8 ways ChatGPT shows promise for healthcare
The generative artificial intelligence (AI) tool ChatGPT has made countless headlines since its launch at the end of November.
Generative AI ranks among the top digital health trends of 2023, and natural language processing (NLP) has begun to demonstrate its value in use cases ranging from automating documentation to powering customer service chatbots. Meanwhile, a Morning Consult poll found that 52% of Americans believe generative AI is “here to stay,” while workers in healthcare were twice as likely to say they have more to gain from generative AI than they have to lose.
Ashutosh Kaul, M.D., a professor at New York Medical College and a pioneer in robotic surgery, said tools like ChatGPT can play a pivotal role in augmenting clinical care as well. “As a surgeon, you can improve life one person at a time,” Kaul said, “but modern technology can do it on a vast scale in a cost-effective way. It’s so exciting to think about it.”
Three use cases represent low-hanging fruit, he said:
Task automation. Providertech estimated that 40% of the work done by healthcare support staff can be automated; Accenture pegged the figure at 17% for doctors and 51% for nurses. “That opens up opportunities, especially given the industry’s workforce shortages and skill gaps,” Kaul said.
Skin and eye care diagnostics. AI-assisted triages for rashes or pink eye, particularly after brick-and-mortar physician’s offices have closed, can save patients a late-night trip to the emergency department (ED).
Mental health. More than 20% of Americans experience a mental illness, but only 45% of these individuals receive treatment, according to Mental Health America. “We don’t have the manpower to take care of those numbers,” Kaul said – but generative AI models can better tailor interventions and broaden access.
Five other areas show promise.
Decision support. A preprint study from researchers at Vanderbilt University Medical Center found that ChatGPT “could be an important complementary part” of clinical decision support “and may even be able to assist experts in formulating their own suggestions for CDS improvement.” Of particular interest to Kaul is generative AI’s potential to parse and summarize patient notes in the ED. “Even if the records and labs are in the system, it’s going to take 15 minutes to go through them in detail,” he said.
Palliative care. One of the more difficult decisions at hospital discharge is whether a patient needs specialized medical care. With 1 in 6 Americans over 65, it’s also a decision that confronts a growing number of care teams. Using generative AI to scan a patient’s record and recommend a palliative care consult can increase access to objective assessments and improve decision-making, Kaul said.
The hospital at home. In the inpatient and ICU setting, automated alerts from remote monitoring devices help clinical teams monitor multiple patients and prioritize their responses. Kaul sees similar possibilities for home-based remote monitoring – along with a potential model for reimbursing clinicians for providing these services.
Wellness. Generative AI could address many questions patients have about their health that may not otherwise require an in-person appointment, Kaul said. Personal wellness is one area, as clinically approved applications could help patients develop healthier eating habits tailored to their dietary needs, cultural backgrounds, or other food preferences.
Continuing education. Standardized, in-person training programs don’t always meet the needs of busy healthcare workers, but resources available on demand can keep clinical staff up to date on changing guidelines or provide administrative staff with skill-building opportunities, Kaul said.
“It’s about transforming care so we can provide better, faster, more cost-effective and on-demand care,” he said.
Brian Eastwood is a Boston-based writer with more than 10 years of experience covering healthcare IT and healthcare delivery. He also writes about enterprise IT, consumer technology, and corporate leadership.