Risk-Stratified Outreach Under Hypothetical Service-Capacity Constraints: A Study Protocol for Secondary Analysis of Real-World Diabetes Hospital Encounters

Study Protocol This study protocol presents a reproducible framework for a retrospective secondary analysis of real-world hospital encounters involving patients with diabetes. The proposed research uses the publicly available Diabetes 130-US Hospitals dataset from the UCI Machine Learning Repository, containing 101,766 historical hospital encounters recorded between 1999 and 2008. The study aims to develop and evaluate a risk-stratified prioritization approach for potential patient follow-up under hypothetical healthcare service-capacity constraints. The planned methodology includes data preprocessing, prediction of observed 30-day hospital readmission, comparison of prioritization strategies, and exploratory assessment of equity-related performance. Research status: This document is a study protocol. Empirical analysis of the original dataset has not yet been completed or independently validated. No observed effects of war, crisis-related service disruption, or clinical interventions are claimed. Capacity-constrained scenarios are hypothetical. Data source: UCI Machine Learning Repository. DOI: 10.24432/C5230J. Transparency statement: OpenAI ChatGPT assisted with initial document structuring and analytical code drafting. The author is responsible for reviewing, validating, and verifying the research before publication. Peer-review status: Not peer reviewed. Author: Elnaz Abedini, Doctoral Student in Biomedical Engineering, Karabük University, Türkiye.