Review of Hospital Outpatient No-Show Explainable Prediction Using Machine Learning

Patient no-shows present significant challenges and costs to healthcare systems, prompting a focused exploration of hospital outpatient no-show prediction using explainable machine learning (ML) in this review. The systematic methodology aligns with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, ensuring transparency, replicability, and thoroughness. A comprehensive search across major literature databases yielded 92 potential studies on ML models for outpatient no-show prediction. Following screening and eligibility assessments, 51 studies met inclusion criteria, emphasizing outpatient no-shows, leveraging explainable ML, featuring predictive modeling, and reporting quantitative metrics. Examining the publication distribution from 2010 to 2024 reveals a discernible upward linear trend, indicating the growing relevance of this research domain. The preference for more interpretable ML models over time, such as regression, decision trees, and ensemble methods, underscores their transparency and actionability. Despite challenges like limited model generalizability and clinician hesitation, proposed solutions, including collaborative learning and feature engineering pipelines, aim to enhance the reliability and applicability of outpatient noshow prediction models. This comprehensive strategy addresses existing challenges, facilitating the effective integration of explainable ML in healthcare settings and contributing to improved outpatient appointment attendance. In conclusion, this systematic review provides a robust foundation for understanding and implementing explainable ML approaches in predicting hospital outpatient no-shows, offering valuable insights for healthcare practitioners and researchers alike.