An adaptive Mantel-Haenszel test for sensitivity analysis in observational studies

Addressing unmeasured confounding bias with a prior knowledge guided approach: coronary artery bypass grafting (CABG) versus percutaneous coronary intervention (PCI) in patients with stable ischemic heart disease

Unmeasured confounding undermines the validity of observational studies. Although randomized clinical trials (RCTs) are considered the "gold standard" of study types, we often observe divergent findings between RCTs and empirical settings. We present the "L-table", a simulation-based, prior knowledge (e.g., RCTs) guided approach that estimates the true effect adjusting for the potential influence of unmeasured confounders when using observational data. Using electronic health record data from Kaiser Permanente Southern California, we compare the effectiveness of coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) on endpoints at 1, 3, 5, and 10 years for patients with stable ischemic heart disease. We applied the L-table approach to the propensity score adjusted cohort to derive the omitted-confounder-adjusted estimated effects. After the L-table adjustment, CABG patients are 57.6% less likely to encounter major adverse cardiac and cerebrovascular event (MACCE) at 1 year (OR [95% CI] 0.424 [0.396, 0.517]), 56.4% less likely at 3 years (OR [95% CI] 0.436 [0.369, 0.527]), and 48.9% less likely at 5 years (OR [95% CI] 0.511 [0.451, 0.538]). CABG patients are also 49.5% less likely to die by the end of 10 years than PCI patients (OR [95% CI] 0.505 [0.446, 0.582]). We found the estimated true effects all shifted towards CABG as a more effective procedure that led to better health outcomes compared to PCI. Unlike existing sensitivity tools, the L-table approach explicitly lays out probable values and can therefore better support clinical decision-making. We recommend using L-table as a supplement to available techniques of sensitivity analysis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10742-022-00282-y.

A Semiparametric Approach to Model-Based Sensitivity Analysis in Observational Studies

When drawing causal inference from observational data, there is almost always concern about unmeasured confounding. One way to tackle this is to conduct a sensitivity analysis. One widely-used sensitivity analysis framework hypothesizes the existence of a scalar unmeasured confounder U and asks how the causal conclusion would change were U measured and included in the primary analysis. Work along this line often makes various parametric assumptions on U, for the sake of mathematical and computational convenience. In this article, we further this line of research by developing a valid sensitivity analysis that leaves the distribution of U unrestricted. Compared to many existing methods in the literature, our method allows for a larger and more flexible family of models, mitigates observable implications (Franks et al., 2019), and works seamlessly with any primary analysis that models the outcome regression parametrically. We construct both pointwise confidence intervals and confidence bands that are uniformly valid over a given sensitivity parameter space, thus formally accounting for unknown sensitivity parameters. We apply our proposed method on an influential yet controversial study of the causal relationship between war experiences and political activeness using observational data from Uganda.

Estimating Unhealthy Food Effects on Childhood Overweight in Malawi Using an Observational Study

INTRODUCTION

Consumption of unhealthy foods in children contributes to high levels of childhood obesity globally. In developing countries there is paucity of empirical studies on the association.  This study employed propensity-score methods to evaluate the effect of unhealthy foods on overweight among children in Malawi using observational data.

METHODS

Data on 4625 children aged 6 to 59 months from the 2015-16 Malawi Demographic and Health Survey (MDHS) were analyzed. A multivariable logistic regression model of unhealthy foods (yes or no) on purported confounders of childhood overweight was used to obtain a child's unhealthy food propensity score. The propensity scores were then used to form matched sets of healthy and unhealthy fed children. The association between unhealthy foods and childhood overweight was assessed using the conditional logistic regression model.

RESULTS

The prevalence of overweight (body mass index (BMI) z-score > 2 standard deviations) was estimated at 4.5% (3.8%, 5.3%). The proportion of children who consumed unhealthy foods was estimated at 14.6% (95% CI: 13.1%, 16.2%). Our propensity score matching achieved a balance in the distribution of the confounders between children in the healthy and unhealthy food groups. Children fed unhealthy foods were significantly more likely to be overweight than those fed healthy foods (OR = 2.5, 95% CI: (1.2, 5.2)).

CONCLUSION

The findings suggest the adverse effects of unhealthy foods on childhood overweight in Malawi. Thus, efforts to reduce unhealthy food consumption among children should be implemented and supported to address the problem of childhood overweight in Malawi and the sub-Saharan African region.

A statistic with demonstrated insensitivity to unmeasured bias for 2 × 2 × S tables in observational studies

Are weak associations between a treatment and a binary outcome always sensitive to small unmeasured biases in observational studies? This possibility is often discussed in epidemiology. The familiar Mantel-Haenszel test for a 2 × 2 × S $$ 2\times 2\times S $$ contingency table exaggerates sensitivity to unmeasured biases when the population odds ratios vary among the S $$ S $$ strata. A statistic built from several components, here from the S $$ S $$  strata, is said to have demonstrated insensitivity to bias if it uses only those components that provide indications of insensitivity to bias. Briefly, such a statistic is a d $$ d $$ -statistic. There are 2 S - 1 $$ {2}^S-1 $$ candidate statistics with S $$ S $$ strata, and a d $$ d $$ -statistic considers them all.  To have level α $$ \alpha $$ , a test based on a d $$ d $$ -statistic must pay a price for its double use of the data, but as the sample size increases, that price becomes small, while the gain may be large. The price is paid by conditioning on the limited information used to identify components that are insensitive to a bias of specified magnitude, basing the test result on the information that remains after conditioning. In large samples, the d $$ d $$ -statistic achieves the largest possible design sensitivity, so it does not exaggerate sensitivity to unmeasured bias. A simulation verifies that the large sample result has traction in samples of practical size. A study of sunlight as a cause of cataract is used to illustrate issues and methods. Several extensions of the method are discussed. An R package dstat2x2xk implements the method.

Evaluating the Effect of Appropriate Complementary Feeding Practices on Child Growth in Malawi Using Cross-Sectional Data: An Application of Propensity Score Matching

Introduction: Appropriate complementary foods have been found to provide infants and young children with nutritional needs for their growth and development. In the absence of a randomized control trial (RCT), this study used observational data to evaluate the effect of appropriate complementary feeding practices on the nutritional status of children aged 6–23 months in Malawi using a propensity score matching statistical technique.

Methods: Data on 4,722 children aged 6 to 23 months from the 2015–16 Malawi Demographic and Health Survey (MDHS) were analyzed. Appropriate complementary feeding practices were assessed using the core indicators recommended by the World Health Organization (WHO)/United Nations Children's Fund (UNICEF), and consist of the introduction of complementary feeding, minimum dietary diversity, minimum meal frequency and minimum acceptable diet based on a dietary intake during a most recent 24-h period.

Results: The prevalence of stunting (height-for-age z-score < −2 SD) was 31.9% (95% CI: 29.3%, 34.6%), wasting (weight-for-height z-score < −2 SD) 3.5% (95% CI: 2.6%, 4.7%) and underweight (weight-for-age z-score < −2 SD) 9.9% (95% CI: 8.4%, 11.8%). Of the 4,722 children, 7.7% (95% CI: 6.9%, 8.5%) were provided appropriate complementary foods. Appropriate complementary feeding practices were found to result in significant decrease in stunting (OR = 0.7, 95% CI: 0.4, 0.95). They also resulted in the decrease of wasting (OR = 0.4, 95% CI: 0.1, 1.7) and underweight (OR = 0.6, 95% CI: 0.2, 1.7).

Conclusion: Appropriate complementary feeding practices resulted in a reduction of stunting, wasting, and underweight among children 6 to 23 months of age in Malawi. We recommend the continued provision of appropriate complementary foods to infants and young children to ensure that the diet has adequate nutritional needs for their healthy growth.

Compositional and genetic alterations in Graves' disease gut microbiome reveal specific diagnostic biomarkers

Graves' Disease is the most common organ-specific autoimmune disease and has been linked in small pilot studies to taxonomic markers within the gut microbiome. Important limitations of this work include small sample sizes and low-resolution taxonomic markers. Accordingly, we studied 162 gut microbiomes of mild and severe Graves' disease (GD) patients and healthy controls. Taxonomic and functional analyses based on metagenome-assembled genomes (MAGs) and MAG-annotated genes, together with predicted metabolic functions and metabolite profiles, revealed a well-defined network of MAGs, genes and clinical indexes separating healthy from GD subjects. A supervised classification model identified a combination of biomarkers including microbial species, MAGs, genes and SNPs, with predictive power superior to models from any single biomarker type (AUC = 0.98). Global, cross-disease multi-cohort analysis of gut microbiomes revealed high specificity of these GD biomarkers, notably discriminating against Parkinson's Disease, and suggesting that non-invasive stool-based diagnostics will be useful for these diseases.

The uniform general signed rank test and its design sensitivity

A sensitivity analysis in an observational study tests whether the qualitative conclusions of an analysis would change if we were to allow for the possibility of limited bias due to confounding. The design sensitivity of a hypothesis test quantifies the asymptotic performance of the test in a sensitivity analysis against a particular alternative. We propose a new, nonasymptotic, distribution-free test, the uniform general signed rank test, for observational studies with paired data, and examine its performance under Rosenbaum’s sensitivity analysis model. Our test can be viewed as adaptively choosing from among a large underlying family of signed rank tests, and we show that the uniform test achieves design sensitivity equal to the maximum design sensitivity over the underlying family of signed rank tests. Our test thus achieves superior design sensitivity, indicating it will perform well in sensitivity analyses on large samples. We support this conclusion with simulations and a data example, showing that the advantages of our test extend to moderate sample sizes as well.

Ticagrelor-based antiplatelet regimens in patients treated with coronary artery bypass grafting: a meta-analysis of randomized controlled trials

OBJECTIVES

The optimal antiplatelet strategy in patients undergoing CABG remains unclear. This is the first meta-analysis investigating the clinical outcomes associated with ticagrelor-based antiplatelet regimens in patients receiving CABG.

METHODS

Relevant scientific databases were searched for studies investigating antiplatelet regimens after CABG from inception until April 1, 2019. Studies which randomly assigned CABG patients to either ticagrelor-based or control antiplatelet regimens were eligible. The primary outcome of this analysis was all-cause death. The main secondary outcome was MI. Other outcomes of interest were cardiac death, major adverse cardiac events, stroke and bleeding. This study is registered with PROSPERO, number CRD42019122192.

RESULTS

Five trials comprising 3996 patients (2002 assigned to ticagrelor-based and 1994 to control antiplatelet regimens) were eligible for quantitative synthesis. The median follow-up was 12 months. Control antiplatelet regimens consisted of either aspirin or clopidogrel or both. As compared to control, ticagrelor-based regimens reduced the risk of all-cause death [0.61 (0.43-0.87); P = 0.007], cardiac death [0.58 (0.39-0.86); P = 0.007] and major adverse cardiac events [0.79 (0.63-0.98); P = 0.03], without difference in the risk of MI [0.76 (0.50-1.18); P = 0.22], stroke [0.99 (0.56-1.78); P = 0.98] or bleeding [1.04 (0.95-1.14); P = 0.41]. There was a treatment effect modification for the primary outcome associated with trials enrolling predominantly patients with acute coronary syndrome (P for interaction = 0.038).

CONCLUSIONS

In patients receiving CABG, ticagrelor-based regimens reduce mortality and major adverse cardiac events without excess bleeding risk as compared with aspirin monotherapy or the combination of aspirin and clopidogrel. The benefit of ticagrelor-based regimens is more relevant in those studies enrolling predominantly patients with acute coronary syndrome. These findings require further confirmation in randomized trials focused on this subset of patients and powered for clinical outcomes.