![](jpg/gettyimages-1280895783-150x150.jpg)
“AI may be on its way to your doctor’s office, but it’s not ready to see patients,” a Los Angeles Times article proclaimed this spring. It’s not the first to remind us: For all the hype around ChatGPT and other new AI tools, we’re still a long way from widespread adoption.
But there are AI tools available for healthcare now that can demonstrate return on investment (ROI) in months, not years. These solutions strengthen referral processes and relieve pressure on overstretched healthcare teams while increasing capacity where it matters most. They also help eliminate the administrative demands of care that fuel burnout, speed information transfer, and ease transitions in care.
Here are five critical AI investments that provide “right now” value for health systems, clinicians and patients.
1. Natural language processing (NLP) and artificial intelligence (AI)-powered tools that strengthen referral processes. When physicians make referrals to specialists, the latest NLP AI advancements combined with digital fax technology ensures that no one worries about what will happen to their e-faxed referral request. For instance, when NLP and AI are applied to digital faxes, these unstructured documents can be transformed into structured, searchable data that EHR applications can digest. Then, using an integration engine, the structured data is automatically matched to the right patient’s record so providers can act on the information.
Capabilities like these put providers on another plane. They help build stronger relationships with referral partners because all the information they need is routed in their preferred workflows and in a structured format. As a result, referral workflows no longer take hours, days, or weeks. It’s an approach that works even when a partially illegible, digital fax is submitted. At one hospital, use of AI and NLP for gastroenterology referrals, which are particularly complex, allowed the hospital to automatically triage 40% to 50% of urgent-suspicion-of-cancer referrals.
2. NLP and AI solutions that reduce workforce burden. Today, 45% of inpatient nurses say they are likely to leave their role in the next six months, in part due to an unmanageable workload. That’s in part due to the amount of time nurses spend per shift hunting for information, equipment or supplies (43 minutes), communicating about care for patient hand-offs (60 minutes), and completing administrative and logistical tasks (97 minutes). Today, NLP and AI can transform handwritten or text data into a structure that can be consumed by any IT system—including the EHR—and conveyed to staff in ways that complement their workflow.
Given the staffing crisis health systems currently face, reducing ineffective and redundant workflows can be crucial. It’s an area where NLP and AI hold strong potential to make a difference in improving nurse workload and their ability to manage patients requiring complex care. In fact, nurses believe 42% of the time they spend per shift could be reduced by nearly half through tech-enabled processes, including the use of intelligent automation.
3. NLP AI solutions that extract matching data for clinical trials. Recently, the Fred Hutchinson Cancer Center in Seattle leveraged NLP AI technology to match patients with clinical cancer studies, combing through unstructured data at the rate of 10,000 medical charts per hour to find patients who met clinical trial inclusion criteria. According to one expert, fewer than 5% of patients match the recruitment criteria for these trials. One barrier to participation: identifying the right patients amid large volumes of unstructured data.
4. NLP and AI technology that strengthens collaboration among medical teams. At Children’s Hospital of Philadelphia, NLP AI is used to structure clinical, genomic, and imaging data, enabling researchers to cross-analyze diseases and intelligently extract medical insights for children across a wide spectrum. These new medical discoveries can change lives. And at Boston Children’s Hospital, an NLP AI laboratory explores use cases for applying this form of AI to research projects, including pharmacogenetics research, and answering clinical questions at the point of care.
5. AI-powered solutions that speed information transfer during transitions in care. The handoff of patients from one care setting to another is one of the most difficult challenges providers face. It’s also the point at which the potential for error dramatically increases, especially as the volume of high-acuity referrals for post-acute care rises, leading to more complex cases. Yet most skilled nursing facilities and post-acute care facilities lack meaningful connectivity with their referral partners—typically hospitals and health systems. More than half say they receive patient information after the patient is in their care. Even when the information does arrive, 76% say at least a portion of the data isn’t usable or it’s incomplete. This not only delays admissions, but also prevents patients from receiving critically needed care.
That’s why NLP AI solutions that speed information transfer and enable clinicians to easily extract actionable structured data from unstructured digital faxes can improve patient outcomes.
It’s difficult to believe that in today’s post-Covid era, seven out of 10 healthcare organizations still rely on paper faxes to exchange patient data. Digital cloud fax solutions can be combined with NLP AI technology to flag specific actions needed—such as orders for hospice care—and enable post-acute staff to promptly receive and act upon the extracted information appropriately. Since printed paper discharge plans may be up to four inches thick, digital cloud fax solutions help accelerate the intake process, and increase efficiency, despite limited staff.
One recent study also shows NLP AI technology can be used to identify barriers to post-acute care referrals, including patient preferences, prior to hospital discharge and work to address them in collaboration with families.
Making the right moves for AI value today
In 2016, Geoffrey Hinton, a British-Canadian cognitive psychologist known as “the godfather of AI” insisted that within five years, radiologists would be replaced by AI. “People should stop training radiologists now. It’s just completely obvious that, within five years, deep learning is going to do better,” he said during a conference. That’s an AI hype prediction that didn’t hold up with time.
For healthcare leaders, it’s a cautionary reminder: Don’t bet so big on AI solutions with “someday potential” that you ignore tools that can solve challenges and generate ROI now. By examining NLP AI tools that relieve pressure on clinicians and ensure access to the right information at the right time, your organization can make a substantial positive impact in care quality, safety, and value—and increase satisfaction among all stakeholders.
Photo: Natali_Mis, Getty Images