Automated Health Economic Model using Artificial Intelligence for Evaluating the Large Language Models

Abstract

Unprecedented conversation about the possible applications of large language models (LLMs) in a variety of fields, including healthcare, has resulted from the recent attention on LLMs. Although LLMs have shown great promise in completing jobs that humans can complete, they have also shown serious disadvantages, such as producing false information, manipulating data, and encouraging plagiarism. Artificial Intelligence has now turned into the fourth pillar upon which health resource allocation decision-making rests. This work presents an Automated Health Economic Model (AHEM) applying machine learning methods for economic evaluation and performance assessment of LLM) for healthcare-oriented NLP tasks. With the proposed framework, coupling predictive modeling and cost-utility analysis-based evaluation, we bridge clinical outcomes and economic feasibility. Using a dataset of NLP-related healthcare benchmarks and randomized simulated treatment scenarios, we implemented Random Forest, Gradient Boosting, and Neural Network algorithms to predict clinical benefits and operational costs of LLM deployment. This study shows that an automated AI-based economic model leads to better predictive accuracy (RMSE: 0.075) and somewhat improved consistency in decision-making when compared to classical models built manually. This presents a replicable method for healthcare technology assessment in AI-driven medicine.

Ähnliche Arbeiten