Healthcare systems across the globe are grappling with the stress of an aging population. The World Health Organization has announced a worldwide shortage of healthcare workers. Technologists see generative Artificial Intelligence (AI) as a means to reduce the burden on the medical profession and allow physicians to focus on care that only human doctors can perform.
Plunging Workforce Insufficient to Address New Healthcare Demand
In the United States, 72,000 American doctors exited the workforce between 2021 and 2022, pushing the healthcare crisis to the brink, especially in more rural and underserved communities. Countries like South Korea are relaxing entrance standards for medical schools to increase the student population to care for the elderly.
Impact on Patient Care
Patients are the main losers in the current unbalanced healthcare marketplace. Doctors have increasing workloads, and the quality of care patients can expect is decreasing proportionally.
Promise of AI in Healthcare Workforce
Some believe that generative Artificial Intelligence is ripe for adapting to the needs of the healthcare space, with a promise of saving time and resources of overextended medical professionals. Specifically, AI could potentially increase efficiency in standard uniform practices, allowing physicians to focus on improving clinical outcomes.
Understanding Generative AI
AI has been utilized in healthcare for many years, as technologies have allowed practitioners to diagnose and forecast treatment success with computer vision and predictive algorithms. Generative AI has not yet been widely applied in the healthcare industry. Generative AI would allow neural networks to identify patterns from existing data and create entirely new phenomena. This possibility holds promise to expand the efficacy of healthcare, just as it is likely to transform other industries as well.
AI’s Role in Healthcare
Practitioners of AI could potentially foresee the AI capability simplifying patient journeys and clinical documentation, providing relevant information during procedures and surgeries, and augmenting telehealth capabilities. AI cannot completely replace physicians, but it could simplify and create pathways in a bottlenecked medical care pipeline.
AI Assistants for Medical Guidance
Since the onset of the COVID-19 Pandemic, many healthcare practitioners began to offer remote and telehealth visits. Like many other industries that increasingly rely or offer the option of lower-cost robo-advisors, it is possible that generative AI can create assistants to handle non-specialized and general medical cases. These AI assistants could guide patients to appropriate medical treatments, creating a gatekeeper for physicians while still addressing patient needs.
Pharmaceutical Companies Employing AI Assistants
Drug companies such as Sanofi, Bayer, and Novartis have already started utilizing AI assistants.
Streamlining Administrative Work
For non-physician-specific tasks, Generative AI chatbots can and sometimes already do complete automated tasks such as booking appointments and sending reminders. Healthcare organizations such as Mercy Health, Baptist Health, and Intermountain Healthcare utilize such AI assistance for patient registration, scheduling, and prescription refills.
Enhancing Clinical Documentation
An area beyond the administrative task list in which AI can assist in the healthcare field is where generative AI assistance can listen to physician conversations with patients and summarize clinical notes. This reduces the amount of tedious manual work by physicians. These AI-generated notes can then be used to create instructions for patients and a personalized treatment plan, which includes previous notes on the patient’s record.
Data Retrieval in Workflow
Healthcare organizations will quickly be able to craft AI assistants that are “smart” and self-instructing without retraining, as well as be able to provide contextual answers. This can be achieved by utilizing large language models (LLMs) and retrieval augment generation (RAG). These contextual answers will access additional data resources without the human interaction required to retrain the assistant. The assistant should be able to produce evidence-based recommendations.
Data Analysis and Report Generation
When it comes to analyzing CT scans, MRI’s and X-rays, generative AI capabilities could automate the basic level analysis, thereby creating time efficiencies and reducing effort required from over-taxed medical professionals. Experimentation at this early stage reveals AI-generated reports return results at least comparable to the quality and accuracy of analyses produced by radiologists.
Advancing Drug Development
Pharmaceutical developments could also be potentially impacted by generative AI capabilities, as these technologies are able to analyze complex medical data to identify new chemical combinations and molecular structures. This capability could result in new drug candidates. The technology could also predict side effects as well as interactions with other pharmaceutical products. As an example, Insilico Medicine currently has a drug product for pulmonary fibrosis in clinical trials that was generated by AI.
The Need for Caution
There are two main areas in which great caution must be exercised in the use of generative AI in healthcare. One is the truth that generative AI is only as reliable as the data with which it has been trained. The healthcare inputs must be of the highest accuracy and quality, and that certainty is nearly impossible to ensure.
Balancing Technology and Human Expertise
Second, generative AI and chatbots will never be able to replace the human touch in patient care. Both a listening, empathetic and attuned ear of a physician is unreplicable, as is the expertise of a surgeon in emergency surgery. Medical technologies must always be overseen and approved by actual trained medical professionals.
State Actions to Regulate AI in Medical Care
In California, the legislature has taken up measures requiring healthcare providers to inform patients when they are interacting with AI assistants. The state of California also is looking into limiting insurance providers’ ability to use AI when making decisions on coverage.
Georgia, Utah, Colorado
In Georgia, bills are making their way through the legislature to prohibit medical providers from relying exclusively on AI when making decisions on patient care. Other states such as Utah and Colorado are also considering policy implications of AI on patient care.
