Next steps and resources
This guide helps you use AWS services to automate medical NLP and generative AI tasks for real-world applications in production environments. It describes how you can use Amazon Comprehend Medical, supported LLMs in Amazon Bedrock, pretrained medical LLMs, or fine-tuned LLMs to achieve your healthcare and life science business objectives. This guide describes the advantages and limitations for the following approaches:
Using Amazon Comprehend Medical independently
Providing Amazon Comprehend Medical results to an LLM
Using a pretrained general LLM or a medical LLM in a Retrieval Augmented Generation (RAG) approach
Fine-tuning a general LLM or medical LLM
Use the decision tree and the business maturity considerations in this guide to choose between these approaches based on your organization's AI/ML maturity level. Although Amazon Comprehend Medical and Amazon Bedrock LLMs provide powerful capabilities, they are only successful if you properly implement and evaluate them. Use the evaluation information and metrics described in this guide to validate the performance of your solution.
For next steps, we recommend that healthcare IT managers, architects, and technical leads work with AI/ML practitioners to identify their NLP medical task. Use this guide to choose a development path, and then use the appropriate AWS services and features to successfully implement an automated solution on AWS.
AWS resources
Other resources
