Artificial intelligence will continue to advance in health care administration and delivery
January 14, 2025
This is the fourth in a series on issues facing employer-sponsored health insurance in 2025. See the bottom of the post for links to previous posts.
Artificial intelligence, it seems, is everywhere. Artificial intelligence can monitor biologic waste in our toilets and provide a holographic nurse. Some promised advances from artificial intelligence include a Will smart umbrellas that tell us when it’s going to rain or a smart mirrors that measure health by assessing facial blood flow be a hit? Stay tuned.
There have already been substantial advances in artificial intelligence in health care delivery over the last year. AI is rapidly replacing scribes, listening to office visits, and creating notes, although these occasionally include invented information. Some physicians are using AI tools to prepare first drafts of notes and replies to patients. AI tools show more empathy than physicians, and mental health chatbots abound. AI tools can improve imaging readings, and an AI tool was able to make more accurate diagnoses than either physicians or physicians aided by AI, and can often suggest unusual potential diagnoses months earlier than treating clinicians..
In terms of health care benefits and finance, medical carriers and pharmacy benefit managers have used artificial intelligence to vet prior authorizations, which can speed decision-making and decrease resource costs. However, programs that do not include meaningful human oversight of denials have come under fire. Startups are offering AI tools to providers and patients to obtain approvals or complete appeals of denials -- suggesting that an AI arms race might not lead to meaningful decrease in administrative costs. AI can assist with fraud detection, and chatbots can help members understand their health plan benefits and find physicians or clinical trials to meet their needs.
Implications for employers:
Employers can inventory where AI is being used within human resources and their health benefit plans and deploy processes to protect member privacy and system security.
AI can sustain or even amplify bias from its training data, so employers should seek models designed not to increase disparities.
Meaningful oversight systems can identify and suppress “hallucinations” from AI models.
Many employees will need to be upskilled to maintain their edge in a world where productive employees must learn how to identify automation opportunities and effectively “prompt” AI to get a good answer.
Thanks for reading. You can find previous posts in the Employer Coverage archive
Please subscribe, “like” and suggest this newsletter to friends and colleagues. Thanks!
The posts in this series:
Employers and health systems will continue to face labor shortages
Regulation and policy changes could reshape employer sponsored health insurance
Coming later this week:
Artificial intelligence will continue to advance in health care delivery and administration
Climate change will adversely impact health and increase health care costs
Mental health will continue to require employer attention
Infectious disease will continue to pose serious risks
Illustration by ChatGPT