The rise of artificial intelligence in healthcare has become increasingly widespread, but with myriad choices and no specific framework, how can hospitals and health systems ensure its successful deployment and efficiency?
Avishkar Sharma, MD, director of artificial intelligence and a radiologist at Philadelphia-based Einstein Healthcare Network, says there is a one-word answer to ensuring the successful adoption of AI strategies into a health system: education.
Dr. Sharma emphasized the importance of immediate education during the initial onboarding stage of a new solution, then prioritizing reinforcement to prevent utilization from fizzling out as time passes. Along with initial education and reinforcement, internal conversations that offer the vendor continuous clinician feedback enables the product to be successful in user experience and patient outcomes.
"I know that AI is here to stay and will be part of my future career amongst many others," Dr. Sharma told Becker's. "It's imperative for people to become very informed and educated about the opportunities of AI, understanding its pros, cons and limitations, so people can make informed decisions about it."
"What will the impact be on day one or day one thousand?" Dr. Sharma told Becker's. "Is it something that's going to be actually timeless and something that will be everlasting, or something that's a short term, cool thing to have?"
The role of an AI director or leader within a hospital or health system is to become an expert in the subject matter surrounding each new technology. Having a designated source of information to answer any questions clinicians may have will assist in building trust and continued utilization of the tool.
After the initial onboarding process, Dr. Sharma said it is crucial to demonstrate the effective ways the tool can be used and pinpoint the specific areas it can be used within each clinician's workflow. Dr. Sharma estimates that around half of his colleagues are "on board" to try new AI technologies, but most medical professionals are data-driven and want to know explicit ways the product will make an impact.
Dr. Sharma accomplishes this by "gathering usage statistics and then showing the endpoints, whether its turnaround time, efficiency, improved patient outcomes." He says that "getting more personal anecdotal experiences" goes a long way, resonates with people and motivates individuals who have shown hesitancy to take the first step in adopting the technology.
Not only should clinicians be a part of the conversations by providing feedback, but Dr. Sharma said they should be governing the whole process.
"Clinicians need to be in the driver's seat with the integration of AI," Dr. Sharma said.
Clinician governance entails an organized structure with a step-by-step method that confirms new technology is screened, verified, properly implemented and continuously monitored, assuring the product remains as widely utilized and effective as it was "day one out of the box," Dr. Sharma said.
Executive buy-in is also important. Oftentimes, C-suite leaders at large organizations do not have "boots on the ground," making it nearly impossible to know what is making a clinical impact and make informed, educated decisions surrounding creating a framework without prioritizing clinician feedback and experiences.
In addition to ongoing internal conversations among C-suite leaders, vendors and clinicians being vital, Dr. Sharma said these relationships are most advantageous when it is informal within formalized collaboration. Essentially, he said the relationship should act as a "constant pipeline of communication," allowing clinicians to feel comfortable reaching out at any moment to explain what is and is not working.
Dr. Sharma said leaders and clinicians must keep in mind that a very small fraction of staff using the tool at the beginning does not mean it is a failure. Instead, it presents a new collaboration opportunity in which vendors should use the experiences and data to help and nurture the usage of the product.
Einstein Health partners with Aidoc, a healthcare-focused AI organization, and Dr. Sharma attributed its implementation success to the willingness and excitement Aidoc displayed in receiving clinical feedback from deployment. Extensive and uninterrupted collaboration has allowed the effective utilization of various Aidoc tools that "span the entire imaging lifecycle," Dr. Sharma said.
Successful AI implementation at Einstein means improving clinicians' mental health and cognitive load, while having patient care at the forefront. Dr. Sharma said the interpretation of imaging and automation of tasks acts as a "safety net" or "second pair of eyes" that boosts clinicians' confidence in decision-making, in turn easing their cognitive load and improving their mental health.
Dr. Sharma credited the upcoming generation with a heightened awareness of the need for a healthy work-life balance. He said with "full confidence" that AI is "not going anywhere," and he is hopeful that the cognitive load and mental health of clinicians will be a priority for the next generation of key decision-makers when it comes to AI integration.