reliably predict vaccine efficacy (VE) against the clinical endpoint and hence is accepted as a surrogate endpoint that can be used for accelerated approval or guide use of vaccines. In randomized, placebo-controlled trials, CoR analysis is limited by not assessing a causal vaccine effect. To address this limitation, we construct the controlled risk curve of a biomarker, which provides the causal risk of an endpoint if all participants are assigned vaccine and the biomarker is set to different levels. Furthermore, we propose a causal CoP analysis based on controlled effects, where for the important special case that the biomarker is constant in the placebo arm, we study the controlled vaccine efficacy curve that contrasts the controlled risk curve with placebo arm risk. We provide identification conditions and formulae that account for right censoring of the clinical endpoint and two-phase sampling of the biomarker, and consider G-computation estimation and inference under a semiparametric model such as the Cox model. We add modular approaches to sensitivity analysis that quantify robustness of CoP evidence to unmeasured confounding. We provide an application to two phase 3 trials of a dengue vaccine indicating that controlled risk of dengue strongly varies with 50$\%$ neutralizing antibody titer. Our work introduces controlled effects causal mediation analysis to immune CoP evaluation.

How to Choose between Different Bayesian Posterior Indices for Hypothesis Testing in Practice.

Hypothesis testing is an essential statistical method in experimental psychology and the cognitive sciences. The problems of traditional null hypothesis significance testing (NHST) have been discussed widely, and among the proposed solutions to the replication problems caused by the inappropriate use of significance tests and p-values is a shift toward Bayesian data analysis. However, Bayesian hypothesis testing is concerned with various posterior indices for significance and the size of an effect. This complicates Bayesian hypothesis testing in practice, as the availability of multiple Bayesian alternatives to the traditional p-value causes confusion which one to select and why. In this paper, various Bayesian posterior indices which have been proposed in the literature are compared and their benefits and limitations are discussed. The comparison shows that conceptually not all proposed Bayesian alternatives to NHST and p-values are beneficial, and the usefulness of some indices strongly depends on the study design and research goal. However, the comparison also reveals that there exist at least two candidates among the available Bayesian posterior indices which have appealing theoretical properties and are widely underused in the cognitive sciences.

Timely treatment and higher compliance to traditional Chinese medicine: New influencing factors for reducing severe COVID-19 based on retrospective cohorts in 2020 and 2021.

OBJECTIVE: From the beginning of 2020, our study team formulated a traditional Chinese medicine (TCM) prescription named Sanhanhuashi formula (SHHS) to treat COVID-19 patients. Then we conducted several studies to explore the effectiveness of SHHS formula and other influencing factors on prognosis of disease. The purpose of this study was to describe the trends of patients' characters from 2020 to 2021 based on two separate cohorts, and to explore the influencing factors on incidence of severe COVID-19 conditions, especially the contributions of timely treatment and higher compliance to SHHS formula. METHODS: A retrospective cohort study was conducted in Wuhan, Hubei province and Tonghua, Jilin province. Participants were hospitalized mild to moderate COVID-19 consecutive enrolled patients in Wuhan hospital of traditional Chinese and western medicine (from Feb 13, 2020 to March 8, 2020) and Tonghua central hospital (from Jan 17, 2021 to Feb 5, 2021). Age, sex, time waiting to be hospitalized, medical history, initial symptoms, concomitant medication, and severity of disease were collected. Univariate and multivariate logistic regression were used to explore the associations between various exposures and the outcome, ie. the proportion of patients who were converted to severe status. E-values and its lower control limit (LCL) were calculated for sensitivity analysis. RESULTS: Totally, 176 COVID-19 patients in two hospitals were enrolled. 81 patients were from Wuhan hospital of traditional Chinese and western medicine and 95 from Tonghua central hospital. 42 patients used SHHS formula arrival or exceed 7 days, and 2 (4.8%) progressed to severe condition. Among 134 patients who were exposed SHHS less than 7 days, 18 (13.4%) were converted to severe situation. Compared with those diagnosed in 2020, cases in 2021 were characterized as lower rates of initial symptoms (88.9% vs 35.8%, P < 0.001) and concomitant medications ever widely used, eg. antiviral medicine (71.6% vs 43.2%, P < 0.001), antibiotics (61.7% vs 13.7%, P < 0.001) and Chinese patent medicine (76.5% vs 44.2%, P < 0.001). They also waited less time for hospitalization (median: 12 vs 2 days, P < 0.001). The final multivariate logistic regression model showed that age (> 60 yrs) (OR: 3.943; 95% CI: 1.402-11.086; P = 0.009; E-value = 7.35, LCL:2.15), diagnosis year (OR: 0.165; 95% CI: 0.050-0.551; P = 0.003; E-value=11.6, LCL: 3.03) and SHHS exposure (OR: 0.118; 95% CI: 0.014-0.992; P = 0.049; E-value = 16.43, LCL:1.1) were independent risk factors for predicting severe status. CONCLUSIONS: The profile of COVID-19 patients has changed after one year. In addition to age, diagnosis year and SHHS exposure are two new factors to predict the prognosis of disease. The patients diagnosed in 2021 were mainly benefited from timely treatment. Subsequently, adhere to use SHHS formula a quite longer time reduced the number of severe cases. Therefore, both the current epidemic prevention and control measures and increasing compliance to traditional Chinese medicine are effective ways to reducing severe cases and improving public health.

Risk of large for gestational age births at early, full and late term in relation to pre-pregnancy body mass index: Mediation by gestational diabetes status.

BACKGROUND: Maternal pre-pregnancy body mass index (BMI) is strongly associated with infant birthweight and the risk differs in pregnancies complicated by gestational diabetes (GDM). OBJECTIVES: To examine the risk of large for gestational age (LGA) (≥97th percentile) singleton births at early term, full term and late term in relation to maternal pre-pregnancy BMI status mediated through GDM. METHODS: We analysed data from the 2018 U.S. National Vital Statistics Natality File restricted to singleton term births (N = 3,229,783). In counterfactual models for causal inference, we estimated the total effect (TE), natural direct effect (NDE) and natural indirect effect (NIE) for the association of pre-pregnancy BMI with subcategories of LGA births at early, full and late term mediated through GDM, using log-binomial regression and adjusting for race/ethnicity, age, education, parity and infant sex. Proportion mediated was calculated on the risk difference scale and potential unmeasured confounders were assessed using the E-value. RESULTS: Overall, 6.4% of women had GDM, and there were 3.6% LGA singleton term births. The highest prevalence of GDM was among pre-gestational overweight/obesity that also had the highest rates of LGA births at term. The TE estimates for the risk of LGA births were the strongest across women with higher pre-pregnancy BMI compared to women with normal pre-pregnancy BMI. The NDE estimates were higher than the NIE estimates for overweight/obese BMI status. The proportion mediated, which answers the causal question to what extent the total effect of the association between pre-pregnancy BMI and LGA births is accounted for through GDM, was the highest (up to 16%) for early term births. CONCLUSIONS: Term singleton births make up the largest proportion in a cohort of newborns. While the percentage mediated through GDM was relatively small, health risks arising from pre-pregnancy overweight, and obesity can be substantial to both mothers and their offspring.

Low density lipoprotein cholesterol is associated with increased risk of cardiovascular disease in participants over 70 years old: A prospective cohort study.

BACKGROUND AND AIMS: Previous studies have indicated that the association of elevated low-density lipoprotein cholesterol (LDL-C) with cardiovascular disease (CVD) varies greatly with age, with the association being much stronger in younger than older individuals. To estimate the relationship between LDL-C and CVD risk in a contemporary population aged over 70 years in China. METHODS AND RESULTS: In this analysis, participants of China Health and Retirement Longitudinal Study (CHARLS) who did not take statins and did not have heart disease and stroke in 2011 were include and were followed up to 2018. The outcome of this analysis was the occurrence of CVD. Cox regression was used to assess the effect of LDL-C on CVD. We calculated E-values to quantify the effect of unmeasured confounding. In the 9,631 participants, 15.2% (N = 1,463) were aged over 70 years. During follow-up of 7 years, 1,437 participants had a first CVD attack. The Risk of CVD increased with each 10 mg/mL elevation in LDL-C in whole participants and all age groups. We noted a U-shaped relationship between LDL-C and risk of CVD in group over 70 years old, however, we further found that in the left side of U-shape curve, LDL-C was not associated with CVD, which indicated that a lower level of LDL-C could not increase the risk of CVD. E-value analysis suggested robustness to unmeasured confounding. CONCLUSIONS: In a contemporary society of China, elevated the level of LDL-C also increased the risk of CVD in participants over 70 years old. These results should strengthen guideline recommendations for the use of lipid-lowering therapies in those elderly.

fbst: An R package for the Full Bayesian Significance Test for testing a sharp null hypothesis against its alternative via the e value.

Hypothesis testing is a central statistical method in psychology and the cognitive sciences. However, the problems of null hypothesis significance testing (NHST) and p values have been debated widely, but few attractive alternatives exist. This article introduces the fbst R package, which implements the Full Bayesian Significance Test (FBST) to test a sharp null hypothesis against its alternative via the e value. The statistical theory of the FBST has been introduced more than two decades ago and since then the FBST has shown to be a Bayesian alternative to NHST and p values with both theoretical and practical highly appealing properties. The algorithm provided in the fbst package is applicable to any Bayesian model as long as the posterior distribution can be obtained at least numerically. The core function of the package provides the Bayesian evidence against the null hypothesis, the e value. Additionally, p values based on asymptotic arguments can be computed and rich visualizations for communication and interpretation of the results can be produced. Three examples of frequently used statistical procedures in the cognitive sciences are given in this paper, which demonstrate how to apply the FBST in practice using the fbst package. Based on the success of the FBST in statistical science, the fbst package should be of interest to a broad range of researchers and hopefully will encourage researchers to consider the FBST as a possible alternative when conducting hypothesis tests of a sharp null hypothesis.

Effect of maternal smoking during pregnancy on childhood type 1 diabetes: a whole-of-population study.

AIMS/HYPOTHESIS: Evidence of an association between maternal smoking during pregnancy (prenatal smoking) and childhood type 1 diabetes is mixed. Previous studies have been small and potentially biased due to unmeasured confounding. The objectives of this study were to estimate the association between prenatal smoking and childhood type 1 diabetes, assess residual confounding with a negative control design and an E-value analysis, and summarise published effect estimates from a meta-analysis. METHODS: This whole-of-population study (births from 1999 to 2013, participants aged ≤15 years) used de-identified linked administrative data from the South Australian Early Childhood Data Project. Type 1 diabetes was diagnosed in 557 children (ICD, tenth edition, Australian Modification [ICD-10-AM] codes: E10, E101-E109) during hospitalisation (2001-2014). Families not given financial assistance for school fees was a negative control outcome. Adjusted Cox proportional HRs were calculated. Analyses were conducted on complete-case (n = 264,542, type 1 diabetes = 442) and imputed (n = 286,058, type 1 diabetes = 557) data. A random-effects meta-analysis was used to summarise the effects of prenatal smoking on type 1 diabetes. RESULTS: Compared with non-smokers, children exposed to maternal smoking only in the first or second half of pregnancy had a 6% higher type 1 diabetes incidence (adjusted HR 1.06 [95% CI 0.73, 1.55]). Type 1 diabetes incidence was 24% lower (adjusted HR 0.76 [95% CI 0.58, 0.99]) among children exposed to consistent prenatal smoking, and 16% lower for exposure to any maternal smoking in pregnancy (adjusted HR 0.84 [95% CI 0.67, 1.08]), compared with the unexposed group. Meta-analytic estimates showed 28-29% lower risk of type 1 diabetes among children exposed to prenatal smoking compared with those not exposed. The negative control outcome analysis indicated residual confounding in the prenatal smoking and type 1 diabetes association. E-value analysis indicated that unmeasured confounding associated with prenatal smoking and childhood type 1 diabetes, with a HR of 1.67, could negate the observed effect. CONCLUSIONS/INTERPRETATION: Our best estimate from the study is that maternal smoking in pregnancy was associated with 16% lower childhood type 1 diabetes incidence, and some of this effect was due to residual confounding.

Low birthweight in term singletons mediates the association between maternal smoking intensity exposure status and immediate neonatal intensive care unit admission: the E-value assessment.

BACKGROUND: Research investigating the wellbeing of term neonates in the United States is scarce. The objectives of this study were to estimate the prevalence of low birthweight (LBW) and neonatal intensive care unit (NICU) admission among term singletons in association with maternal smoking intensity exposure status, to explore LBW as a mediator linking smoking to immediate newborn NICU transfer/admission, and to assess the potential impact of unmeasured confounding in effect estimates. METHODS: The Natality File of live births registered in the United States in 2016, the first year that all 50 states implemented the revised 2003 standard birth certificate, was restricted to singleton term births (37-41 completed weeks gestation). The prevalence of LBW (< 2500 g) and NICU transfer/admission was estimated across maternal demographic characteristics and smoking intensity status in early and in late pregnancy. Mediation analyses, based on the counterfactual approach, were conducted to examine the total effect (TE), controlled direct effect (CDE), natural direct (NDE) and indirect effects (NIE), and the percentage mediated through LBW. The E-values based on effect size estimates and on lower-bounds of 95% confidence intervals (CIs) assessed the potential impact of unmeasured confounding. RESULTS: Nearly 6.8% of women smoked in early and in late pregnancy, most (36.4%) smoked at high intensity (≥ 10 cigarettes /day) and had the highest prevalence of LBW (6.7%) and NICU transfer/admission (7.0%). For the largest smoking intensity exposure category, the estimate of the TE was 1.68 (95% CI: 1.63, 1.73), of the NDE was 1.56 (95% CI: 1.51, 1.61), of the NIE was 1.08 (95% CI:1.07, 1.09), and the percentage mediated by LBW was 17.6%. The E-values for association estimates and for the lower-bounds of 95% CIs demonstrated the minimum strength of the potential unmeasured confounding necessary to explain away observed associations. CONCLUSIONS: These findings fill a gap on the prevalence of LBW and NICU transfer/admission in term neonates of mothers who smoke and on the role of LBW linking to NICU placement, which could be used to update practitioners, to implement smoking cessation interventions, monitor trends, and to inform planning and allocation of healthcare resources.

A whole genome gene content phylogenetic analysis of anopheline mosquitoes.

Construction of stringent gene content matrices was accomplished for 21 Anopheline mosquito species and strains and four outgroups species. The presence absence matrix using e-75 as a cutoff in single linkage clustering had over 17,000 ortholog groups. We used the gene content matrix to generate a phylogenetic hypothesis that is in general agreement with gene sequence based phylogenies. In addition to establishing a congruent gene content phylogeny we examined the consistency of three methods for analyzing presence absence data - unweighted parsimony, dollo parsimonly and maximum likelihood using a BINGAMMA model. An examination of the chromosomal location of the gains and losses in the presence absence matrix revealed a low frequency of gains and losses at centromeres and tips of chromosomes.

Integrated Proteomic Pipeline Using Multiple Search Engines for a Proteogenomic Study with a Controlled Protein False Discovery Rate.

In the Chromosome-Centric Human Proteome Project (C-HPP), false-positive identification by peptide spectrum matches (PSMs) after database searches is a major issue for proteogenomic studies using liquid-chromatography and mass-spectrometry-based large proteomic profiling. Here we developed a simple strategy for protein identification, with a controlled false discovery rate (FDR) at the protein level, using an integrated proteomic pipeline (IPP) that consists of four engrailed steps as follows. First, using three different search engines, SEQUEST, MASCOT, and MS-GF+, individual proteomic searches were performed against the neXtProt database. Second, the search results from the PSMs were combined using statistical evaluation tools including DTASelect and Percolator. Third, the peptide search scores were converted into E-scores normalized using an in-house program. Last, ProteinInferencer was used to filter the proteins containing two or more peptides with a controlled FDR of 1.0% at the protein level. Finally, we compared the performance of the IPP to a conventional proteomic pipeline (CPP) for protein identification using a controlled FDR of <1% at the protein level. Using the IPP, a total of 5756 proteins (vs 4453 using the CPP) including 477 alternative splicing variants (vs 182 using the CPP) were identified from human hippocampal tissue. In addition, a total of 10 missing proteins (vs 7 using the CPP) were identified with two or more unique peptides, and their tryptic peptides were validated using MS/MS spectral pattern from a repository database or their corresponding synthetic peptides. This study shows that the IPP effectively improved the identification of proteins, including alternative splicing variants and missing proteins, in human hippocampal tissues for the C-HPP. All RAW files used in this study were deposited in ProteomeXchange (PXD000395).

A semi-mechanistic aging model of Pb added to soil by a modified stable isotope dilution technique.

The aging of Pb added to soils has not been studied by the isotopic technology because of difficulties in determination of isotopically exchangeable Pb in soil, so that a set of 10 typical agricultural soils in China and a one-year aging experiment with the addition of water-soluble Pb to the soils were carried out. A modified stable isotope dilution technique to determine isotopically exchangeable Pb in soil was developed where 0.2 mM EDTA (ethylenediaminetetraacetic acid) as the extractant. When water-soluble Pb was added to soil, the isotopically exchangeable Pb (Eadd%, the percentage of isotopically exchangeable Pb to total Pb added to soil) initially decreased rapidly and gradually slowly. A semi-mechanistic aging model of Pb added to soils, including precipitation/nucleation (Y1), micropore diffusion (Y2), and organic matter encapsulation processes (Y3) was developed with the root mean square error 8.3% where Y1, Y2, and Y3 accounted for 0.02~26.9%, 1.4~21.8% and 3.8~11.3%, respectively, when the pH 4.0~8.0 and organic matter 2.0~6.0%. Soil pH was a vital factor affecting the aging rate. When the pH increased by 1 unit, the Eadd value decreased by approximately 9%. The model could be used to scale ecotoxicological data of Pb in soil generated in different aging times.

The use of the E-value for sensitivity analysis.

Observational research designs enable clinicians to investigate topics for which randomized-controlled trials may be difficult to conduct. However, the lack of randomization in observational studies increases the likelihood of confounders introducing bias to study results. Analytical methods such as propensity score matching and regression analysis are employed to reduce the effects of such confounding, mainly by determining characteristics of patient groups and adjusting for measured confounders. Sensitivity analyses are subsequently applied to elucidate the extent to which study results could still be affected by unmeasured confounding. The E-value is one such approach. By presenting a value that quantifies the strength of unmeasured confounding necessary to negate the observed results, the E-value is a useful heuristic concept for assessing the robustness of observational studies. This article provides an introductory overview of how the E-value can be evaluated and presented in clinical research studies.

Risk of suicide after a diagnosis of sleep apnea: A nationwide longitudinal study.

Previous studies have presented evidence on the association between sleep apnea and suicidal ideation and planning, but the relationship between a clinical diagnosis of sleep apnea and suicide attempts remains unknown. We investigated the risk of suicide after a diagnosis with sleep apnea using data from a nationwide community-based population database, i.e., the Taiwan National Health Insurance Research Database. We recruited 7,095 adults with sleep apnea and 28,380 age-, sex-, and comorbidity-matched controls between 1998 and 2010 and followed them up until the end of 2011. Individuals who exhibited any (once or repeated) suicide attempts were identified during the follow-up period. The E value was calculated for unmeasured bias. Sensitivity analysis was conducted. Patients with sleep apnea were more likely to carry out any suicide attempt (hazard ratio: 4.53; 95% confidence interval: 3.48-5.88) during the follow-up period than the controls after adjusting for demographic data, mental disorders, and physical comorbidities. The hazard ratio remained significant after excluding individuals with mental disorders (4.23; 3.03-5.92). The hazard ratio was 4.82 (3.55-6.56) for male patients and 3.86 (2.33-6.38) for female patients. Consistent findings of increased risk of repeated suicide attempt were found among patients with sleep apnea. No association was found between continuous positive airway pressure treatment and suicide risk. The calculated E values support suicide risk after the diagnosis of sleep apnea. The risk of suicide was 4.53-fold higher in patients diagnosed with sleep apnea than in their counterparts without sleep apnea.

Association of maternal personality traits with medication use during pregnancy to appraise unmeasured confounding in long-term pharmacoepidemiological safety studies.

Maternal personality is a possible confounder on the association between prenatal medication exposure and long-term developmental outcomes in offspring, but it is often unmeasured. This study aimed to (i) estimate the association between five maternal personality traits and prenatal use of acetaminophen (including extended use), opioid analgesics, antidepressants, benzodiazepines/z-hypnotics, and antipsychotics; (ii) evaluate, using an applied example, whether unmeasured confounding by maternal neuroticism would make the association between prenatal antidepressant-child ADHD null, using the E-value framework. We used data from 8,879 pregnant women and recent mothers who participated in the Multinational Medication Use in Pregnancy Study, a web-based cross-sectional study performed within the period from 1-Oct-2011 to 29-Feb-2012 in Europe, North America and Australia. Medication use in pregnancy was self-reported by the women. Personality was assessed with the Big Five Inventory, capturing the dimensions of neuroticism, extraversion, openness, agreeableness, and conscientiousness. Adjusted logistic regression analyses were conducted for each trait-medication pair, using the survey weighting. There was a strong association between having high neuroticism and prenatal use of antidepressants (Odds Ratio (OR): 5.63, 95% Confidence Interval (CI): 3.96-8.01), benzodiazepines/z-hypnotics (OR: 6.66, 95% CI: 4.05-10.95), and analgesic opioids (OR: 2.24, 95% CI: 1.41-3.56), but not with antipsychotics. Among women with mental illness, this association attenuated for benzodiazepines/z-hypnotics, but decreased to the null for antidepressants. High neuroticism (OR: 1.31, 95% CI: 1.08-1.59) and high openness (OR: 0.77, 95% CI: 0.64-0.93) were associated with extended use of acetaminophen. The E-value for the Hazard Ratio 1.93 in the applied example was 3.27. If the example study was conducted using a population comparison group, high maternal neuroticism could have explained away the association antidepressant-ADHD. Because the example study included only women with a mental illness, this risk of bias was assessed as minimal. Various personality dispositions in the mother are associated, with a different degree, to prenatal use of medication. The strength of these association can aid researchers in evaluating the influence of uncontrolled confounding by maternal personality in long-term safety studies in pregnancy, using the E-value. This assessment should always be performed in addition to a rigorous study design using approaches to triangulate the evidence.

An E-value analysis of potential unmeasured or residual confounding in systematic reviews of post-tuberculosis mortality, respiratory disease, and cardiovascular disease.

INTRODUCTION: Unmeasured confounding poses a serious threat to observational studies of post-TB health outcomes. E-values have been recently proposed as a method to assess the magnitude of unmeasured confounding necessary to nullify, or to render non-significant, relative effect estimates from observational studies. METHODS: We calculated E-values for both the risk ratio (RR) point estimates and their lower 95% confidence limits (LCL) from studies of post-TB mortality, respiratory disease, and cardiovascular disease (CVD) included in published systematic reviews within and across post-TB outcome domains. We also employed a meta-analytic E-value approach to estimate the proportion of unconfounded study RRs greater than 1.1 at different levels of unmeasured confounding. RESULTS: Across post-TB health outcome domains, we observed a median E-value of 5.59 (IQR = 3.19-7.35) for RRs, and 2.95 (IQR = 1.71-4.61) for LCLs. Post-TB mortality studies had higher median E-values (E-valueRR = 6.90 and E-valueLCL = 4.54) than studies of respiratory disease (E-valueRR = 5.59, E-valueLCL = 2.94) or CVD (E-valueRR = 3.90, E-valueLCL = 1.81). The E-value at which the estimated proportion of studies with unconfounded RRs greater than 1.1 would remain over 0.7 was 3.45 for post-TB mortality, 3.96 for post-TB respiratory disease, and 1.71 for post-TB CVD meta-analyses. CONCLUSIONS: Unmeasured confounding with an association of 2.95 or greater with both the exposure (TB) and outcome, on the risk ratio scale, could render most post-TB health studies' findings statistically non-significant. Post-TB mortality and respiratory disease studies had higher E-values than TB-CVD studies, indicating that either (a) TB-CVD studies may be more susceptible to unmeasured confounding bias, or (b) the true effect of TB on CVD is lower.

Assessing Techniques for Quantifying the Impact of Bias Due to an Unmeasured Confounder: An Applied Example.

PURPOSE: To compare the magnitude of bias due to unmeasured confounding estimated from various techniques in an applied example. PATIENTS AND METHODS: We examined the association between dibutyl phthalate (DBP) and incident estrogen receptor (ER)-positive breast cancer in a Danish nationwide cohort (N=1,122,042). Cox regression analyses were adjusted for age and active drug compounds contributing to DBP exposure. We estimated the hazard ratios (HRs) that would have been observed had one of the DBP sources been unmeasured and calculated the strength of confounding by comparing to the fully adjusted HR. We performed a quantitative bias analysis (QBA) of the "unmeasured" confounder, using external information to specify the bias parameters. Upper bounds on the bias were estimated and E-values were calculated. RESULTS: The adjusted HR for incident ER-positive breast cancer among women with high-level (≥10,000 cumulative milligrams) versus no DBP exposure was 2.12 (95% confidence interval 1.12 to 4.05). Removing each DBP source in isolation resulted in negligible change in the HR. The bias estimates from the QBA ranged from 1.00 to 1.01. The estimated maximum impact of unmeasured confounding ranged from 1.01 to 1.51. E-values ranged from 3.46 to 3.68. CONCLUSION: The impact of bias due to simulated unmeasured confounding was negligible, in part, because the unmeasured variable was not independent of controlled variables. When a suspected confounder cannot be measured in the study data, our exercise suggests that QBA is the most informative method for assessing the impact. E-values may best be reserved for situations where uncontrolled confounding emanates from an unknown confounder.