Artificial Intelligence usage in Healthcare

Since the COVID-19 pandemic, the healthcare industry has continued to adopt different functionalities of Artificial Intelligence. As telemedicine and virtual care have been at the forefront of technology during the pandemic, Artificial Intelligence’s utilization in healthcare has exploded as well. The integration of AI in Healthcare was expected to have a growth rate of 43.9% from 2021 to 2027 (Dicuonzo et al., 2023). As this technological explosion continues, it is time to understand the different uses of AI and the current applications of these cases within the industry since the end of the pandemic. This will also highlight considerations to make when searching and utilizing AI products.

Artificial Intelligence is the ability of computers to learn human-like functions or tasks (Microsoft, 2024). The technology’s utilization in healthcare involves three applications and expert systems which are natural language processing, speech recognition, and machine learning.

Natural Language Processing – AI’s ability to understand, analyze, and derive human language as it’s either spoken and/or written. (Gillis et al., 2024)

Speech recognition – AI’s ability to convert speech into text (Lutkevich & Kiwak, 2021).

Machine Learning – AI’s ability to utilize data and algorithms in order for the AI to imitate and respond the way humans learn and respond. The more the application progresses in its use, the more it improves its accuracy (IBM, 2021).

Examples of Speech recognition in healthcare are the most recent introduction of Dax or Ambient Assist (Nuance, 2024)(NextGen, 2023). These programs generate SOAP notes for providers after recording the conversation between them and their patients. The information for the SOAP notes will be extracted from the conversation and will then create a SOAP Note Draft for the provider to review for accuracy. NexGen’s Ambient Assist was able to demonstrate a 90% accuracy rate in the SOAP note as a result of using Speech recognition (NextGen, 2023).

An example of Natural Language Processing in Healthcare is the integration of chatbots. Most websites have a chatbot that can communicate and assist in directing the patients through navigating the website or answering basic questions. Another example of a chatbot was a study by AllianceChicago which performed NLP reminders to pediatric patient's parents reminding them to schedule a well-child check or informing them that their child is due for immunizations (AllianceChicago, 2023).

Machine Learning (ML) is probably the most diverse application in healthcare. Examples of ML being used are for coding and utilization management accuracy, radiology and imaging interpretations, utilizing data analytics to help simplify and create efficiency in processes, etc. (Johns Hopkins Medicine, 2023).

Although this technology is evolving and intruding on human applications within the industry, there are reasons for caution such as reviewing the results for accuracy. The ambient assist and Dax technology do not have a 100% success rate and it is important to understand that these notes need to be reviewed by a provider for accuracy as these notes are relied upon for reimbursement and are shared with other organizations if requested by the patient.

With Machine Learning, systems such as ChatGPT can be fallible and the accuracy is not 100 percent (Eliot, 2023). The reason is that this is a predictive technology with a learning algorithm that guarantees that the system will never be accurate all the time. There are times when these applications may create a false answer to a question, which should generate a concern for radiologists or other providers who are utilizing AI to interpret imaging scans. There is foresight for efficiency, a decrease in administrative burden, and accuracy when using AI-based Clinical Decision Support Systems, though until there are higher accuracy rates, providers should always question the AI-generated results before finalizing the documents. Below are resources for the applications discussed earlier in this article.

DAX: https://www.nuance.com/healthcare/dragon-ai-clinical-solutions/dax-copilot/explore-dax-for-clinicians.html (Nuance, 2024).

Ambient Assist: https://webadmin.nextgen.com/-/media/DAM/Collateral/Brochures/MB_GiveYourProvidersAnAdvantage_Brochure.pdf (NextGen, 2023).

Alliance Chicago: https://alliancechicago.org/2023/01/17/alliancechicago-publishes-digital-intervention-study-using-chatbots-to-increase-well-child-visits-and-immunizations/ (AllianceChicago, 2023).

Johns Hopkins Exploration on AI in the Reading Room: https://www.hopkinsmedicine.org/news/articles/2023/11/johns-hopkins-radiology-explores-the-potential-of-ai-in-the-reading-room (Johns Hopkins Medicine, 2023).

The development of regulations for Artificial Intelligence utilization in Healthcare is continuing to be investigated and generated as we move forward with the increased frequency of these products throughout the industry. ONC, the national coordinator for Health IT, released a regulation under the HTI-1 final rule surrounding the use of Decision Support Interventions by creating certification criteria for Health IT developers to comply with (ONC, 2023).

We are not recommending any products, though below is a website to G2 that provides evaluations of technology among different industries (G2, 2024). G2’s healthcare software page includes different options to choose from when looking for ratings (G2, 2024). This could help if a practice is interested in adopting an AI application to increase efficiency and decrease physician burnout. According to the AMA, in 2022, 63% of physicians report signs of burnout such as emotional exhaustion and depersonalization at least once per week (AMA, 2023). Medical Economics states that providers spend about 20 hours a week on paperwork alone, sometimes taking more time to document than actually caring for patients (Monger, 2023). The need for technology to increase efficiency and decrease this burden for providers is critical, and it is time to consider options that can help.

G2: https://www.g2.com/categories/health-care

References:

1. Dicuonzo, G., Donofrio, F., Fusco, A., & Shini, M. (2023). Healthcare system: Moving forward with artificial intelligence. Technovation, 120, 102510. https://doi.org/10.1016/j.technovation.2022.102510 
2. Microsoft. (2024, January 1). What is Artificial Intelligence?: Microsoft Azure. What is Artificial Intelligence? | Microsoft Azure. https://azure.microsoft.com/en-us/resources/cloud-computing-dictionary/what-is-artificial-intelligence#self-driving-cars 
3. Gillis, A. S., Lutkevich, B., & Burns, E. (2024, February 15). What is natural language processing?: Definition from TechTarget. Enterprise AI. https://www.techtarget.com/searchenterpriseai/definition/natural-language-processing-NLP 
4. Lutkevich, B., & Kiwak, K. (2021, September 13). What is speech recognition?. Customer Experience. https://www.techtarget.com/searchcustomerexperience/definition/speech-recognition 
5. IBM. (2021, September 22). What is machine learning (ML)? https://www.ibm.com/topics/machine-learning 
6. Nuance. (2024, May 1). Automatically document care with Dragon® Ambient eXperience (DAXTM ) Copilot. https://www.nuance.com/asset/en_us/collateral/healthcare/data-sheet/ds-ambient-clinical-intelligence-en-us.pdf 
7. NextGen. (2023, November 16). NextGen Ambient Assist Give Your Providers an Experience Advantage. NextGen Ambient Assist. https://www.nextgen.com/-/media/dam/collateral/brochures/mb_giveyourprovidersanadvantage_brochure.pdf 
8. AllianceChicago. (2023, January 17). AllianceChicago publishes digital intervention study using Chatbots to increase well-child visits and immunizations. https://alliancechicago.org/2023/01/17/alliancechicago-publishes-digital-intervention-study-using-chatbots-to-increase-well-child-visits-and-immunizations/ 
9. Johns Hopkins Medicine. (2023, November 1). Johns Hopkins Radiology explores the potential of ai in the reading room. https://www.hopkinsmedicine.org/news/articles/2023/11/johns-hopkins-radiology-explores-the-potential-of-ai-in-the-reading-room 
10. Eliot, L. (2023, October 5). Ai ethics battling stubborn myth that AI is infallible, including that autonomous self-driving cars are going to be unfailing and error-free. Forbes. https://www.forbes.com/sites/lanceeliot/2022/05/02/ai-ethics-battling-stubborn-myth-that-ai-is-infallible-including-that-autonomous-self-driving-cars-are-going-to-be-unfailing-and-error-free/ 
11. ONC. (2023, December 11). HTI-1 Decision Support Interventions (DSI) fact sheet. HTI-1 fact Sheet. https://www.healthit.gov/sites/default/files/page/2023-12/HTI-1_DSI_fact%20sheet_508.pdf 
12. G2. (2024, March 29). Best health care software in 2024. https://www.g2.com/categories/health-care 
13. AMA. (2023, February 16). What is physician burnout?. American Medical Association. https://www.ama-assn.org/practice-management/physician-health/what-physician-burnout 
14. Monger, M. (2023, August 2). How doctors and practices can rein in administrative burden. MedicalEconomics. https://www.medicaleconomics.com/view/how-doctors-and-practices-can-rein-in-administrative-burden