ehrmonize 0.1.1

ehrmonize is package to abstract medical concepts using large language models.

EHRmonize

Welcome to EHRmonize, a Python package to abstract medical concepts using large language models.

Citation

TBD, stay tuned!

Motivation

Processing and harmonizing the vast amounts of data captured in complex electronic health records (EHR) is a challenging and costly task that requires clinical expertise. Large language models (LLMs) have shown promise in various healthcare-related tasks. We herein introduce EHRmonize, a framework designed to abstract EHR medical concepts using LLMs.

Rationale

EHRmonize is designed with two main components: a corpus generation and an LLM inference pipeline. The first step entails querying the EHR databases to extract and the text/concepts across various data domains that need categorization. The second step employs LLM few-shot prompting across different tasks. The objective is to leverage the vast medical text exposure of LLMs to convert raw input medication data into useful, predefined classes.

Dataset

Our curated and labeled dataset is accessible on HuggingFace.

Current supported tasks

Type	Task
Free-text	task_generic_drug
	task_generic_route
Multiclass	task_multiclass_drug
Binary	task_binary_drug
Custom	task_custom

Current supported models / engines / APIs

API	model_id
OpenAI	gpt-4
	gpt-4o
	gpt-3.5-turbo (discouraged!)
AWS Bedrock	anthropic.claude-3-5-sonnet-20240620-v1:0
	meta.llama3-70b-instruct-v1:0
	mistral.mixtral-8x7b-instruct-v0:1