Chapter 36: Further Reading — AI in Healthcare Decision-Making

Foundational Clinical AI References

1. Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464), 447–453. The landmark study documenting racial bias in a commercial care management algorithm through use of cost as a proxy for health need. The most-cited empirical paper on algorithmic bias in healthcare and essential reading for anyone working on clinical AI equity.

2. Topol, E. J. (2019). High-performance medicine: The convergence of human and artificial intelligence. Nature Medicine, 25, 44–56. A balanced overview of clinical AI's capabilities and limitations by a leading cardiologist and researcher, covering AI in radiology, pathology, cardiology, and genomics. Provides accessible technical grounding without excessive optimism.

3. Rajpurkar, P., Chen, E., Banerjee, O., & Topol, E. J. (2022). AI in health and medicine. Nature Medicine, 28, 31–38. An updated assessment of the state of AI in clinical medicine, covering both the advances since Topol's 2019 paper and the persistent challenges of clinical validation, bias, and deployment.

Automation Bias and Human Oversight

4. Goddard, K., Roudsari, A., & Wyatt, J. C. (2012). Automation bias: A systematic review of frequency, effect mediators, and mitigators. Journal of the American Medical Informatics Association, 19(1), 121–127. The foundational systematic review of automation bias in healthcare, documenting its prevalence, conditions under which it occurs, and approaches to mitigation. Essential background for understanding why "human-in-the-loop" requires substantive conditions, not nominal presence.

5. Cresswell, K., et al. (2020). Qualitative analysis of vendor experiences with artificial intelligence implementation in healthcare. NPJ Digital Medicine, 3, 91. A qualitative study examining how AI vendors and hospital partners experience AI deployment, including insights into governance gaps, communication failures, and the disconnect between vendor and clinician perspectives.

IBM Watson for Oncology

6. Ross, C., & Swetlitz, I. (2017, September 5). IBM's Watson supercomputer recommended 'unsafe and incorrect' cancer treatments, internal documents show. STAT News. The investigative report that brought Watson's clinical limitations to public attention. Based on internal IBM documents, this article documents specific cases of inappropriate recommendations and the gap between IBM's marketing and clinical reality.

7. Strickland, E. (2019). How IBM Watson overpromised and underdelivered on AI health care. IEEE Spectrum. A post-mortem analysis of what went wrong with Watson Health, drawing on interviews with clinicians, IBM employees, and health system partners. Provides detail on the business and organizational dynamics that allowed deployment to proceed without adequate clinical evidence.

8. Maddox, T. M., et al. (2019). Questions for artificial intelligence in health care. JAMA, 321(1), 31–32. A concise editorial articulating the core questions that must be answered before clinical AI can be deployed responsibly — many of which Watson failed to answer. Useful as a governance checklist framework.

FDA Regulation of Clinical AI

9. Food and Drug Administration. (2021). Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan. The FDA's primary policy document on AI/ML medical device regulation, introducing the concept of predetermined change control plans and laying out the regulatory roadmap for adaptive AI devices. Essential reading for healthcare regulatory compliance.

10. Food and Drug Administration. (2023). Marketing Submission Recommendations for a Predetermined Change Control Plan for Artificial Intelligence-Enabled Device Software Functions. Draft Guidance. The FDA's draft guidance operationalizing PCCP requirements — what manufacturers must submit and what standards they must meet. The key regulatory document for organizations developing or procuring adaptive clinical AI.

11. Gerke, S., Minssen, T., & Cohen, G. (2020). Ethical and legal challenges of artificial intelligence-driven healthcare. In Artificial intelligence in healthcare (pp. 295–336). Academic Press. A comprehensive legal and ethical analysis of AI in healthcare, covering FDA regulation, liability, privacy, and equity. The most thorough legal analysis of the regulatory framework as it existed at the time of publication.

Algorithmic Bias and Health Equity

12. Cirillo, D., et al. (2020). Sex and gender differences and biases in artificial intelligence for biomedicine and healthcare. NPJ Digital Medicine, 3, 81. A systematic analysis of sex and gender bias in biomedical AI, documenting how underrepresentation of women in training data creates models with differential performance across sexes.

13. Adamson, A. S., & Smith, A. (2018). Machine learning and health care disparities in dermatology. JAMA Dermatology, 154(11), 1247–1248. An early and influential paper documenting that dermatology AI trained predominantly on images of lighter skin tones performed significantly worse on images of darker skin tones — a finding with direct clinical implications for dermatology diagnosis.

14. Vyas, D. A., Eisenstein, L. E. G., & Jones, D. S. (2020). Hidden in plain sight — Reconsidering the use of race correction in clinical algorithms. New England Journal of Medicine, 383(9), 874–882. The comprehensive medical journal analysis of race correction in clinical algorithms, including eGFR, spirometry, and other formulas, providing the scientific framework that supported the subsequent elimination of race corrections in multiple clinical tools.

End-of-Life and Mental Health AI

15. Meisel, Z. F., & Rubin, E. B. (2020). Death, dying, and the algorithms of resuscitation. Annals of Internal Medicine, 172(4), 292–293. An essay examining the ethical dimensions of AI mortality prediction in the context of resuscitation decision-making, raising concerns about algorithmic prognosis and patient dignity.

16. Abd-Alrazaq, A. A., et al. (2020). Perceptions and opinions of patients about mental health chatbots: Scoping review. Journal of Medical Internet Research, 22(3), e17828. A systematic review of what patients think about mental health chatbots — providing user-perspective evidence that should inform the design and governance of these tools.

17. Torous, J., et al. (2020). Digital mental health and COVID-19: Using technology today to accelerate the curve on access and quality tomorrow. JMIR Mental Health, 7(3), e18848. An analysis of the potential and limitations of digital mental health tools, including AI chatbots, with attention to evidence quality, privacy, and governance.

Governance and Institutional Frameworks

18. Shah, N. H., Milstein, A., & Bagley, S. C. (2019). Making machine learning models clinically useful. JAMA, 322(14), 1351–1352. A practical framework for clinical AI model development, validation, and deployment from the Stanford clinical informatics community. Articulates the standards that clinical AI should meet before influencing patient care.

19. The Joint Commission. (2023). Artificial Intelligence in Healthcare: Current and Future Applications. The accreditation body's framework for how hospitals should govern AI in clinical settings — important for understanding what governance standards accreditors may eventually require.

20. National Academy of Medicine. (2023). Artificial Intelligence in Health Care: The Hope, the Hype, the Promise, the Peril. A comprehensive assessment of AI in healthcare from the National Academy of Medicine, covering the full range of applications, governance challenges, and policy recommendations. The most authoritative multi-disciplinary review available.