The growing use of artificial intelligence and the explosive presence of generative AI (like ChatGPT) are lighting up the realm of healthcare information technology. The keys are where to use AI and how to safely use it.
Dr. Nirav S. Shah is medical director of quality innovation and clinical practice analytics at NorthShore – Edward-Elmhurst Health. He is attending HIMSS23 this week. His organization uses AI – and uses it quite successfully.
We interviewed Shah to get his expert views on and predictions for artificial intelligence, and what his expectations are for HIMSS23.
Q. As healthcare provider organization executives such as yourself gather for HIMSS23, what do you feel is the most important issue in healthcare information technology?
A. There are a number of macro trends occurring in healthcare today, which are converging to bring specific IT issues to the forefront.
Recently, the president authorized the end of the COVID-19 public health emergency, and the aftereffects of the pandemic have had wide-reaching implications that have rippled across the healthcare system.
Trends that existed prior to the pandemic have been accelerated. Clinicians still are struggling with burnout, and many continue to leave the field, and this attrition is exacerbating the strain and burden on providers and health systems – affecting their ability to deliver care.
Meanwhile, the population is aging, has gotten sicker, and there are more mental health issues.
So, the lapse in care over the last few years during the COVID emergency has created a void that is hard to fill in an environment facing a workforce shortage and an epidemic of burnout. In the end, it all comes down to care enablement, or empowering and optimizing our care providers to work at the top of license and deliver safe, seamless and personal care.
Healthcare information technology has a major role to play in enabling our care providers, and with the recent advances in AI, specifically with large language models and generative AI, we are seeing the automation and augmentation of not only low-value tasks, but increasingly higher-value tasks.
Q. What issues are you facing in your organization that you’re hoping to find help for from the HIMSS23 educational sessions and exhibit hall?
A. So, circling back to those macro issues, we’re looking for tools leveraging AI that will enable the workforce to be more efficient and more effective – and that will bring back the joy in clinical care. A key element is how these tools are integrated into workflows and solve the last-mile challenge.
A highly accurate tool is only as good as its ability to integrate into provider workflow and augment care processes.
Recently, we’ve partnered with IQVIA on their natural language processing platform, and our analytics and informatics team from NorthShore – Edward-Elmhurst Health gave a presentation at HIMSS23 entitled “Implementing AI NLP Tool to Address SDOH Needs.”
This talk is focused on our work solving for the last-mile challenge and enabling frontline clinicians to be able to take care of patients more effectively with the assistance of AI.
Our first use case using this platform was in social determinants of health – and the ability to extract SDOH care gaps from our clinical notes, pushing that insight back into the electronic health record, and bringing it to the forefront for our clinicians – specifically our ED social workers – so they can be more efficient and effective in delivering care.
Our ED social workers traditionally leveraged proxy measures for SDOH gaps, such as predictive analytics around readmission and mortality. While this method identifies high-risk patients, these patients don’t necessarily have clear and addressable SDOH needs and as such our ED social work team was spending a considerable amount of time trying to identify patients that require their assistance.
With the NLP platform and implementation, which our ED social workers have labeled as a game changer, they are finding they are better able to target patients who have SDOH gaps and are spending more of their time making a difference.
Q. What do you think are a couple of the major health IT challenges in the year ahead for CIOs, CMIOs, CISOs and other health IT leaders at provider organizations?
A. Nationally we are seeing strains in our workforce’s ability to care for patients. These strains are manifesting in access and the ability of our overburdened care teams to address additional needs and challenges that may arise.
We need to think thoughtfully about ways to lessen clinical burden, whether through automation and augmentation with AI, scaling digital tools to outsource and make care teams more efficient, and even de-implementation of low-value practices.
This year, there will be continued interest and focus on the applications of generative AI. These tools have the ability to save time in documenting clinical notes and patient instructions, assisting with medical coding and patient scheduling, and summarization of large volumes of data.
The abilities of these models are astounding, and, while the potential is vast, they will bring challenges to health IT leaders. How do you manage a life cycle where these models are becoming exponentially more powerful within months? What are the security implications of using these tools? How best should these be integrated into workflows? How do you deal with the black-box nature of these models, and how do you safely deploy these tools in patient-facing applications?
It’s an exciting area, and although it’s still just getting started, its importance will only increase in the year ahead.
Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email the writer: firstname.lastname@example.org
Healthcare IT News is a HIMSS Media publication.