The American College of Physicians (ACP) has issued a policy position statement outlining 10 recommendations regarding the use of artificial intelligence (AI) and machine learning techniques in healthcare applications such as clinical documentation, diagnostic imaging, and clinical decision support. Annals of Internal Medicine.
While the use of AI tools in healthcare has steadily increased since the 1970s, the U.S. Food and Drug Administration approved more AI technologies for clinical application between January 2020 and October 2023 than in the previous 25 years. ACP recognizes that the recent proliferation and broad applicability of generative AI tools has led to increased interest and enthusiasm for their clinical use, creating a need for oversight and regulation.
In its first advisory statement, the ACP stated that AI-based technologies “complement, not replace” the reasoning and decision-making of physicians and other clinicians, whose training and observation must remain the primary focus of patient care. Additionally, the ACP clarified that AI technologies must be consistent with principles of patient care, clinical decision-making, the patient-physician relationship, and medical ethics to enhance equity and justice in health care.
The ACP recommends that patients, physicians, and other clinicians be informed when AI tools are being used in treatment or decision-making, and therefore advocates for transparency, clarity, and education for health care providers about where AI-generated information enters their workflow. The group states that privacy and confidentiality of patient and clinician data should be prioritized, and clinical safety, effectiveness, and health equity should be the top priorities of AI developers, researchers, implementers, and regulators. The ACP also recommends implementing a continuous improvement process that includes feedback based on end-user testing in real-world clinical settings, using diverse patient demographics and peer-reviewed studies.
“
Ensuring the safe, effective, and ethical use of these technologies requires best practices as well as research, regulatory guidance, and oversight.
The ACP notes that AI and other emerging technologies should be designed to reduce health and healthcare disparities, and advocates for Congress, the U.S. Department of Health and Human Services, and other entities to support research and analysis of AI data to identify disparities and discriminatory impacts.
The ACP states that AI developers should be held accountable for the performance of their models and develop a federal AI strategy in coordination with government and non-government regulatory agencies. Collaborative research and development efforts by these agencies should focus on “ways to mitigate bias in existing or future algorithmic technologies.” Existing and future AI-related policies and guidance should be enforced and reporting of AI-related adverse events should be enabled.
ACP also said AI tools should be leveraged to reduce physician workload: For example, using AI to perform patient admissions, scheduling, and prior authorization functions can reduce the cognitive burden on healthcare providers.
The ACP recommends providing training to individuals at all levels of medical education to help physicians practice effectively in an AI-enabled health system. “If physicians do not understand or trust the output of an AI system, they will be significantly less likely to use the AI tool,” the authors state. “Therefore, creating and distributing clear, comprehensive educational materials for clinicians and end-users of AI is critical to increasing and improving the use and usefulness of AI.”
Finally, according to the ACP, the environmental impacts of AI need to be explored and mitigated throughout the AI process, as the lack of standardized measures limits our ability to address AI's potential negative impacts on the climate.
“These advanced AI [machine learning] “These tools are not yet fully understood by physicians, and the full extent of their risks to patients has yet to be determined,” the ACP authors wrote. “Best practices, as well as research, regulatory guidance, and oversight, are needed to ensure the safe, effective, and ethical use of these technologies.”
This article originally appeared on Dermatology Advisor.
