From Biomedical Datasets to Fairness-Aware Recommendations: An Integrated Data Orchestration Pipeline for Binary Clinical Predictions.

Many problems in biomedicine can be posed as binary classification. When they are addressed using artificial intelligence methods, though, average performance alone does not show whether a dataset is artificial intelligence ready, whether the endpoint is clinically valid, or whether errors are unevenly distributed across patient subgroups. This article presents the Fairness-Aware Data Orchestration Pipeline (FADOP), a reusable workflow that analyzes biomedical datasets, trains baseline binary classifiers, audits subgroup error patterns, tests mitigation strategies, and generates a documented recommendation. Such a pipeline is intended for systematic evaluation before clinical translation, not as an automatic deployment tool. Two publicly available case studies illustrate its use: the HIV-related ACTG175 dataset was repurposed from a treatment-comparison trial into a 1-year baseline mortality-prediction task, with death by day 365 as the positive class rather than the cid AIDS/failure composite endpoint; then, a stroke-risk dataset was analyzed as direct event prediction. The case studies show how the same workflow can generate cohort, performance, fairness, mitigation, and recommendation evidence across different rare-event clinical datasets.