Forecasting Urgent Dialysis Needs: From Predictive Accuracy to Clinical Actionability

Return visit admissions (RVA), which are instances where patients discharged from the emergency department (ED) rapidly return and require hospital admission, have been associated with quality issues and adverse outcomes. We developed and validated a machine learning model to predict 72-hour RVA using electronic health records (EHR) data. Study data were extracted from EHR data in 2019 from three urban EDs. The development and independent validation datasets included 62,154 patients from two EDs and 73,453 patients from one ED, respectively. Multiple machine learning algorithms were evaluated, including deep significance clustering (DICE), regularized logistic regression (LR), Gradient Boosting Decision Tree, and XGBoost. These machine learning models were also compared against an existing clinical risk score. To support clinical actionability, clinician investigators conducted manual chart reviews of the cases identified by the model. Chart reviews categorized predicted cases across index ED discharge diagnosis and RVA root cause classifications. The best-performing model achieved an AUC of 0.87 in the development site (test set) and 0.75 in the independent validation set. The model, which combined DICE and LR, boosted predictive performance while providing well-defined features. The model was relatively robust to sensitivity analyses regarding performance across age, race, and by varying predictor availability but less robust across diagnostic groups. Clinician examination demonstrated discrete model performance characteristics within clinical subtypes of RVA. This machine learning model demonstrated a strong predictive performance for 72- RVA. Despite the limited clinical actionability potentially due to model complexity, the rarity of the outcome, and variable relevance, the clinical examination offered guidance on further variable inclusion for enhanced predictive accuracy and actionability.

Chronic kidney disease (CKD) represents a major and growing global public health burden, with substantial proportions of cases remaining undiagnosed until advanced stages. Emerging precision public health approaches leveraging multi-omic risk scores, integrating genomic, epigenomic, proteomic, and metabolomic data offer significant promise for improving early detection, risk stratification, and targeted prevention strategies. This paper examines the current evidence base supporting the use of multi-omic risk scores in CKD, highlighting their potential to enhance predictive accuracy beyond conventional clinical models and enable timely, individualized interventions at the population level. Despite these advances, substantial challenges remain in translating multiomic risk scoring into routine public health practice. Key implementation barriers include limited external validation across diverse populations, inadequate health system infrastructure for data integration, and unresolved questions regarding clinical utility and actionability. Equity concerns are particularly salient, as underrepresentation of diverse populations in omics datasets risks exacerbating existing health disparities, while unequal access to testing and care may skew benefits toward more advantaged groups. Ethical and governance considerations including data privacy, consent, and fair data use further complicate large-scale deployment. Addressing these challenges requires a coordinated, interdisciplinary approach that integrates methodological rigor, inclusive data generation, robust governance frameworks, and health system readiness. Strengthening population diversity in datasets, improving interoperability of health data systems, and aligning policy and funding mechanisms will be essential to ensure equitable and effective implementation. Ultimately, multi-omic risk scores have the potential to transform CKD prevention and management within a precision public health framework, provided that scientific innovation is matched with ethical, equitable, and context-sensitive implementation strategies. Keywords: Chronic kidney disease (CKD), Multi-omic risk scores, Precision public health, Health equity and Implementation science.