AI Could Perform Up to 22% of U.S. Outpatient Care, McKinsey Analysis Finds

Artificial intelligence could potentially perform services represented in 16% to 22% of outpatient claims in the United States, according to a new analysis by McKinsey & Co. These services account for an estimated 13% to 19% of total outpatient spending.

An analysis released Sept. 28 by McKinsey & Co. examined 2024 commercial, Medicare and Medicaid claims and found that approximately 2 billion to 3 billion outpatient claims involve services that clinical artificial intelligence could potentially perform.

McKinsey describes this category as “discursive care,” referring to reasoning- and dialogue-based clinical activities such as patient intake, triage, diagnosis, referrals and follow-up that do not require a hands-on physical examination.

Low-Complexity Visits Represent a Significant Share

Evaluation and management visits involving no additional care, or only low-complexity care, represented approximately 11% to 15% of outpatient claims and accounted for 8% to 12% of outpatient spending, according to the analysis.

The interpretation of diagnostic and imaging results represented another 5% to 7% of both outpatient claims and spending.

McKinsey said physician offices, clinics, virtual care settings and urgent care facilities are likely to experience the largest share of AI-driven changes as clinical AI capabilities continue to develop.

AI Could Transform Tasks Across the Healthcare Workforce

The analysis also estimates that AI could perform a substantial portion of core tasks currently carried out by approximately 11 million of the 19 million healthcare workers included in the analysis.

However, McKinsey does not anticipate that the overall healthcare workforce will necessarily shrink. The firm projects that net demand for healthcare services will continue to increase, potentially offsetting productivity gains created by AI.

The impact could nevertheless vary considerably depending on healthcare payment models.

Fee-for-service providers could face reduced volumes of low-acuity visits as patients increasingly turn to AI-supported tools for certain types of care. By contrast, organizations operating under value-based care arrangements could potentially use AI to improve care-management economics and allocate clinical resources more efficiently.

McKinsey also suggests that payers could benefit significantly if AI becomes a patient's “front door” to the healthcare system, potentially changing how patients enter, navigate and access care.

Deskilling Identified as a Potential Risk

The analysis also highlights a potential clinical workforce risk: deskilling.

According to McKinsey, clinicians who rely heavily on AI without deliberately maintaining and practicing their own clinical skills could gradually lose proficiency in tasks that AI increasingly performs.

The findings suggest that the impact of clinical AI may extend beyond automation. As AI becomes integrated into intake, triage, diagnosis, interpretation, referrals and follow-up, healthcare organizations may need to reconsider workflows, workforce roles, payment models and approaches to maintaining clinical expertise.

Kaynak: Becker's Hospital Review