Artificial Intelligence (AI) in Home Health: What You Need to Know in 2024

Introduction

If you speak with forward-thinking home health executives and CIOs about the wave of AI in 2024, one thing is clear: we are in the midst of an industry-redefining technology shift in home health.

Unlike the cloud and mobile shifts of previous decades, AI is moving at a faster pace and will produce new winners and losers in home health in ways beyond what we can imagine today.

Brief History of AI

The beginnings of AI can be traced back to the 1950s, when British mathematician Alan Turing conceptualized a machine capable of advancing past its original programming. Dartmouth professor John McCarthy would later coin the term “artificial intelligence.”

It wasn’t until the late 1990s, when IBM’s Deep Blue defeated then-world chess champion Gary Kasparov, that the field gained more funding. With renewed interest, students of Professor Geoffrey Hinton in 2012 released the idea of a system that processes information in a way similar to the human brain.

In 2017, Google researchers released a new architecture called the “transformer”, changing the way AI processes text, images, audio, and video to generate content. By 2020, the research company OpenAI had trained a generative transformer model, GPT-3, on billions of data points from the internet. The release of ChatGPT in 2022 brought generative AI to the forefront and triggered the AI arms race of 2023-2024.

For more information, see The History of AI: A Timeline of Artificial Intelligence.

Understanding Generative AI

Generative AI is a type of intelligence that learns from vast volumes of past information, including medical or legal textbooks, to generate realistic content. By 2023, these models were powerful enough to perform at an expert physician level and pass the bar exam.

A few types of tasks that generative AI is capable of performing:

• Summarization, such as summarizing a conversation between a clinician and a patient.
• Information Extraction, such as extracting key patient demographic fields from a document or audio recording.‍
• Action Initiation, such as determining when to initiate a billing action in an application.

Applications of AI in Home Health

AI Agents in Home Health

An AI agent is a software program that can interact with its environment, collect data, and use the data to perform self-determined tasks to meet predetermined goals. See What are AI Agents for more information on different types of AI agents.

For a given goal, an AI agent may perform several types of generative AI tasks. Consider an AI Scribe that wants to complete documentation for a home health clinician.

• The AI Scribe may recognize when a clinician has started providing care in the patient’s home and uses action initiation to remind the clinician of the planned goals and interventions for that visit.
• As the clinician speaks to the patient, the AI Scribe may use information extraction to fill out the structured data in the note.
• At the end, the AI Scribe may summarize all of the structured data into a narrative.

Other examples of AI agents in home health:

• AI Care Planner can analyze patient data from medical records and assessments to create personalized care suggestions.
• AI QA Reviewer can read through clinical documentation and flag deficiencies for regulatory compliance.
• AI Intake can monitor incoming referrals and accept patients based on agency capacity.‍
• AI Biller can monitor claims, validates for readiness, and submits to payers.

AI Safety in Home Health

To ensure the safe and ethical adoption of AI in home health, it is critical for early AI agents not to take action in a “black box” but rather build trust with its user. This means an AI agent should be expected to:

• Cite its source: You should be able to trace back to the source content used by your AI agent to perform its work.
• Shows its work: You should be able to view the steps your AI agent took to arrive at its work.
• Enable ease of review: You should be equipped with a user experience that highlights where human review is needed and reveals the confidence level the AI agent had when performing an action.

Evaluating AI in Home Health

If you are evaluating AI solutions, here are a few questions to ask your technology partner:

• How does the AI solution improve current healthcare processes or outcomes?
• How was the AI model trained, and what datasets were used?
• What is the accuracy rate of the AI solution? How does the solution handle false positives and false negatives?
• Is the AI solution compliant with relevant healthcare regulations (e.g. HIPAA)? What measures are in place to ensure data privacy and security?

Conclusion

AI cannot replace human judgment, emotional connection, or treatment delivery in the home. However, AI is rapidly taking on the administrative burden that has grown to levels no longer sustainable in the home health industry today.

Make no mistake: the AI wave has already arrived, and the time to prepare your organization for the education and adoption of AI is now.

Looking to begin your journey with AI in home health? Book a personalized meeting with our team today to understand how you can use Narrable Health AI agents to augment your team in their daily workflows.