This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American
Is there a place for artificial intelligence (AI) in the field of medicine?
Will doctors one day be replaced by robots?
The answers are “yes” and “no,” respectively.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
AI—the capability of a machine to imitate intelligent human behavior—is poised to transform the practice of medicine as we know it. But it is a complementary technology, designed to enhance the performance of humans—including physicians, nurses and medical researchers—in doing their jobs.
Currently, health care is trailing other industries in the acceptance and application of AI. But many experts predict it will be the industry most disrupted by AI in the coming decade, thanks to the widespread adoption of electronic healthcare records (EHRs) and the huge amounts of data at our disposal.
At our new Center for Clinical Artificial Intelligence, and at other AI-centric research centers around the world, researchers are developing machine learning capabilities that, ultimately, will improve how we diagnose and treat patients. These capabilities also create cost and time efficiencies that improve the overall patient experience and help break down barriers to care.
With AI as an instrument inside our medical bags, we will have the opportunity to save and enhance more patient lives. Indeed, the possibilities are endless.
PROMISING PREDICTIVE MODELS
Currently, our team is using AI to find the hidden pearls of wisdom buried inside massive reams of data. At the same time, we are striving to create a new, hybrid role—what we call “physician data scientists”—who understand machine learning, AI and how these technologies can be applied to medical research and clinical practice. Our goal is to improve patient outcomes and drive down costs.
Results, to date, have been extremely promising. Our researchers are building machine learning models that use AI not only to predict certain patient outcomes but to lead directly to actions that improve our patients’ health. For instance, we have been able to identify (at high rates of accuracy) patients at high risk of death within 48 to 72 hours of hospital admission, which enables clinicians to take proactive steps to treat them in ways that mitigate further risk.
In another project, we have developed a personalized prediction model that surpassed existing prediction models for myelodysplastic syndromes (MDS). We can determine, with high degrees of accuracy, an MDS patient’s risk of mortality, as well as the risk of transformation to acute myeloid leukemia (AML), a more aggressive type of bone marrow cancer.
By understanding the likelihood of a patient’s prognosis, we will be able to develop a treatment plan that is more appropriate for his/her situation. That means fewer instances of over- or undertreatment, better counsel to patients and more personalized care.
MOVING FROM PUBLISHING TO PRACTICING
But AI in health care has it challenges, too, given the level of complexity and nuance in this field. Also, given a lack of regulatory and clinical standards in AI research to date, the field can produce inconsistent or flawed studies that could lead to improper or irresponsible implementation of the findings.
AI is not a panacea. That’s why, in my view, you’ll never see “machine” doctors, because the human factors of empathy, common sense and instinct so often play a critical role in medical decision-making. What we’re doing with AI, in essence, is striving to better harness data to gain critical additional insights that could lead to improved care and outcomes.
Our work is progressing, but for us to truly move this effort forward we must get more physicians engaged. And we have to train them in how to better understand these algorithmic models and what the results mean for research or patient care.
As we move forward, we at the Center for Clinical Artificial Intelligence are eager to move beyond academic research that leads to published studies; and instead, generate research outcomes that can be more broadly reviewed, assessed and adopted as common medical practices.
Our goal is to save and improve lives.
Isn’t that why we got into this profession in the first place?
