AI in primary care: Opportunities and challenges for preventive healthcare

Background: Primary care systems face growing pressure from rising patient demand, increasing multimorbidity, expanding volumes of clinical data, and persistent workforce constraints. Artificial intelligence (AI)—including machine learning (ML) and deep learning (DL)—is increasingly deployed to support diagnosis, risk stratification, clinical decision support, and operational workflows in primary care.¹–⁵ However, concerns remain regarding external validity, bias, clinical integration, safety, and trust.

Objective: This systematic review synthesises evidence on how AI improves healthcare delivery in primary care settings compared with conventional clinician-only approaches, focusing on diagnostic performance, efficiency, workload, and patient-relevant outcomes.

Methods: A systematic literature search was conducted across PubMed, ResearchGate, Cochrane Library, and Google Scholar for publications from 2015 onwards using predefined keywords and Medical Subject Headings related to “artificial intelligence,” “machine learning,” “diagnosis,” “patient outcomes,” and “primary care.” Eligibility criteria were structured using a PICO framework and screening was conducted in accordance with PRISMA 2020 principles.⁶ Data were extracted into Microsoft Excel using a structured framework, and findings were synthesised narratively due to heterogeneity in study designs and outcomes. Study quality was appraised using the Critical Appraisal Skills Programme (CASP) checklists.

Results: Evidence indicates that AI can enhance primary care delivery by improving diagnostic and prognostic accuracy, accelerating clinical decision-making, enabling personalised risk assessment and treatment support, and reducing administrative burden through documentation and inbox-management assistance.³–⁵,¹³–¹⁵ However, substantial limitations persist, including inadequate external validation of many models, high risk of bias in model development and evaluation, and implementation challenges related to explainability, data governance, and clinician and patient trust.

Conclusion: AI has meaningful potential to strengthen primary care by improving speed, precision, and operational sustainability, supporting early intervention and more person-centred care.³–⁵ Real-world impact depends on rigorous validation across diverse populations, transparent governance, careful workflow integration, and sustained clinician oversight to ensure safety, equity, and trust.