AI Safari

Author: Ayomide Owoyemi

The biggest rave right now in the tech world is artificial intelligence, and within the space of artificial intelligence, Language Models (LMs) – especially the large ones – are the most interesting tools. Whether we have or will achieve AGI or not is a different question, but right now, LMs are revolutionizing tasks, creativity, and the delivery of services. There have been a lot of proposed use cases in healthcare, from automation of notes to automation of summaries to use cases in offering care to patients. However, these use cases have some nagging challenges due to some inherent challenges that LMs have.

 

While the focus is on how to use LMs to cause a significant paradigm shift in healthcare, I believe that some of the most helpful and easiest to implement use-cases are of otherwise mundane things but with a significant improvement in how work is done.

I will highlight two use cases, one for health workers and one for patients.

Health workers use-case: Health workers working at healthcare facilities have a lot of information to deal with as it relates to administrative aspects of their work, how to navigate care delivery, and how to effectively use some of the digital tools that are part of their day-to-day workflow. There are usually different documents, notes, links, and sources of information for them to deal with any time they need clarification or guidance. They can call or ask someone, but being able to quickly find the needed insight on the go will be a useful solution in a lot of cases. Healthcare organizations can curate all these disparate pieces of information into a single database and then train an LLM on it and create a front end that’s based on a chat-like user interface for their staff to interact with.

Patient use-case: A common thing at healthcare facilities after you receive care is to be given a printed document with some information, education, and insights on what you were managed for. While I believe some people read those printed documents, a lot of people find them too much to sit down and read or spend time on. People deal much better with tailored information that is in small bits, and after every consultation session, patients will have questions about their health and care that might come to mind later, some of the answers are usually in those education materials but there is too much to deal with. Healthcare organizations can curate all this education information into a single database and then train an LLM on it and create a front end that’s based on a chat-like user interface for patients to interact with. They can text the patient a link to the chat page after the consultation with some initial prompts suggested for the patient. This will help patients to learn better and find easier ways to digest the information in those documents.

These are just two examples of otherwise mundane tasks that can have a significant impact on how work is done for these two groups of users.

I believe that augmenting simple tasks like this is easier and has far-reaching benefits in the short and long term before we expend energy on other use cases with more significant workflow disruption and unclear benefits and adoption.

Subscribe
Notify of
0 Comments
Inline Feedbacks
View all comments
Left Menu Icon
AI Safari