Review
Artificial Intelligence in Medical Practice: The Question to the Answer?

https://doi.org/10.1016/j.amjmed.2017.10.035Get rights and content

Abstract

Computer science advances and ultra-fast computing speeds find artificial intelligence (AI) broadly benefitting modern society—forecasting weather, recognizing faces, detecting fraud, and deciphering genomics. AI's future role in medical practice remains an unanswered question. Machines (computers) learn to detect patterns not decipherable using biostatistics by processing massive datasets (big data) through layered mathematical models (algorithms). Correcting algorithm mistakes (training) adds to AI predictive model confidence. AI is being successfully applied for image analysis in radiology, pathology, and dermatology, with diagnostic speed exceeding, and accuracy paralleling, medical experts. While diagnostic confidence never reaches 100%, combining machines plus physicians reliably enhances system performance. Cognitive programs are impacting medical practice by applying natural language processing to read the rapidly expanding scientific literature and collate years of diverse electronic medical records. In this and other ways, AI may optimize the care trajectory of chronic disease patients, suggest precision therapies for complex illnesses, reduce medical errors, and improve subject enrollment into clinical trials.

Section snippets

Answerable Questions

Some questions about AI's role in modern society have been answered:

  • Why has AI emerged as useful in several diverse sectors (business, science, government)?

  • How does AI differ from standard biostatistics?

  • What is “big data”? How does AI enable big dataset analysis?

  • How do AI applications differ from smart technologies (medical devices, digital diagnostics, data management systems) already used in medical practice?

Inside AI's Black Box

While AI encompasses a wide range of symbolic and statistical approaches to learning and reasoning (Figure), recent advances in algorithms, computational power, and access to large datasets have enabled artificial neural networks to emerge as the leading AI method. Artificial neural networks are flexible mathematical models that use multiple algorithms to identify complex nonlinear relationships within large datasets (analytics). Machines learn when errors encountered in response to minor

Works in Progress

Questions remain about the applicability, practicality, and value of AI in medical practice:

  • How is AI use in medical practice distinguished from big data analytics applications for health care delivery and population health?

  • Can AI address medical practice “pain points,” providing more efficient and efficacious care while de-escalating physician burnout?

  • Will AI improve patient outcomes when used at the point of care?

  • Can Internet-of-Things health care facilities and medical homes become a

Use Cases for Cognitive Medical Practice

Simple neural networks have been used in medicine since the early 1990s to interpret electrocardiograms,6 diagnose myocardial infarction,7 and predict intensive care unit length of stay following cardiac surgery.8 AI's scientific applications have proliferated, including image analysis (radiographic, histologic), text recognition with natural language processing, drug activity design, and prediction of gene mutation expression.9, 10 Recent AI applications provide proof of concept for AI use in

Potential Jeopardies

Concerns about cognitive medical practice are largely the result of existing information deficits:

  • Will providers perceive AI as another technology barrier to direct patient care?

  • Does AI enhance or dis-intermediate patient–physician engagement?

  • What nonmedical barriers exist to the use of AI in direct patient care (eg, reimbursement, regulatory)?

  • Will AI put some physicians out of work (obsolescence) and/or reduce physician compensation (relative value)?

  • Are physicians using AI at risk for skill

Technology Insertion

Tracey Kidder's 1981 Pulitzer Prize-winning book The Soul of a New Machine22 underscored how imperfect humans remained critical to intelligent computer design. The current AI medical literature reproducibly supports a widely held tautology—that collaborative human-machine tasking improves performance over either alone. While AI's technology displacement curve is paralleled by an opportunity curve, concerns abide that AI will dislocate highly skilled health professionals from their jobs.

A tool

“Is That Your Final Answer?”

Final Jeopardy! answer: An 1816 medical instrument invented by Dr. René Laennec to avoid patient contact. Correct question: “What is the stethoscope?”

There are 2 reasons why medical schools still teach students to use a centuries-old tool. The first is that the stethoscope reveals diagnostic information helpful to patient care. The second is that the hand-held device requires learners to physically contact the precordium, a connection that is both humanistic of doctors and reassuring to

References (25)

  • TuJ.V. et al.

    Use of a neural network as a predictive instrument for length of stay in the intensive care unit following cardiac surgery

    Proc Annu Symp Comput Appl Med Care

    (1992)
  • M.K. Leung et al.

    Deep learning of the tissue-regulated splicing code

    Bioinformatics

    (2014)
  • Cited by (437)

    View all citing articles on Scopus

    Funding: None.

    Conflicts of Interest: DDM: None; EWB: Employed by IBM; the employment relationship did not create direct or indirect financial or scientific conflicts in the preparation of this paper.

    Authorship: All authors had access to the data and a role in writing this manuscript.

    View full text