odds ratio interpretation

In common outcome scenarios, logistic regression’s odds ratio (OR) substantially deviates from the risk ratio (RR), leading to biased effect interpretation. Method: This paper proposes using complementary log-log (cloglog) regression to directly estimate RR and establishes, for the first time, a rigorous theoretical result: under any outcome prevalence, the “complementary log-ratio” induced by cloglog approximates RR with uniformly smaller absolute bias than OR. We further introduce the Aranda-Ordaz family of link functions to construct a unified theoretical framework enabling comparable effect estimation across models. Contribution/Results: Through analytical error-bound derivation and Monte Carlo simulations, we demonstrate that the proposed approach significantly improves RR estimation accuracy across low-to-high prevalence settings. The method is readily implementable in standard statistical software (e.g., R or SAS), combining theoretical rigor with practical feasibility.

This paper addresses the problem of identifying **interpretable, high-effect heterogeneous subgroups** from conditional average treatment effect (CATE) estimation. We propose a rule-set-based subgroup discovery framework—e.g., “treatment effect is significant if (X_1 > 0) and (X_2 < 1)”—that jointly optimizes subgroup size and effect magnitude, yielding a Pareto-optimal rule frontier; sample splitting ensures valid statistical inference. Our contributions are threefold: (i) explicit encoding of high-dimensional interaction effects into human-readable logical rules; (ii) the first application of multi-objective optimization to CATE subgroup identification, balancing interpretability, statistical significance, and representativeness; and (iii) theoretical guarantees of asymptotic unbiasedness and coverage for the derived rule sets. In extensive simulations and real-world policy evaluation datasets, our method substantially outperforms existing black-box subgroup detection approaches, achieving both strong statistical power and decision transparency.

In spatial logistic regression, incorporating random effects to account for spatial dependence shifts coefficient interpretation from population-averaged to subject-specific, thereby forfeiting marginal interpretability. To address this, we propose a bridge-process-based spatial logistic regression model that embeds spatially structured random effects without compromising marginal interpretability. This bridge process is the first spatial random-effects formulation that simultaneously preserves both marginal and conditional interpretations, and admits a scale-mixture-of-normals representation with favorable theoretical properties. Using Bayesian inference and an efficient MCMC algorithm, our model achieves superior predictive accuracy, computational efficiency, and interpretability in simulation studies and analysis of Gambian childhood malaria data. The framework establishes a new paradigm for modeling spatial binary data—rigorous from a statistical standpoint while retaining practical, policy-relevant interpretability.

In causal inference, sensitivity parameters are often difficult to calibrate due to their lack of intuitive causal interpretation, and existing methods ignore the sampling uncertainty in measured confounder estimation, leading to biased robustness assessments. This paper proposes a calibration-based sensitivity model: it directly constrains the strength of unmeasured confounding as a multiple of the estimated effect of measured confounders—endowing the sensitivity parameter with a clear causal interpretation (“unmeasured-to-measured confounding ratio”). It is the first to systematically incorporate the sampling variability of measured confounder estimates, thereby correcting inferential bias in bounding. Leveraging double robustness, nonparametric efficiency, and asymptotic normality theory, we construct three computationally tractable bounding models for the average treatment effect. Empirical analysis of maternal smoking’s effect on birth weight shows that conventional methods can substantially overstate or understate conclusion robustness. Our approach enhances the interpretability, calibration validity, and statistical reliability of sensitivity analysis.

Clinical prediction models are frequently developed from observational data that include early treatments, rendering them vulnerable to confounding, selection bias, mediation effects, and dynamic treatment regimes—collectively termed “causal blind spots”—which lead to miscalibrated risk estimates and suboptimal clinical decisions. This paper formally defines “causal blind spots” for the first time and demonstrates that conventional modeling strategies—treating treatment as a covariate, stratifying by treatment, or omitting treatment—are all unreliable. We propose an intervention-oriented framework centered on the *interventional prediction estimand*, integrating causal diagrams, do-calculus, and potential outcomes theory. This framework mandates embedding causal inference into both model development and validation. By shifting predictive modeling from associative pattern recognition to causal intervention modeling, our approach provides a principled foundation for revising clinical prediction guidelines to ensure causal validity, thereby enhancing the scientific rigor and safety of treatment decisions.

Health economic evaluations often struggle to accurately estimate the joint distribution of potential outcomes under interventions due to the absence of a single comprehensive data source, leading to structural and parametric biases in decision-analytic models. This study proposes a unified framework that integrates decision-analytic modeling within causal inference by leveraging the potential outcomes paradigm to synthesize causal parameters from multiple data sources and approximate intervention effects. It systematically decomposes estimation bias into structural bias and target parameter bias, elucidating how nonstandard parameter specifications induce bias and demonstrating that model credibility critically depends on underlying causal assumptions. By clarifying these mechanisms, the work enhances the transparency and rigor of health policy decision-making.

This work addresses the limitations of existing explainable AI methods, which predominantly focus on associative predictions and fall short in supporting decision-making that requires causal reasoning and counterfactual analysis. To bridge this gap, the paper proposes a novel framework that integrates causal machine learning with intrinsically interpretable models—such as additive models and symbolic regression—by explicitly embedding causal inference mechanisms within the model architecture. This approach enables the explicit recovery of causal structures and functional forms among variables directly from cross-sectional data. While maintaining high predictive accuracy, the method achieves comprehensive transparency in system structure, causal relationships, and response mechanisms, thereby substantially enhancing both interpretability and causal reliability for trustworthy “What-if” analyses.

Traditional heterogeneity analyses rely on pre-specified subgroups, limiting their ability to uncover complex effect modification mechanisms, while existing data-driven approaches often lack interpretability. This study proposes a novel paradigm grounded in causal transportability, treating population composition as a continuous variable to model the relationship between effect modifier distributions and the overall exposure effect. The framework enables estimation of intervention effects across diverse populations and identification of key vulnerability features. It integrates causal transportability, effect modifier selection—combining prior knowledge with data-adaptive strategies—and effect surface modeling, facilitating both effect attribution and ranking of modifier importance. Empirical application to child stunting and drought exposure successfully identifies critical modifiers, and an open-source Shiny interactive tool is provided to support broad adoption.

Traditional subgroup analyses in observational biomedical data often yield unstable and difficult-to-interpret results due to individuals experiencing only a single exposure, non-identifiable true causal effects, and uncertain confounding structures. This work proposes an integrated framework that first selects covariates via causal discovery, then constructs exposure- and outcome-agnostic pretreatment subgroups using unsupervised clustering methods—including K-means, fuzzy C-means, and Bayesian Gaussian mixture models. Subsequently, it evaluates hypothetical intervention strategies through uncertainty-aware screening combined with doubly robust estimation. The approach uniquely unifies unsupervised subgroup discovery with policy evaluation and introduces empirical Bernstein gating and Bayesian pooling to control risk. Applied to PIMA and NHANES datasets, the optimal policies achieved utilities of 0.735–0.799, though risk differences became nonsignificant after multiple testing correction.

This study addresses the challenge of cross-wave missing data in large-scale complex surveys, where certain variables are measured only in specific years. To enable effective prediction of unobserved outcomes for the current population, the authors propose a weighted conformal prediction framework that jointly estimates density ratios and subgroup proportions to approximate the likelihood ratio between historical samples and the target population. This approach corrects for temporal shifts in covariate distributions while preserving representativeness under complex sampling designs. Both theoretical analysis and empirical evaluations demonstrate that the method achieves valid prediction sets with coverage close to the nominal level and substantially improves prediction efficiency compared to existing approaches, as evidenced in simulation studies and real-world prediction of low-density lipoprotein cholesterol (LDL-C) levels in the U.S. population.