Clinical accuracy still requires clinical intelligence.
With decades of experience providing clinical care and running health organizations, our team at Onpoint is enthusiastic about advancements in AI. The ability to streamline EHR orchestration, synthesize mountains of data, and augment real-time care is a future we’re all hopeful for, especially as health systems are exhausting human resources to prop up broken workflows. From pre-visit through post-visit, the opportunities for intelligent automation are vast but the risks of getting it wrong are too great to ignore.
From pre-visit through post-visit, the opportunities for intelligent automation are vast but the risks of getting it wrong are too great to ignore.
When it comes to black box solutions, if it sounds too good to be true, it probably is. Every day there’s a new point solution out there aiming to solve one piece of the puzzle, and adding a whole new dimension of problems for our clinicians, administrators, and IT departments. And beyond the operational traffic jams, let’s not forget that patient lives are on the line. Outside-in point solutions tend to over promise on capabilities and brush the risks under the carpet. They don’t see the forest for the trees and their business models depend on rapid scaling. It’s marketshare now, fix the issues later.
Take what just happened with Microsoft’s Nuance, an AI-only solution. A class action lawsuit has been filed against Geisinger, a Pennsylvania-based healthcare provider using the tech, following a data breach that affected approximately 1 million patients1. The lawsuit alleges that both Geisinger and Nuance were negligent in protecting patient data and the plaintiff is seeking over $5 million in damages.
The risk to patient data is just one of many safety concerns—there are countless examples of AI impacting patient outcomes2. AI systems are susceptible to errors, both in their generation of outputs and their interpretation of data. These errors, known as “hallucinations,” can lead to misdiagnoses or inappropriate treatment plans3. Furthermore, AI algorithms are only as good as the data they’re trained on. Biases in the training data can lead to “errors of omission” where critical information is overlooked4.
The immediate benefits of smart AI integration are undeniable and unbelievably exciting. At a time when our health systems are dealing with doctor burnout and rising costs, AI, when safely implemented, offers a path to survive and then really thrive.
All of this said, the immediate benefits of smart AI integration are undeniable and unbelievably exciting. At a time when our health systems are dealing with doctor burnout and rising costs, AI, when safely implemented, offers a path to survive and then really thrive5. AI can automate repetitive administrative tasks such as scheduling appointments, processing insurance claims, and aggregating data and can free up valuable time for us to focus on patient care. Additionally, AI can analyze large volumes of data to identify patterns and trends, enabling more informed decision-making and optimizing resource allocation.
Clinically, AI decision trees can help standardize healthcare delivery by providing consistent clinical guidelines and protocols across complex patient data. This standardization ensures that all patients receive the same level of care, regardless of their complexity, the healthcare provider, or geographic location.
But all of this promise won’t be delivered on if we don’t take a measured approach. The complexity of healthcare demands professional oversight and this powerful tool can’t yet be fully trusted to replace the critical thinking and judgment of trained healthcare professionals—even if startups and corporations will have us think otherwise.
AI is already becoming a valuable ally in improving our healthcare systems but it must be integrated responsibly, and offer unmatched clinical accuracy. That’s why, at Onpoint, we provide an AI solution that keeps clinicians in the loop. Continuous adaptation and clinical intelligence are non-negotiable as we transform our health systems to meet the challenges of today with the promise of tomorrow. After all, in healthcare, we are under oath. First, do no harm.
About the author
Clay Hall is a seasoned Primary Care Nurse Practitioner and clinical leader at Onpoint Healthcare Partners with a deep-rooted commitment to safely advancing AI in healthcare. With firsthand experience navigating the complexities of modern healthcare delivery, he recognizes the transformative potential of AI while understanding its challenges. Leveraging his extensive clinical background, Clay is dedicated to guiding healthcare systems towards successful AI integration with a safety-first mindset.
References
- Geisinger sued by patients over vendor data breach (Becker’s Health IT)
- How AI Errors Impact Patient Care (Healthcare IT News)
- How AI Hallucinations Looked in 2023 (Forbes)
- AI Model Training Challenges (Oracle)
- What Will AI do for Healthcare? (Boston Consulting Group)
