Artificial intelligence
Unlocking AI’s full potential in healthcare: A tailored approach
September 30, 2024
Imagine a world where high-quality data is seamlessly provided for your clinical decisions, administrative tasks disappear, and your attention is fully on your patients. That world could rapidly take shape, driven by the introduction of artificial intelligence (AI) in healthcare.
But as exciting as this vision sounds, AI is a fast-evolving, moving target – what works today could be completely transformed in just six months.
To truly harness AI’s potential, healthcare providers must cut through the noise, adopt tools that address specific challenges, understand the shortcomings, and ensure proper use of the technology to safely enhance care. The key is a thoughtful, selective approach, based on each healthcare practitioner’s individual needs.
Exploring the AI landscape: At MEDFAR, Canada’s fastest-growing electronic medical record (EMR) provider, we piloted several promising AI solutions. These tools showed significant potential, especially in areas like documentation, with many physicians offering positive feedback.
However, some physicians found that, in certain situations, non-AI solutions were still more efficient. “When a physician has a well-organized library of pre-made clickable texts for medical conditions, the assessment and plan of a note can be completed much faster than with a generative AI response,” noted family physician Dr. David Harrison from British Columbia.
Overall, physicians expressed varied preferences for different AI solutions, with no single tool emerging as the clear favorite across the board. Each AI solution demonstrated unique strengths, raising a key question: which one should be integrated into EMRs?
A study published in NEJM Catalyst earlier this year emphasized that targeted AI use is associated with improvements in documentation efficiency and favorable feedback from both patients and physicians without EMR integration. Hence, AI seems to offer tangible benefits now, irrespective of EMR integration, provided one chooses the right tool and applies it safely and effectively.
A key finding from the article highlighted that, although 10,000 physicians and staff were granted access and offered training for the ambient AI scribe, only 34.4 percent had used the application within the 10-week study period.
On average, those who used the AI scribe enabled it for 88 patient encounters during this time.
The authors noted that the rollout encountered several barriers to adoption, which must be carefully identified and addressed to ensure a successful implementation.
The Future of AI in healthcare: While AI-driven scribing tools have gained significant attention, the future of AI in healthcare extends far beyond documentation. AI holds the potential to elevate every stage of the patient journey, from streamlining triage and paperwork to extracting and organizing data, enhancing patient education, and optimizing follow-up care.
Dr. Tahmeena Ali, a family physician and medical director at Catalyst Kinetics Medical in Surrey, British Columbia, highlights AI’s ability to organize patient information in one place, just in time for the physician.
This allows doctors to fully grasp the patient’s condition and make more informed decisions. “AI could review a patient’s reason for visit and medical history to provide useful insights for the physician’s consideration,” she explained. “It could also send educational materials to patients after their visit, helping them better understand and manage their health.”
AI’s ability to sift through the fast-growing medical literature and instantly deliver relevant, personalized insights for each patient could prove to be a game-changer. “Physicians can’t keep up with all the new evidence emerging in real-time, but AI can,” said Dr. Mark Karanofsky, a family physician and unit director at the Goldman Herzl Family Practice Centre in Montreal, Quebec.
“These tools can guide us in identifying the best treatments and risks, optimizing our time and improving the quality of care.”
Re-centering the doctor-patient relationship: One of AI’s greatest potentials lies in re-centering the doctor-patient relationship. Today, physicians are overwhelmed with administrative tasks like charting, data entry, clicking through disjointed IT systems to fill out forms and requests.
They are also managing inboxes, coordinating follow-ups and transfers, navigating through all sorts of distracting system alerts, answering non-urgent patient and administrative queries, reconciling requests with results, which detract from their face time with patients. AI can help alleviate this burden by automating much of the administrative work, freeing up more time for doctors to engage directly with their patients.
“As AI tools become more reliable, I believe they will not only save time but also improve the advice and treatments we offer,” Dr. Karanofsky states. “AI will allow doctors to focus on their patients rather than their computer screens, helping regain the human connection that many feel has been lost in the transition to digital systems.”
Dr. Ali shares this vision, noting that AI can streamline workflows, reduce errors, and improve patient experience. “AI can automatically update charts, file lab reports, and prepare consultation letters, all of which contribute to more efficient care,” she explains. “In the long term, AI could reduce clinical errors and increase provider satisfaction by lightening the administrative load.”
Addressing concerns and challenges: Despite AI’s potential, significant concerns persist – especially regarding data integrity, accuracy, privacy, and tool reliability. Physicians like Dr. Karanofsky emphasize the need for robust validation and transparency in AI applications as regulatory agencies begin voicing their stance.
“Both clinicians and the public must have confidence in these tools. How they’re validated and who will assume responsibility for their recommendations are critical issues,” he noted.
The failure of past attempts to revolutionize healthcare has shown that technology alone cannot drive success. The adoption of AI must be carefully managed, with healthcare providers fully supported in adapting their practices to make the most of new tools. This will require collaboration between tech companies and healthcare providers to ensure that AI solutions are practical, safe, and tailored to meet real clinical needs.
Dr. Robert Amyot is the Chief Operating Officer at MEDFAR.