In November 2019, I presented to medical students at the Drexel University College of Medicine on the topic, “Will There Be a Role for the Physician in the New World of Augmented Intelligence?”
The question is increasingly pertinent. Augmented intelligence, also called artificial intelligence or AI, continues to advance. The volume of medical knowledge is growing faster than we can assimilate it. Meanwhile graduating medical students who invest in their education and training – more than ten thousand hours over at least seven years with greater than $200,000 in average total student loan debt – deserve reassurance that they will have a meaningful future.
About a decade ago, medical educator and author Peter Densen, MD observed that medical knowledge doubled in a few years. He predicted that for medical students who were graduated in 2020, the volume of medical knowledge would double within 73 days. How can we keep up?
What if we utilize technology to help us to benefit our patients? Nearly 70 years ago, the computer scientist Alan Turing asked, “Can machines think?” He anticipated human-like intelligent machines that we call AI.
The dramatic growth of healthcare AI is driven, in part, by faster and bigger computers, the rise of Big Data, and use of the “cloud.” AI has the potential to streamline the opportunities for us to efficiently absorb information.
Daily Dose of AI
AI is based on algorithms, or decision trees, where computers complete tasks in a manner typically associated with a rational human being. AI is something we already often use in our daily lives.
Have you asked Siri or Alexa a question? That involves Natural Language Processing (NLP) where your voice is changed into a form understood by a computer that generates a response. NLP can help physicians as it automatically reads, prioritizes, and digests journal articles into readily accessible personalized daily doses of the most relevant findings.
Have you searched for something on Google? Google increasingly uses Deep Learning where layers of algorithms have been designed in networks to simulate pathways in the human brain.
AI: Artificial or Augmented Intelligence?
“Augmented” Intelligence has become the preferred term among key technology companies, other innovators, and physician AI experts. The American Medical Association (AMA) House of Delegates adopted policy on AI using the term “Augmented Intelligence” to emphasize that AI’s role should be to enhance – not replace – human intelligence. This policy guides the AMA’s AI-related advocacy to promote that practicing physicians should shape healthcare AI to safely and effectively improve patient care. AMA resources on healthcare AI are found at an AMA microsite: www.ama-assn.org/amaone/augmented-intelligence-ai
Clinical Cases and Concerns
AI can help inform prognosis, therapy, screening, or diagnosis. Data can be derived from online sites, electronic health records, wearables, apps, and other tools. Machines can scan vast quantities of data and pinpoint subtle signs and symptoms that physicians might otherwise miss. If AI is embedded within the workflow, it could enhance patient care in real-time.
AI has the potential to decrease morbidity and mortality. Here are vignettes of patients given fictitious names.
Physician involvement early in the development process is essential to help shape the algorithms, identify biases, ensure equity, and build confidence. AI can be transformative only if physicians are instrumental to its design, validation, and implementation.
Role of the Physician with AI
Some question whether AI has the potential to replace the physician’s role.
Informatics expert John Mattison, MD wrote about the need to incorporate “highly variable and contextual human considerations into the treatment plan [that] really requires thoughtful and empathic discussion between doctor and patient.”
What human elements does the physician bring to the physician-patient relationship?
The education and training of physicians is especially valuable in the era of AI, where clinical acumen is paramount to determine when an AI-guided recommendation is valid for the patient you are seeing.
When we ask, “Will there be a role for the physician in the new world of Augmented Intelligence?” be assured that the answer is yes. Patients never care how much you know until they know how much you care.
Augmented intelligence, if applied correctly, with physician leadership, gives us an opportunity to maximize time to do what we love most: meaningful and rewarding care of patients.
Marilyn Heine, MD, MJH, MD, FACEP, FACP, FCPP is Clinical Assistant Professor of Medicine at Drexel University College of Medicine. She is in active clinical practice in southeast Pennsylvania.
The question is increasingly pertinent. Augmented intelligence, also called artificial intelligence or AI, continues to advance. The volume of medical knowledge is growing faster than we can assimilate it. Meanwhile graduating medical students who invest in their education and training – more than ten thousand hours over at least seven years with greater than $200,000 in average total student loan debt – deserve reassurance that they will have a meaningful future.
About a decade ago, medical educator and author Peter Densen, MD observed that medical knowledge doubled in a few years. He predicted that for medical students who were graduated in 2020, the volume of medical knowledge would double within 73 days. How can we keep up?
What if we utilize technology to help us to benefit our patients? Nearly 70 years ago, the computer scientist Alan Turing asked, “Can machines think?” He anticipated human-like intelligent machines that we call AI.
The dramatic growth of healthcare AI is driven, in part, by faster and bigger computers, the rise of Big Data, and use of the “cloud.” AI has the potential to streamline the opportunities for us to efficiently absorb information.
Daily Dose of AI
AI is based on algorithms, or decision trees, where computers complete tasks in a manner typically associated with a rational human being. AI is something we already often use in our daily lives.
Have you asked Siri or Alexa a question? That involves Natural Language Processing (NLP) where your voice is changed into a form understood by a computer that generates a response. NLP can help physicians as it automatically reads, prioritizes, and digests journal articles into readily accessible personalized daily doses of the most relevant findings.
Have you searched for something on Google? Google increasingly uses Deep Learning where layers of algorithms have been designed in networks to simulate pathways in the human brain.
AI: Artificial or Augmented Intelligence?
“Augmented” Intelligence has become the preferred term among key technology companies, other innovators, and physician AI experts. The American Medical Association (AMA) House of Delegates adopted policy on AI using the term “Augmented Intelligence” to emphasize that AI’s role should be to enhance – not replace – human intelligence. This policy guides the AMA’s AI-related advocacy to promote that practicing physicians should shape healthcare AI to safely and effectively improve patient care. AMA resources on healthcare AI are found at an AMA microsite: www.ama-assn.org/amaone/augmented-intelligence-ai
Clinical Cases and Concerns
AI can help inform prognosis, therapy, screening, or diagnosis. Data can be derived from online sites, electronic health records, wearables, apps, and other tools. Machines can scan vast quantities of data and pinpoint subtle signs and symptoms that physicians might otherwise miss. If AI is embedded within the workflow, it could enhance patient care in real-time.
AI has the potential to decrease morbidity and mortality. Here are vignettes of patients given fictitious names.
- Darlene has diabetes and is at risk for blindness. She needs to be screened for diabetic retinopathy but is unable to readily access an ophthalmologist. What if Darlene’s primary care physician can screen her for diabetic retinopathy using AI?
- John has an irregularity on imaging that is too early to identify as cancer. What if his early cancer can be found using AI?
- George presents to the emergency department with evidence of a urinary tract infection, but appears to be stable. What if his emerging sepsis could be detected early to help improve his chance of survival by using AI?
- Gloria’s breast cancer treatment is based on pathology studies and generally done tests. What if her breast cancer treatment could be tailored more specifically using AI?
- Bias: We must ensure equity. Whose data forms the basis of the AI algorithms? Are unconscious biases baked in without the developers being aware? Can AI risk-stratify and personalize care, or will it entrench existing biases and exacerbate disparities? How can we optimize population health?
- Privacy: How is patients' protected health information (PHI) being used? Is patient consent obtained prior to sharing sensitive data? Is PHI being monetized? Are patients informed that even if PHI is de-identified, the data may later be re-identified? Our online interactions generate an abundance of potentially trackable data.
- Black box: Who is influencing the algorithms? Are the algorithms adequately tested? Do the algorithms stay constant or do they change? What is the level of transparency?
- Liability: Who is responsible for using or not using AI? Who is held accountable if AI provides misinformation?
Physician involvement early in the development process is essential to help shape the algorithms, identify biases, ensure equity, and build confidence. AI can be transformative only if physicians are instrumental to its design, validation, and implementation.
Role of the Physician with AI
Some question whether AI has the potential to replace the physician’s role.
Informatics expert John Mattison, MD wrote about the need to incorporate “highly variable and contextual human considerations into the treatment plan [that] really requires thoughtful and empathic discussion between doctor and patient.”
What human elements does the physician bring to the physician-patient relationship?
- Compassion: The ability of the physician to provide comfort and reassurance is central to the patient-physician relationship.
- Empathy: The physician’s role to assure that support is conveyed and assess that patients understand.
- Social determinants of health: An appreciation for the patient’s unique context.
- An open mind: The ability to adapt, demonstrate cultural competence, and understand the patient’s level of health literacy. A human can recognize and address a variance, but AI is based on pattern recognition and might overlook conflicting clinical information.
The education and training of physicians is especially valuable in the era of AI, where clinical acumen is paramount to determine when an AI-guided recommendation is valid for the patient you are seeing.
When we ask, “Will there be a role for the physician in the new world of Augmented Intelligence?” be assured that the answer is yes. Patients never care how much you know until they know how much you care.
Augmented intelligence, if applied correctly, with physician leadership, gives us an opportunity to maximize time to do what we love most: meaningful and rewarding care of patients.
Marilyn Heine, MD, MJH, MD, FACEP, FACP, FCPP is Clinical Assistant Professor of Medicine at Drexel University College of Medicine. She is in active clinical practice in southeast Pennsylvania.
