A Multi-Head Attention Soft Random Forest for Interpretable Patient No-Show Prediction

TL;DR

This paper introduces a novel Multi-Head Attention Soft Random Forest model that improves patient no-show prediction accuracy and interpretability by integrating attention mechanisms with probabilistic soft splits.

Contribution

The paper presents a hybrid MHASRF model that combines attention mechanisms with random forests using soft splits, enhancing adaptability and interpretability in patient no-show prediction.

Findings

Achieved 93.72% accuracy and 97.87% AUC, outperforming traditional models.

Enabled deeper insights into key predictors through dual-level feature importance.

Demonstrated high performance and interpretability in healthcare prediction tasks.

Abstract

Unattended scheduled appointments, defined as patient no-shows, adversely affect both healthcare providers and patients' health, disrupting the continuity of care, operational efficiency, and the efficient allocation of medical resources. Accurate predictive modeling is needed to reduce the impact of no-shows. Although machine learning methods, such as logistic regression, random forest models, and decision trees, are widely used in predicting patient no-shows, they often rely on hard decision splits and static feature importance, limiting their adaptability to specific or complex patient behaviors. To address this limitation, we propose a new hybrid Multi-Head Attention Soft Random Forest (MHASRF) model that integrates attention mechanisms into a random forest model using probabilistic soft splitting instead of hard splitting. The MHASRF model assigns attention weights differently across the trees, enabling attention on specific patient behaviors. The model exhibited 93.72% accuracy, 94.77% specificity, 90.23% precision, 89.38% recall, a 91.54% F1 score and AUC 97.87%, demonstrated high and balance performance across metrics, outperforming decision tree, random forest, logistic regression, and naive bayes models overall. Furthermore, MHASRF was able to identify key predictors of patient no-shows using two levels of feature importance (tree level and attention mechanism level), offering deeper insights into patient no-show predictors. The proposed model is a robust, adaptable, and interpretable method for predicting patient no-shows that will help healthcare providers in optimizing resources.