Asheville – Two months ago, Mission Health HCA published a blog discussing the use of AI in cancer therapy. The article explained how computers can process pathology reports and analyze biopsies more thoroughly, accurately, and quickly than humans. AI can even enhance images to identify and analyze elements that are invisible to the human eye. For treatment planning, it can perform predictive modeling through statistical analyses based on databases of patient genetics, medical histories, and lifestyles. However, the post did not indicate that Mission was employing any of these technologies.
One reason nurses at Mission Health HCA went on strike last year was their desire for assurances in their contract that the hospital would not attempt to replace their skills and training with AI. The public justifiably fears AI in healthcare, viewing it as too nascent to make life-and-death decisions. Characteristics such as “bedside manner,” “trust,” and “relationships” have long been associated with good doctors, with “discernment” and “discretion” being paramount. In the realm of computers, it’s a case of GIGO (garbage in, garbage out), and patients often do not know who is writing the algorithms. They can only hope that the AI used in hospitals is superior to the AI on their personal devices.
Nonetheless, hospitals are discovering effective ways to utilize AI for positive outcomes. Several institutions have assigned AI the task of completing prior authorizations. These forms, which are a source of frustration for the AMA, must be filled out to convince insurance companies that the doctor genuinely intended to prescribe what was prescribed. The paperwork is tedious, and the time spent on hold is excessive; large institutions have even funded dedicated staff, while patients have died waiting. The ideal solution would be to reform or eliminate prior authorizations altogether. Until then, they remain a burden that drives small hospitals unable to afford compliance out of business.
AI can interpret vast amounts of data in real time and compare it with high precision against a database of known possibilities. It can rapidly analyze every bump on an ECG, catch nuances that humans might overlook, and access a database of documented differential diagnoses. Spokesperson Nancy Lindell was asked how Mission was adopting AI to improve patient outcomes and reduce healthcare costs. She stated she would look into it but did not respond before press time. Consequently, the question was directed to AI itself.
Disappointingly, AI appeared to conflate itself with ordinary robotics, automation, and electronic communications. For instance, search results included news about a new electrophysiology lab from 2012 and increased telehealth usage from 2017. In 2018, when Mission was still a nonprofit, it received accolades for using machine learning to predict readmission risks. Last year, HCA promoted how it is proactively adopting AI systemwide—not necessarily at Mission. Its Digital Transformation and Innovation (DT&I) department is integrating AI into telehealth and utilizing virtual assistants to triage calls, access medical records, and schedule appointments.
HCA employs robotic process automation to assist with other “repetitive, rule-based tasks” like billing; uses AI with mathematical optimization techniques for nurse scheduling; and applies machine learning, anomaly detection, predictive analytics, and more for fraud detection. It is also leveraging AI to track and manage supply chains. This not only reduces waste but also prevents patients from waiting during times of trade wars, pandemics, and hurricanes.
For clinicians, HCA stated that it is integrating “AI-driven decision support systems” into its electronic health records (EHR) systems. These systems utilize patient medical records and clinical guidelines to recommend subsequent diagnostic tests and treatments. Additionally, AI is expected to extract summaries from clinical notes to “ensure more accurate and comprehensive patient records.” AI is also employed to analyze radiological images with greater speed and accuracy than human counterparts. Last year, Mission announced its use of AI with CT scans in its pediatric emergency room, providing improved analysis with reduced radiation exposure.