It’s hard to escape the talk of artificial intelligence in the news. AI is being used in various ways, and the medical industry is not exempt from these advancements. We’ll talk about how artificial intelligence is being used in medicine, the purpose of AI in medical devices, recent treatment of AI by the FDA, and what manufacturers need to consider when building a device.
How is AI used in medicine?
Artificial intelligence, while still a new technology, is being used in the medical field in many ways. We are likely to see these uses continue to grow and change as AI develops.
- Diagnostics: Classification models powered by artificial intelligence can help diagnose patients. For example, clinical imaging data has been used in dermatology to create visual classifications that identify skin lesions, skin cancer, and psoriasis.
- Drug discovery & development: We will likely see an increase in personalized care, particularly in drug discovery and development. Verge Genomics is working on applying machine-learning algorithms to the human genome and is working to pair drugs with various neurological diseases.
- Physician & patient communication: One of the most consumer-facing uses for AI has been in lessening the administrative burden of various tasks. Aiva Health in Los Angeles has improved the patient experience and reduced the administrative time by using a voice-powered care assistant to connect patients with the proper physician for communication.
- Transcribing documents: Bot assistants can also lighten the administrative load, transcribing documents for medical professionals, organizing images and files automatically, and answering clinical questions.
- Remotely treating patients: Remote care saw an uptick in response to the pandemic but has been historically limited by services that need to be conducted in person. However, remote diagnosis and treatment could be a future development.
AI in Medical Devices
Artificial intelligence has touched many aspects of the medical process but has particularly compelling implications for medical devices.
According to Sterling Medical Devices, the three main functions AI will likely take on in medical devices include chronic disease management, medical imaging, and Internet of Things (IoT) devices.
One major struggle for patients living with chronic diseases is that their conditions are long-ranging, but appointments can be so short and far between. Often, patients are tasked with self-monitoring and reporting, which can be hard to maintain or unreliable. AI has the potential to monitor and deliver treatment when patients need it. This could look like doling out insulin to patients with diabetes when a sensor indicates they need to regulate their blood sugar levels.
In addition to the development of classification systems, medical imaging with artificial intelligence also has the potential to be clearer and more high-quality, which could cut down on exposure time to radiation.
Internet of Things (IoT) devices are often associated with smart homes – bulbs, locks, and fitness trackers are a few examples. IoT medical devices are interrelated wireless devices that are connected to one another in a way that allows medical professionals to monitor their patients, keep them in the loop, manage their data, and cut down on resources used to treat, diagnose, and work with a patient. This constant interaction afforded by IoT devices can help improve patient outcomes.
The goal of AI in medical devices is to automate repetitive tasks, process and analyze data, improve the ability to predict disease risk and other outcomes and evaluate and monitor more effectively based on larger pools of data that can be collected because of this technology.
The FDA and AI Medical Devices
On September 28, 2022, the FDA provided its final guidance on clinical decision support (CDS) software. While it has been used in different contexts, according to the FDA, CDS includes “computerized alerts and reminders for providers and patients; clinical guidelines; condition-specific order sets; focused patient data reports and summaries; documentation templates; diagnostic support; and contextually relevant reference information.” Software that incorporates artificial intelligence is often categorized as CDS software.
If this CDS software is deemed a device, it needs to meet all FDA regulations for medical devices. The organization has created a helpful graphic for manufacturers and developers to determine whether the software they are using can be treated like a medical device.
To be a device, the CDS software needs to fall into one of four categories:
- Acquisition, processing, and analysis of medical patterns, images, or signals
- Analysis, printing, or display of medical information, such as waveforms or continuous signals/patterns
- The software provides information such as the probability of a condition, risk scores, or outputs that are time-critical
- Does not provide a basis for recommendations
The goal of much of the CDS software currently available is to personalize care to patients in an evidence-based manner. Software that falls under the FDA’s oversight needs to ensure compliance to be marketed and sold.
AI Medical Device Companies
In addition to specific guidance around CDS software, the FDA has a list of approved artificial intelligence and machine learning-enabled medical devices on its site. In October 2022, the agency added 178 devices to the list and made it possible to download all names and details in an Excel file. As of April 2023, the list has 521 entries and includes devices across all specialties.
For example, Iterative Health incorporates AI into its medical devices. Driven by precision medicine, the company is bringing AI-driven tools to gastroenterology, helping physicians detect, localize, and segment better when conducting endoscopies.
What AI medical device manufacturers need to know about certified parts
Medical device manufacturers using AI to improve the patient and practitioner experience will continue to grow and develop the field in ways we only see in glimpses. These new devices coming on the market will need to show they meet several compliance requirements, including IEC 60601-1, the standard for medical electrical equipment.
RAM Technologies’ PC-based power supply units are designed specifically for medical devices and are IEC 60601-1 3.2 certified. Keep innovating and leave the compliance to us. Contact us to learn more about our PSUs!