A Deep AI–Orchestrated Ensemble Architecture for End-to-End Pharmacy Automation with Error-Aware and Workflow-Adaptive Intelligence

View at publisher Summarise with VIRAT
Authors: Mathan Stephen, Tissa Mary Mathew
Cited by: 0
References: 15

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Medication errors are a significant patient safety issue, with manual dispensing error rates between 0.08 % and 3.3 %. This study introduces an intelligent pharmacy automation framework that uses deep neural networks (DNNs) to minimize medication errors. The system combines Convolutional Neural Networks (CNNs) for medication verification (99.99% accuracy), Long Short-Term Memory (LSTM) networks for adverse drug event prediction (94.2 % sensitivity), Gradient Boosting Machines (GBM) for prescription error detection (96.8% precision), and Reinforcement Learning (RL) for workflow optimization. The framework integrates automated storage, intelligent dispensing, unit-dose robotics, and barcode administration using HL7 standards. Validation at three healthcare facilities showed 99.97% dispensing accuracy (a 91% reduction in errors), 85.7% fewer administration errors, 67.4 % shorter waiting times, and 94 % fewer near-miss incidents. Comparative analysis demonstrated a 3.2% improvement in error detection, a 5.7 % increase in prediction accuracy, and processing that is 2.3 times faster. Economic analysis indicates a positive return on investment within 2.68 years, with annual benefits exceeding $1.04 million per facility. These results highlight the effectiveness of multi-model DNN integration for medication safety.

Fields of Study & Subjects

Patient Safety and Medication Errors Artificial Intelligence in Healthcare and Education Artificial Intelligence in Healthcare

Publication details

Journal / source
International Conference on Sentiment Analysis and Deep Learning (ICSADL)
Pages
1–9
Publication date
18 February 2026
Language
EN