Dr. Paige Kilian, Chief Medical Officer at Inovalon, bridges clinical medicine and data analytics to improve real-world healthcare outcomes.
Everyone is trying to forecast AI’s effects on healthcare. Some believe it will finally free medical professionals from mundane administrative tasks so they can spend more time with patients.
Others worry it will erode the hands-on expertise required to deliver comprehensive, compassionate medical care. As AI creates the illusion of an easy approach to diagnosis, physicians will have fewer opportunities to learn by experience, potentially causing overreliance on technology at the expense of patient outcomes. Autopilot could not have landed Captain Sully’s plane.
Neither reflects what’s happening in most encounters today. Right now, physicians are overloaded with nonclinical tasks, patients are waiting too long for care and the playing field continues shifting as more regulations are announced and new technologies enter this space.
Fortunately, there are clear examples of AI catalyzing the change healthcare has long needed. I’ve observed them in my own day-to-day work. Models have produced exciting wins in taking on time-consuming, tedious tasks to help skilled experts make patient encounters more meaningful and, in some areas, produce better outcomes. AI scribes and other ambient documentation tools are giving physicians back time to focus on patients, and AI-based alerts are catching patient safety risks earlier.
The challenge is embracing innovation while preserving the art of taking the patient’s history, factoring in the nuance of expression and inflection and performing a solid physical exam; AI disruption has to happen in parallel to continued medical learning.
New Hope For Overburdened Physicians
Forty-two percent of physicians report experiencing burnout, with administrative work as a primary contributor, a top reason why many physicians, nurses and pharmacists are leaving their professions.
Ambient AI can help ease documentation burden and give physicians time back to focus on patients. Doctors can turn away from the computer, look patients in the eye and still capture notes for later. For non-English-speaking patients, ambient AI could also be a valuable tool in improving communication with their providers.
The result is a more meaningful interaction without sacrificing the documentation accuracy needed to inform billing, quality measurement reporting and risk adjustment programs. In one study, clinician burnout fell 13% after 30 days of ambient use, and another study found that ambient scribes reduced cognitive load for physicians.
If agentic AI is added to providers’ toolkit, it can further offset administrative burden by helping appeal denials or complete prior authorizations. This is one area where early wins are signaling for healthcare teams to move faster. Thoughtful increased use of agentic AI can cut the costly inefficiencies taking up time and resources, allowing clinical teams to see their next patient sooner without sacrificing documentation needs for clinical care, quality reporting and prior authorizations.
Improving Patient Safety
In one study comparing AI-based recommendations for empiric antibiotics against what clinicians prescribed, the model chose the correct antibiotic as well as, or better than, the clinicians’ choices—and accounted for resistance profiles and chose narrower-spectrum antibiotics, reducing the risk of antibiotic-resistant infections.
In a cardiovascular health clinical trial, AI-enabled electrocardiogram (ECG) alerts reduced all-cause mortality rates within 90 days. AI alerts sent physicians a report of patients predicted to be at high risk of mortality, producing a 3.6% mortality rate in the intervention group compared to 4.3% in the control group. Separately, a systematic review of 11 randomized cardiovascular AI clinical trials found that five contributed to improvements in clinical events, six produced enhanced diagnostic accuracy and early detection and three yielded better resource utilization.
Balancing AI Advancements With Hands-On Care
As AI is increasingly used in patient care, healthy skepticism must remain to prevent it from curtailing the experience on either side. Clinical experts still have to build their knowledge and apply best practices to each patient’s unique needs while also protecting patient comfort. AI might influence how they express themselves, making the combination of skilled medical practice and attuned bedside manner even more important.
For example, a patient who knows AI is transcribing their conversation may be less confident that the encounter is confidential. They might hold back or answer questions less honestly, and physicians need to catch these subtleties. AI can’t pick up if a patient isn’t looking their doctor in the eye or fidgeting while talking, yet these are the signals that create opportunity to foster connection and uncover important details about a patient’s wellness.
Clinicians must also prevent a diagnostic cascade. If doctors order tests based on AI recommendations, without performing a physical exam as well, one set of imaging could beget another test, more follow-up visits and potentially more invasive testing, which may not be warranted. Tests must be ordered based on sound clinical assessment, including an experienced physical examination that considers the patient’s unique history.
AI can’t replace decision-making in a particular encounter or in the years of practice required to build a personal medical perspective. Physicians must remain dedicated to their craft, with AI as a tool informing them in delivering care—and they should also consider how a patient’s behavior may change when they’re aware that AI is part of their care team.
The Holy Grail Of AI In Healthcare
The more AI takes on healthcare’s operational challenges, the more imperative it is for physicians to reinvest administrative time in continuing to sharpen their clinical skills. AI and human expertise need to progress together, with professionals staying committed to medical principles.
We’re seeing promising examples in lowering burnout and improving patient safety largely because the technology in these instances is taking on the complicated, task-heavy side of care and empowering clinical experts to focus on what the person in front of them actually needs. That’s the ultimate goal of AI use throughout the care journey: building highly skilled AI to collaborate with highly skilled humans to deliver the highest caliber, most compassionate healthcare possible.
Humans must remain at the controls of healthcare delivery, capitalizing on all the benefits technology offers while keeping it in an appropriate, clinical and caring context. There is no role for autopilot in the healthcare journey.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?