Generative AI is artificial intelligence that can create things like text, images, and music. But how would someone use generative AI in healthcare?
This technology has a long list of use cases in healthcare across the board, from drug discovery to improving patient outcomes.
The World Health Organization (WHO) estimates a deficit of 10 million health workers by 2030. Generative AI can play a significant role in addressing this deficit.
As a World Economic Forum article aptly explains:
“Generative AI isn’t just a passing trend; it’s a rapidly evolving ecosystem of tools growing in popularity showing huge potential to revolutionize healthcare in ways we’ve never seen before.”
In this article, we explore generative AI’s use cases in healthcare and their implications. We also help you assess your readiness to use generative AI at your healthcare facility.
Generative AI in Healthcare: Potential Use Cases
The healthcare industry is already thinking of generative AI’s many use cases. In fact, many healthcare experts believe generative AI can transform healthcare.
According to an Accenture report, 98% of healthcare providers and 89% of healthcare payer executives believe that generative AI developments will bring a new era of healthcare intelligence.
Below are some exciting generative AI use cases for the healthcare industry.
Drug Discovery
You’re probably familiar with how time-consuming and expensive the drug discovery process is — a vaccine takes an average of 10 to 15 years to develop.
Generative AI can speed up this process in various ways, such as:
- Identifying new drug molecules: Creating new drug molecules that can potentially be used to develop drugs faster using a large dataset consisting of chemical structures and properties.
- Testing the efficacy and safety of drug candidates: Drug discovery involves testing plenty of viable drug candidates for efficacy and safety. Generative AI algorithms can help identify or screen candidates based on very specific criteria that may exist in multiple data sets.
- Creating virtual compounds: AI can create and test virtual drug compounds in a computer simulation, which is cheaper than a lab.
- Identify target-specific molecules: Generative AI can study a molecular structure’s properties from a database to identify new molecules that can better deal with a specific target.
Disease Diagnosis
Generative AI can ingest a large dataset of medical images to learn the identification of patterns that indicate a specific disease.
For example, you can train AI using a large volume of images that show glaucoma or eyes with abnormalities and differentiate them from healthy eyes. This enables AI to learn how glaucoma looks different from other eye infections.
Medical professionals can use AI to make more timely and accurate diagnoses. AI also helps doctors create treatment plans faster, which helps improve patient outcomes.
Effectively, AI can help you provide the right treatment at the right time. AI-powered, personalized patient care reduces treatment misallocation and optimizes the associated resource utilization.
Patient Care and Treatment Plans
Generative AI can help create personalized treatment plans based on the patient’s medical records, lifestyle patterns, and genetic information.
You can also offer patients more engaging treatment, improving the plan’s effectiveness. For example, generative AI can auto-trigger notifications via email or text to help patients stick to their treatment plans, improving patient outcomes.
Medical Imaging
Generative AI can help enhance medical images. A generative AI model can study a large dataset of medical images of a specific body part.
Once the algorithm is trained using this data, you can feed a medical image, and the algorithm will generate a high-resolution image for you.
Think of a CT scan of the lungs, for example. When doctors feed this image into an AI algorithm to make it more high-resolution, they’ll be able to spot otherwise unnoticeable differences in the lungs that could be indicative of a disease.
Reduce the Friction of Data Access
Unsupervised generative AI is a type of generative AI that learns from unstructured data that lacks labels and categorization.
When you have a massive dataset and no guidance for analyzing it, unsupervised generative AI can help.
Interpreting unstructured data can help with various healthcare-related tasks. For example, it helps summarize key facts from unstructured data in EHR (electronic health records), medical notes, and medical images.
This eliminates the need to spend time reading through various types of records you handle on a daily basis. You can quickly review the summary and start testing, diagnosis, or treatment.
Reduce Physical Burnout
According to Dr. Robert Bart, chief medical officer at the University of Pittsburgh Medical Center, lessening physicians’ documentation burden is a top priority for many hospital technology leaders.
Clinicians spend roughly five to six hours a day taking notes in EHRs. This can cause burnout. Generative AI can summarize patient and treatment information. It can also review the patient’s medical records and provide a summary, making the physician’s work easier.
For example, think about how tools like otter.ai have made transcribing and synthesizing notes from virtual meetings a dream. The same can be achieved during a patient consult.
Generative AI can significantly reduce administrative workload. This not only addresses the problem of shortage of healthcare workers but also helps minimize errors — up to 50% of all medical errors in primary care are caused because of administrative reasons.
Automated Health Assistance
Medical professionals often spend a lot of time responding to emails or scheduling appointments. Creating a virtual health assistant (i.e., an AI-powered chatbot) can help patients get basic information and schedule appointments themselves.
Medical Research
Sharing patient data for research and healthcare while ensuring data security and privacy. New drugs and medical devices require years of research and billions of dollars before they pass clinical trials and get regulatory approval.
Gen AI can possibly be used to partially simulate clinical testing. This can help bring drugs to market faster and at a lower cost.
Five-Step Approach to Assess if You Are Ready for Healthcare Generative AI
Assessing your readiness before implementing generative AI ensures your business and team aren’t overwhelmed by its complexities.
Jumping in with complete awareness gives you the confidence to derive generative AI’s full benefits. Here’s how you can assess your business’s readiness for generative AI:
- Start by learning about AI technology: Learn about the AI tech you want to use. What problem will it solve for your business? How good is the tech at solving your problem?
- Identify major areas that need AI intervention: Does the AI tech solve problems related to drug discovery, disease diagnosis, or burnout?
- Barriers to adoption: Will something stand in the way of you successfully implementing this tech? It could be your budget or lack of buy-in from employees who fear they might lose their jobs.
- Prioritize: Study the business dynamics around entering high-priority categories with this tech. Do you expect any disruption during or after implementing the AI tech?
- Find potential for expanding the scope of the solution: Can you extend the scope of the AI solution to include more than just the primary use case? If you’re creating a healthcare assistance solution, you might consider expanding the solution’s scope by including patient education.
This five-step approach will help you understand your readiness for implementing AI solutions for your healthcare business.
The Practical Implications of Using Generative AI in Healthcare
Microsoft and Epic are already testing generative AI for EHRs. Google is developing its own language model Med-PaLM 2, which aims to help diagnose complex diseases.
With multiple big companies already making a move towards using generative AI in healthcare, here are some implications we can expect to see in healthcare:
Clinicians Will Spend Less Time on Clinical Documentation
Generative AI will eliminate the need for clinicians to spend too much time on documentation.
Major clinical documentation companies are starting to integrate generative AI into their solutions. Doing so can improve the solution’s documentation speed and accuracy.
For example, Microsoft’s documentation company Nuance recently integrated GPT-4 into its clinical note-taking solution.
Similarly, Google has partnered with Suki, a documentation company that launched a generative AI voice assistant. Suki can auto-generate clinical notes during a conversation and enter those details into the note.
AI Will Emulate Medical Decision-Making
Generative AI can’t guarantee a future outcome. However, it can “predict” the next best idea when generating text to illustrate a concept.
This can help solve various problems, including making medical decisions. In fact, generative AI uses a similar process to make decisions as doctors.
Doctors learn medical science in college and then from research papers, medical journals, and experience. Similarly, AI learns from digital resources available on the internet.
When a patient walks in, doctors use their knowledge or search through relevant assets to understand and treat a patient’s symptoms. AI tries to find relevant text using billions of parameters.
Doctors predict possible causes and try to narrow things down by comparing diagnoses. Generative AI compares words and sentences, weighs its options, and finds the best possible match from the available resources.
The most significant shortcoming is that AI currently can’t interrogate the patients or order tests for more context. This will likely change in the near future.
Medical Errors Will Reduce Significantly
Medical errors are a major cause of death in U.S. hospitals. A New England Journal of Medicine paper suggests that close to 1 in 4 patients in a hospital may experience some sort of harm during their time in the hospital.
The future generations of AI will be able to monitor medical professionals and trigger alerts in real time when they’re about to commit an error. This includes errors committed when writing a prescription or during surgery.
There has been promising progress toward AI-based video surveillance. Installing these technologies in ICUs and operation theaters significantly improves patient outcomes and minimizes errors.
You can use an AI-based solution with a patient monitor to monitor the patient’s condition. The AI tool can collect this data and over time, learn from this data to become better at detecting abnormalities.
However, it’s also important to remember that AI can make mistakes. Moreover, you can’t quite determine how AI got the data it used to generate the results it did. Keeping these limitations will ensure you’re able to make the best use of generative AI in healthcare.
The Discovery of New Use Cases Will Continue
Generative AI is still in its infancy. As it evolves, healthcare professionals will be able to find new use cases to improve efficiency and minimize errors.
The problem? AI is evolving way too fast. You’ll need constant support from a qualified partner to determine how you can succeed.
Blanc Labs can help you implement customized, turnkey AI solutions that will provide the support you need on your digital transformation journey. Book a discovery call with Blanc Labs to learn more about how generative AI can help your healthcare business.
