This detailed text introduces the article titled “Navigating the doctor-patient-AI relationship – a mixed-methods study of physician attitudes toward artificial intelligence in primary care”. This research, published in BMC Primary Care in 2024, was authored by Matthew R. Allen, Sophie Webb, Ammar Mandvi, Marshall Frieden, Ming Tai-Seale, and Gene Kallenberg.
Background and Rationale: Artificial intelligence (AI) is a rapidly advancing field that is increasingly being integrated into medical practice. Primary care, a foundational component of medicine, faces significant challenges such as physician shortage and burnout, which adversely affect patient care. AI, particularly through digital health applications, is frequently proposed as a potential solution to these issues. Despite the critical role of primary care and the potential impact of AI, there has been a scarcity of research specifically focusing on primary care physician (PCP) attitudes toward AI. Furthermore, the impact of AI and digital health on the essential doctor-patient relationship remains largely underexplored. Existing research often discusses AI in general terms, lacking engagement with clinicians on tangible, specific use cases.
Study Objective: Recognizing this gap, this study aimed to delve into PCP attitudes regarding the transformative influence of AI and the broader shift towards digitalization in primary care. The primary objective was to examine PCP views on AI and its potential impact on crucial aspects pertinent to primary care, such as the doctor-patient relationship and clinical workflow. By doing so, the researchers sought to inform primary care stakeholders to encourage the successful and equitable uptake of future AI tools. A key distinguishing feature of this study is its novel approach: it is the first to explore PCP attitudes using specific primary care AI use cases, rather than discussing AI in general medical terms. The long-term goal is to use these findings to develop AI tools that improve patient care.
Methodology: Conducted from June to August 2023, the study employed a mixed-methods approach. This involved:
- A digital survey administered to 47 primary care physicians affiliated with a large academic health system in Southern California. The survey quantified general attitudes toward AI and specific AI use cases, using Likert scales to assess comfort levels and perceptions.
- Semi-structured interviews with 15 of the survey respondents. These interviews provided an open platform for PCPs to share nuanced perspectives, which were then analyzed thematically. The study specifically highlighted AI use cases relevant to primary care, including:
- AI-enhanced disease screening for conditions like obstructive sleep apnea (OSA).
- AI-facilitated disease management for chronic conditions such as hypertension.
- AI-facilitated administrative tasks, such as using Large Language Models (like ChatGPT) for drafting patient message responses within the Electronic Health Record (EHR).
Key Findings: The study revealed that PCPs generally held largely positive views of AI in medicine, with 76.6% expressing optimism. However, these attitudes were often contingent on the context of AI adoption. While some PCPs reported comfort in communicating the role of AI tools to patients, a significant percentage did not. Concerns reported by PCPs were categorized into two main areas:
- Technological concerns: These included issues such as algorithmic bias (1 participant) and accuracy and safety (7 participants). PCPs also considered the external validity and the ability of AI algorithms to appreciate individual patient nuances.
- People-and-process factors: Many concerns centered on systemic issues rather than technology itself. These included:
- Medicolegal implications of acting or failing to act on AI guidance.
- Resource availability: Concerns that AI tools would not be helpful without adequate resources for follow-up diagnosis and treatment.
- Increased workload and physician burnout: A common concern among 10 participants, who feared AI could lead to more work, a constant need to verify AI output, or an excessive focus on productivity.
- Impact on the doctor-patient relationship: While some expressed hope for AI to improve this relationship (10 participants) by alleviating clinician burden, many worried it could harm it (12 participants) by warping patient expectations or shifting focus away from physician well-being.
- Lack of focus on physician well-being when implementing new technologies.
- Reimbursement issues: A key concern was that current healthcare payment models do not support innovative care delivery methods like AI-powered digital health tools, creating a disconnect between care innovation and practice.
Conclusion and Significance: The study’s findings highlight the “dual nature” or “double-edged sword” of AI in healthcare, indicating its potential to either alleviate or exacerbate existing challenges in primary care. It underscores that for clinicians, concerns about AI often extend beyond the technology itself to its impact on their professional lives, well-being, and patient relationships. The research suggests that AI initiatives that fail to address both technological and people-and-process concerns raised by PCPs may struggle to make a meaningful impact. The authors emphasize the crucial need for primary care stakeholders to align on key issues identified by PCPs, including innovative reimbursement models, scheduled time for digital health engagement, additional team members to support digital care, and comprehensive education for both patients and physicians. Ultimately, the study advocates for a thoughtful and equitable application of AI in primary care to avoid further disillusionment among PCPs and ensure that benefits like time and cost-savings are shared by physicians, aligning with the Quadruple Aim of healthcare.
APA Reference:
Allen, M. R., Webb, S., Mandvi, A., Frieden, M., Tai-Seale, M., & Kallenberg, G. (2024). Navigating the doctor-patient-AI relationship—A mixed-methods study of physician attitudes toward artificial intelligence in primary care. BMC Primary Care, 25, Article 42. https://doi.org/10.1186/s12875-024-02282-y