Healthcare facilities of all sizes are employing AI to reduce administrative workloads and improve memo acquisition. However, adopting technology involves risks.
Business leaders in almost every sector are plagued by the possibilities of AI to automate repetitive tasks, improve efficiency and save money. Healthcare is no exception.
Physicians and other providers are interested in ways that can help use AI to automate patient communication, improve response times, manage tasks such as scheduling and inventory, and improve outcomes of radiology and imaging care.
These tools are promising, but like new technologies, they could put healthcare companies at unexpected risks. For now, how insurance policies cover AI is a bit unclear if you make a claim mistake.
Hospitals and other hospitals in this sector employ other tools, so it is important to carefully understand how AI works, how it works, and what the potential impact is. It is important for everyone involved to remember that these tools are not an alternative to what long-time educational doctors, nurses and other healthcare professionals bring to the table.
“AI is not intended to completely replace a physician’s independent judgment, and there is no space for it to happen anytime soon. JNIFER FREEDEN of CPHRM, Southwest Regional Risk Manager at ProAssurance, says: AI is just another tool in a physician’s toolbox.”
“You cannot replicate human components,” added Proassurance, Bradley E. Byrne, Jr., JD, Southeast Regional Risk Manager.
Risks of AI implementation in healthcare
Jennifer Freeden, JD, CPHRM, Southwest Regional Risk Manager, ProAssurance
Currently, health facilities of various sizes and specializations, and most regions of the country use AI primarily to support “scheduling, healthcare supply inventory, staffing needs, surgery, laboratory availability, and more.” “These implementations of AI are at a low risk of delivering measurable benefits for future planning, while increasing patient care and satisfaction.”
These tools also have advantages in meeting patient demands through the health system’s online portal. For example, if a patient needs a record of a particular treatment they have received, the system can find it and upload it quickly.
“I never sleep. It’s never tired,” Byrne said. “When a patient sends a request for the Patient Portal to get a copy of the record, artificial intelligence can generate a custom response and provide it in real time to the requested record.”
Others see the possibility of AI for more dangerous tasks, such as taking notes during patient visits and evaluating medical images. AI shows some success here, but it is important for the doctor to check its output to make sure there are no errors. Relying on these tools can lead to claims.
“Humans tend to be overly dependent on technology, especially the longer we are exposed to it, the more comfortable we become,” Byrne said. “People, including doctors who use these techniques, can focus primarily on positives without fully understanding their limitations.”
One reason why AI provides inaccurate results is data. An artificial intelligence system is as good as the data it trains. If the dataset is biased or incomplete, it may result in inaccurate results for healthcare providers.
“If the underlying dataset is based on an adult population, pediatricians should not utilize that particular AI solution when serving pediatric patients under the age of 18, and it is the clinician’s responsibility to understand the AI model they are using,” Freeden said.
It is important that doctors ask: Byrne said.
AI and insurance
Bradley E. Byrne, Jr., JD, Southeast Regional Risk Manager, ProAssurance
Many companies rely on insurance contracts to address the risks AI poses in the healthcare sector. Healthcare providers are subject to medical malpractice and have product liability policies covering new technologies. Certainly something protects you from the risks that AI poses.
Currently, it is unclear whether AI risks are subject to general liability or malpractice policies. If the AI system is doing something wrong, is it a product fault? Or a doctor who depends on that decision? Currently, there are no standalone, AI-specific insurance products that can solve some of these questions. Also, AI-specific policy exclusions are not widely practiced.
“The emerging gray area is the intersection of product issues and medical malpractice issues, where liability is reduced in the event of patient harm,” Fridon said. “Through independent judgment using these tools, liability may be allocated between medical device tools and functions that may be specifically identified as defects juxtaposed with whether or not medical standards have been met with independent judgment. Courts face the difficult and nuanced task of distinguishing these liabilities.”
Meanwhile, some healthcare expects AI will help healthcare providers avoid mistakes that could lead to claims. “There is probably a bit of optimism that AI could ultimately help prevent adverse events.
What are the best risk management practices?
Like other new technologies, AI carries risks. Especially in these early times when everything is changing rapidly. However, many in the healthcare industry are optimistic about the potential of technology. This is part of why many hospitals and other health systems, as well as small rural facilities, quickly embraced it.
“It will probably be part of the solution to the problem of a shortage of doctors on the horizon,” Byrne said.
Healthcare companies considering implementing AI may feel flooded with rapid development and product numbers on the market. “It’s very easy to be overwhelmed by the vast amount of new and evolving information about AI in the healthcare field,” Friedon said. “It would be wise to have a manager in the office responsible for staying up to date with AI healthcare development.”
In addition to keeping up with the general trends, it is important for medical companies to know through specific AI systems. Knowing which data is trained and which use cases are appropriate will reduce exposure.
“To know intimately the underlying AI used in clinical decision-making is important for practices and hospitals to move forward. It’s not enough to say, ‘Because the AI told me that’s what it is,” Freeden said. “Doctors should sit with their patients and discuss how they use AI to help develop a diagnosis or treatment plan.”
A doctor using AI as part of the imaging analysis, or a note taker, should inform the patient and ensure that the technology understands and agrees to the parameters used. People also want to know that their sensitive medical data is stored and protected safely.
“The informed consent process does not disappear from the use of AI in patient care. In fact, the ethical obligations of physicians remain the same, and it is to continue to apply to patients the tools that patients use to determine treatment recommendations,” Freeden said.
Most importantly, we ensure that these tools are always used with caution and evaluated by humans. It helps patients feel safer and helps doctors avoid the risks associated with overrely relying on their technology.
“There is certainly reason to be optimistic about what AI brings to healthcare, but most patients think they can be more comfortable with humans,” Friedon said. &
