<h4>Aims</h4>Appropriate analysis of big data is fundamental to precision medicine. While statistical analyses often uncover numerous associations, associations themselves do not convey predictive value. Confusion between association and prediction harms clinicians, scientists, and ultimately, the patients. We analyzed published papers in the field of diabetes that refer to "prediction" in their titles. We assessed whether these articles report metrics relevant to prediction.<h4>Methods</h4>A systematic search was undertaken using NCBI PubMed. Articles with the terms "diabetes" and "prediction" were selected. All abstracts of original research articles, within the field of diabetes epidemiology, were searched for metrics pertaining to predictive statistics. Simulated data was generated to visually convey the differences between association and prediction.<h4>Results</h4>The search-term yielded 2,182 results. After discarding non-relevant articles, 1,910 abstracts were evaluated. Of these, 39% (n = 745) reported metrics of predictive statistics, while 61% (n = 1,165) did not. The top reported metrics of prediction were ROC AUC, sensitivity and specificity. Using the simulated data, we demonstrated that biomarkers with large effect sizes and low P values can still offer poor discriminative utility.<h4>Conclusions</h4>We demonstrate a landscape of confused reporting within the field of diabetes epidemiology where the term "prediction" is often incorrectly used to refer to association statistics. We propose guidelines for future reporting, and two major routes forward in terms of main analytic procedures and research goals: the explanatory route, which contributes to precision medicine, and the prediction route which contributes to personalized medicine.

This primer introduces the domains into which the aims of quantitative health research generally fall and provides tools to improve the methodological quality of observational clinical and population-based research articles—with a special focus on the field of neurology. Generally, research questions can be categorized into one of the following 3 data science domains: description, prediction, and causal inference. A descriptive question aims to quantify and describe the frequency and distribution of a given health condition in a certain population at or during a specific time. A predictive question aims to estimate either the probability of the presence of a given disease or health condition in an individual (diagnostic prediction) or the probability of an individual developing a disease of interest over a specified period (prognostic prediction). A causal question aims to estimate the causal effect of interest (estimand) of an exposure or intervention on an outcome in a given population. Depending on the research question, estimands could be the total causal effect, a mediated indirect effect, or effect (measure) modification by third variables, among others. Each of these domains comes with its own set of research methods, study designs, reporting guidelines, scientific language, strengths, and limitations, whereby the correct attribution of a research domain will have an impact in 3 ways: i) help authors to formulate appropriate research questions and choose and implement suitable study designs and methods; ii) allow reviewers and editors to assess studies with an increased focus on their clinical relevance, methodological advances, and novelty and quality of clinical evidence; and iii) facilitate clear communication of findings and clinical implications to the broader research community in neurology and related fields.

Accurate traffic forecasting is critical for intelligent transportation systems. While recent spatiotemporal Graph Neural Networks (GNNs) have shown strong performance by modeling spatiotemporal dependencies, they primarily capture correlational patterns and lack a causal foundation. This limits their interpretability and robustness, especially under partial observability and structural changes in traffic dynamics. We propose CausalGRIT, a causal spatiotemporal forecasting framework that integrates observational learning with intervention-aware reasoning. Grounded in Structural Causal Models (SCMs), our method constructs dynamic causal graphs that encode directed cause-effect relationship across space and time. To handle latent confounding from incomplete sensor data, we introduce a variational belief encoder with planar flows for uncertainty-aware inference. To further enhance robustness, we develop an Edge Generation via Counterfactuals (EGC) module that simulates interventions to reveal and regularize weak or spurious dependencies during training. On three California PeMS (Performance Measurement System) datasets, CausalGRIT reduces RMSE by 7.2% relative to the next best baseline in average and shows far greater robustness: under perturbations, MAE increases by only 0.8%, compared with 8.8% for non-causal models.

Causal discovery and inference from observational data is an essential problem in statistics posing both modeling and computational challenges. These are typically addressed by imposing strict assumptions on the joint distribution such as linearity. We consider the problem of the Bayesian estimation of the effects of hypothetical interventions in the Gaussian Process Network (GPN) model, a flexible causal framework which allows describing the causal relationships nonparametrically. We detail how to perform causal inference on GPNs by simulating the effect of an intervention across the whole network and propagating the effect of the intervention on downstream variables. We further derive a simpler computational approximation by estimating the intervention distribution as a function of local variables only, modeling the conditional distributions via additive Gaussian processes. We extend both frameworks beyond the case of a known causal graph, incorporating uncertainty about the causal structure via Markov chain Monte Carlo methods. Simulation studies show that our approach is able to identify the effects of hypothetical interventions with non-Gaussian, non-linear observational data and accurately reflect the posterior uncertainty of the causal estimates. Finally we compare the results of our GPN-based causal inference approach to existing methods on a dataset of $A.~thaliana$ gene expressions.

In causal inference, confounding is a form of systematic error (or bias) that can distort estimates of causal effects in observational studies. A confounder In causal inference, confounding is a form of systematic error (or bias) that can distort estimates of causal effects in observational studies. A confounder is traditionally understood to be a variable that (1) independently predicts the outcome (or dependent variable), (2) is associated with the exposure (or independent variable), and (3) is not on the causal pathway between the exposure and the In causal inference, confounding is a form of systematic error (or bias) that can distort estimates of causal effects in observational studies. A confounder is traditionally understood to be a variable that (1) independently predicts the outcome (or dependent variable), (2) is associated with the exposure (or independent variable), and (3) is not on the causal pathway between the exposure and the outcome. Failure to control for a confounder results in a spurious association between exposure and outcome. Confounding is a causal concept rather than a purely statistical one, and therefore cannot be fully described by correlations or associations alone. The presence of confounders helps explain why correlation does not imply causation, and why careful study design and analytical methods (such as randomization, statistical adjustment, or causal diagrams) are required to distinguish causal effects from spurious associations. Several notation systems and formal frameworks, such as causal directed acyclic graphs (DAGs), have been developed to represent and detect confounding, making it possible to identify when a variable must be controlled for in order to obtain an unbiased estimate of a causal effect. Confounders are threats to internal validity.