Modern technological advancements occur at a rapid pace, and innovative organizations are regularly establishing new opportunities to enhance patient care using AI.
How Does AI Work?
Generally, AI systems work by absorbing large amounts of labeled data, analyzing it to identify patterns and correlations, and using its findings to make predictions about future outcomes. An example of this is a chatbot, which is fed examples of human text chats and processes them to learn to produce lifelike exchanges with people1.
Three cognitive skills are key to AI programming1:
- Learning Processes — The ability to acquire data and establish rules, called algorithms, that convert the data into actionable information. The algorithms then govern how a device completes a specific task.
- Reasoning Processes — The ability to choose the right algorithm to achieve the desired outcome.
- Self-Correction Processes — The ability to continually fine-tune and develop algorithms to ensure they achieve the most accurate results possible.
How Is AI Being Used in Healthcare?
AI is extremely versatile, so its applications are diverse and varied. Some help document patient information, others aid in diagnosis.
Generative AI, Such as ChatGPT, Assisting in Admin Tasks
Since its release in November 2022, ChatGPT has gained worldwide attention and sparked much discussion over how it can be utilized in the healthcare industry. The program is an example of “generative AI”, a term which describes algorithms and systems that can be used to create new content, such as text, images, or audio.
Despite the hype that surrounds generative AI and its potential, its current uses in healthcare are fairly limited. An example is its application at the University of Kansas Health System, where the technology is being implemented to aid clinician note taking by summarizing clinical conversations from recorded audio during patient visits2.
Another similar use is the latest program from Microsoft’s Nuance Communications, which will be used in electronic health record systems (EHR)3. Both applications will require users to describe what they’re seeing to function. For example, a clinician will need to provide a specific verbal commentary on what they see during an examination for the program to correctly enter the information into the EHR. The systems will, however, remove conversations not applicable to the care plan2.
Beyond these existing and similar applications, there is trepidation as to generative AI’s suitability for clinical use, at least in its current form.
Chatbots Engaging Patients and Easing Staffing Shortages
To stay in communication with patients, Northwell Health conducted health chats using a chatbot. Among the system’s oncology patients, these chats generated a 94% engagement rate. In addition, 83% of clinicians say the health chats extend the care they can deliver4.
AI Models Analyzing Public Health
By utilizing the ability of AI to comprehend and draw conclusions from large datasets, healthcare organizations can lay the foundation for precision public health. Using data points such as people, places, and time can provide more and better insights on the determinants of disease, both for individuals and population levels. The CDC also suggests this application of AI “can help accelerate public health surveillance and shape public health policies and implementation activities5.”
Remote Sensing Identifying When Medical Equipment Needs Maintenance
Using the capabilities of remote sensing, healthcare providers can learn when their medical equipment is in need of maintenance. Phillips highlights one case of remote sensing allowing them to monitor and analyze more than 500 parameters on an MR machine6. The data empowered them to proactively identify when hardware parts may need maintenance or to be replaced. They say that “as a result, 30% of service cases can be resolved before downtime is caused — preventing avoidable interruptions to clinical practice and unnecessary patient delays.”
AI and Robotics Helping Diagnose Lung Cancer
Atrium Health Wake Forest Baptist in North Carolina is using AI and robotics tools to improve early detection of lung cancer. The tools analyze imaging nodule characteristics to predict the likelihood of lung cancer, categorizing patients into high-risk, intermediate-risk, and low-risk categories7. By utilizing the tool, the health system hopes to reduce unnecessary biopsies for patients classified as low risk and decrease the number of false-positive tests.
AI Scanning Images to Detect Breast Cancer Risk
Massachusetts General Hospital has harnessed the ability of an AI system developed at Massachusetts Institute of Technology (MIT) to detect patients at high risk of developing breast cancer. Discover Magazine recalls the story of one patient who showed no signs of concern visible to human radiologists but was identified by the system as being “high risk for getting breast cancer in the next five years” after it scanned the patient’s images. Four years later, the patient indeed developed breast cancer8.
What Is the Future of AI in Healthcare?
It’s not a matter of if AI is the future of healthcare, it’s a matter of when its integration is complete, with experts predicting it will transform nearly every area of medical practice over time9. And it’s not just large health systems that can benefit, either.
The future of AI in healthcare could include artificial intelligence taking on a variety of tasks, according to Deloitte10, including “answering the phone to medical record review, population health trending and analytics, therapeutic drug and device design, reading radiology images, making clinical diagnoses and treatment plans, and even talking with patients.” In these ways, the adoption of AI can increase the capacity of smaller medical practices, allowing them to achieve more with less and helping them overcome their limitations.
With the volume of investments being made in the industry, progress is likely to be rapid. In 2022, investments in healthcare AI totaled $4.4 billion11. Although that number is considerably down on 2021’s figure, it falls in line with the amount spent in 2020. There’s no reason to think that this investment won’t continue to increase. That thought is validated by research suggesting the market for AI in diagnostics alone will cross $5.123 billion by 203012, which represents a compound annual growth rate (CAGR) of 23.45% between 2023 and 2030.
With financial investment trending upward, and continued advancements in the technology meaning manufacturers and organizations will keep finding new and useful applications for AI, all signs point toward the adoption of AI in healthcare only increasing as time goes on.
Sources
1 https://www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence
3 https://digitalhealth.modernhealthcare.com/digital-health/nuance-microsoft-chatgpt-gpt4-EHR
7 https://healthitanalytics.com/news/nc-health-system-deploys-ai-to-predict-diagnose-lung-cancer
8 https://www.discovermagazine.com/health/taking-a-closer-look-at-ai-and-diagnoses
9 https://www.medicaleconomics.com/view/the-tech-savvy-physician-how-ai-will-transform-your-practice
12 https://finance.yahoo.com/news/artificial-intelligence-ai-diagnostics-market-145500876.html