Search the KHIT Blog

Monday, April 16, 2018

"There is no precision medicine without AI"

A reasonable assertion, I guess. But, need we still be mindful of "AI vs IA" (Artificial Intelligence vs. Intelligence Augmentation)? Or, is the latter essentially morphing increasingly into the former?

Comes a recent paper by the young wizard author of this book I've cited before:

The role of artificial intelligence in precision medicine
Bertalan Mesko
The Medical Futurist Institute; Department of Behavioral Sciences, Semmelweis University, Budapest, Hungary

Accepted 13 September 2017 (pdf link in title)

1. Introduction
The essence of practicing medicine has been obtaining as much data about the patient’s health or disease as possible and making decisions based on that. Physicians have had to rely on their experience, judgement, and problem-solving skills while using rudimentary tools and limited resources.

With the cultural transformation called digital health, disruptive technologies have started to make advanced methods available not only to medical professionals but also to their patients. These technologies such as genomics, biotechnology, wearable sensors, or artificial intelligence (AI) are gradually leading to three major directions. They have been (1) making patients the point-of-care; (2) created a vast amount of data that require advanced analytics; and (3) made the foundation of precision medicine. 

Instead of developing treatments for populations and making the same medical decisions based on a few similar physical characteristics among patients, medicine has shifted toward prevention, personalization, and precision. 

In this shift and cultural transformation, AI is the key technology that can bring this opportunity to everyday practice...
2. The dawn of practicing medicine
In previous centuries, healthcare has focused on working out generalized solutions that can treat the largest number of patients with similar symptoms. If cough syrup was good for the majority of the coughing masses and only a few people had a rash as an allergic reaction to it, there was no question about treating sore throat with cough syrup. Obtaining experience and empirical evidence on a generalized basis was the working method of the medical community since Hippocrates until around the beginning of the twentieth century. 

With the refinement of diagnostic tools, the detection of viruses or bacteria, the development of new pharmaceuticals and medical methods, healthcare has been going through sweep- ing changes since the start of the last century. The experience- based and somewhat ‘trial-and-error’ approach of medicine made place for evidence-based medicine. As a consequence, physicians not only prescribed treatments because their ancestors also used the same methods, but they proved the efficacy of treatments and diagnostic methods in scientific papers and clinical studies…

3. There is no precision medicine without AI
As the National Institutes of Health described it, precision medicine is ‘an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment and lifestyle for each person.’[3] This approach allows doctors and researchers to predict more accurately which treatment and prevention strategies for a particular disease will work in which groups of people. 

It requires significant computing power (supercomputers); algorithms that can learn by themselves at an unprecedented rate (deep learning); and generally, an approach that uses the cognitive capabilities of physicians on a new scale (AI)…

4. AI in clinical practice
A major application of AI in healthcare is collecting, storing, normalizing, and tracing data. The AI research branch of the search giant, Google, launched its DeepMind Health project, which is used to mine the data of medical records in order to provide better and faster health services. In 2016, they launched a cooperative project with the Moorfields Eye Hospital NHS Foundation Trust to improve eye treatment [9]. To investigate how technology could help to analyze eye scans, Moorfields shared a set of one million anonymized eye scans with DeepMind and some related anonymous information about eye condition and disease management…

5. Is precision medicine the end of the human touch?
With advantages will also come ethical considerations and legal issues. Who is to blame if an AI system makes a false decision or prediction? Who will build in safety features? How will the economy respond to the appearance of AI when it starts making certain jobs useless? With driverless cars, there is a global debate about what decisions the algorithms would make in tricky situations. When it comes to health, this becomes a vastly bigger ethical challenge. There are more unanswered questions today than we can deal with and hopefully, with public discussions worldwide, this will clear up as AI is becoming a reality.
AI also has serious limitations in healthcare. Forecasting and prediction are mediated based on precedence in the case of machine learning, but algorithms can be underperforming in novel cases of drug side effects or treatment resistance where there is no prior example to build on. Hence, AI may not replace tacit knowledge that cannot be codified easily…

…Through the cultural transformation called digital health, the hierarchy of traditional medicine is transforming into an equal-level partnership between patients and caregivers. Besides many disruptive technologies, AI has the biggest potential to support this transition by analyzing the vast amounts of data patients and healthcare institutions record in every moment. By taking away the repetitive parts of a physician’s job, it might lead to being able to spend more precious time with their patients, improving the human touch. However, AI can only fulfill its mission if it remains a safe, efficient, and proven aid in treating patients and improving healthcare.
"By taking away the repetitive parts of a physician’s job, it might lead to being able to spend more precious time with their patients, improving the human touch. However, AI can only fulfill its mission if it remains a safe, efficient, and proven aid in treating patients and improving healthcare."

Well, yeah, you can't argue with that closing sentiment. But, for one thing, revisit my prior post "Artificial Intelligence and Ethics." See also my post on remediating "Clinician Burnout."

And, recall, we gotta "Fix the EHRs."

Bertalan's paper is well worth a close read. This stuff is coming, for better and/or worse.


I've run across (via my new issue of Harper's) a breathtakingly marvelous young writer of riveting eloquence and broad perceptivity, Meghan O'Gieblyn: 
There are two kinds of technology critics. On one side are the determinists, who see the history of technology as one of inexorable progress, advancing according to its own Darwinian logic—the wheel, the steam engine, the autonomous car—while humans remain its hapless passengers. It is a fatalistic vision, one even the Luddite can find bewitching. “We do not ride upon the railroad,” Thoreau said, watching the locomotive barrel through his forest retreat. “It rides upon us.” On the opposite side of the tracks lie the social constructivists. They want to know where the train came from, and also, why a train? Why not something else? Constructivists insist that the development of technology is an open process, capable of different outcomes; they are curious about the social and economic forces that shape each invention.
Nowhere is this debate more urgent than on the question of artificial intelligence. Determinists believe all roads lead to the Singularity, a glorious merger between man and machine. Constructivists aren’t so sure: it depends on who’s writing the code. In some sense, the debate about intelligent machines has become a hologram of mortal outcomes—a utopia from one perspective, an apocalypse from another. Conversations about technology are almost always conversations about history. What’s at stake is the trajectory of modernity. Is it marching upward, plunging downward, or bending back on itself? Three new books reckon with this question through the lens of emerging technologies. Taken collectively, they offer a medley of the recurring, and often conflicting, narratives about technology and progress...
I could not recommend Meghan's eclectic writing more enthusiastically.


Erratum, from THCB:

Twitter-Based Medicine: How Social Media is Changing the Public’s View of Medicine

Doctors can be two-faced. This isn’t necessarily a negative attribute. Doctors have distinct personas for our patients and our colleagues. With patients, doctors strive for a compassionate but authoritative role. However, with each other, doctors often reveal a different demeanor: thoughtful and collaborative, but also opinionated and even sometimes petty. These conflicts are often the result of our struggle with evidence-based medicine. The modern practice of evidence-based medicine is more than the scientific studies we read in journals. Medicine doesn’t just change in rational, data-driven increments. Evidence-based medicine is a dialectic, a conversation. Doctors are being continually challenged to reconcile personal experience, professional judgment, and scientific data. Conflict can naturally result.

This struggle has been ongoing since the rise of evidence-based medicine decades ago. There are factions in medicine who are skeptical of clinical trials as the answer to all of medicine’s important questions, while other factions are wary of authority and consensus-driven medicine. These battles have traditionally been confined to the doctor’s lounge, both literal and in the figurative “safe spaces” of academic journals and conferences. But now the doctor’s lounge is going public. Social media is enabling doctors to rapidly communicate with each other. The heated public arguments that often result are in turn raising new questions about the effect of public discourse on the medical profession and the patients we serve.

I think the social media platform that’s doing the most to influence public debate about medicine is Twitter. Twitter, with its character limits, bandwagons, and trolls may seem inhospitable to nuanced medical debate, but the power of Twitter to broadcast physicians’ instinctive and abbreviated thoughts is underappreciated…
Interesting post.

Then there's the biz sided of things. From the NY Times:

Is the doctor in?

In this new medical age of urgent care centers and retail clinics, that’s not a simple question. Nor does it have a simple answer, as primary care doctors become increasingly scarce.

“You call the doctor’s office to book an appointment,” said Matt Feit, a 45-year-old screenwriter in Los Angeles who visited an urgent care center eight times last year. “They’re only open Monday through Friday from these hours to those hours, and, generally, they’re not the hours I’m free or I have to take time off from my job.

“I can go just about anytime to urgent care,” he continued, “and my co-pay is exactly the same as if I went to my primary doctor.”

That’s one reason big players like CVS Health, the drugstore chain, and most recently Walmart, the giant retailer, are eyeing deals with Aetna and Humana, respectively, to use their stores to deliver medical care.

People are flocking to retail clinics and urgent care centers in strip malls or shopping centers, where simple health needs can usually be tended to by health professionals like nurse practitioners or physician assistants much more cheaply than in a doctor’s office. Some 12,000 are already scattered across the country, according to Merchant Medicine, a consulting firm.

On the other side, office visits to primary care doctors declined 18 percent from 2012 to 2016, even as visits to specialists increased, insurance data analyzed by the Health Care Cost Institute shows.

There’s little doubt that the front line of medicine — the traditional family or primary care doctor — has been under siege for years. Long hours and low pay have transformed pediatric or family practices into unattractive options for many aspiring physicians.

And the relationship between patients and doctors has radically changed. Apart from true emergency situations, patients’ expectations now reflect the larger 24/7 insta-culture of wanting everything now…
Lots of moving parts, many of them still moving at cross-purposes. Lots of "disruption."

Announcing the RWJF AI & The Healthcare Consumer Challenge!

…With the advent of advanced and robust AI platforms in marketing, law, and other sectors, we’re observing the vast opportunities for AI solutions in healthcare decision making. As medicine becomes more specialized, talent bottlenecks are developing and leading to increased professional strain on healthcare providers and consumers. To bring clarity and personalization to the healthcare industry, the Robert Wood Johnson Foundation is teaming up with Catalyst @ Health 2.0 to foster change in this space.

The RWJF AI & The Healthcare Consumer Challenge is calling all innovators to create AI enabled tools that support well-informed health decisions. By accelerating image recognition, data analysis, and pattern detection, we can start to remove harmful elements of human error from our systems. With predictive analysis, deep learning, and other AI enabled tools, innovators have the opportunity to help healthcare consumers make more informed and accurate decisions about the best health pathways to explore. With $100,000 in total challenge prizes, the most innovative solutions can bring attention to the benefits of AI in the consumer domain, as well as gain funding to continue tech development. Applicants can submit solutions such as tools to help find the ideal physician, estimate the cost of a health plan, or chatbots that track daily health decisions; it is up to YOU to do your part in solving a multi-billion dollar problem that affects all Americans…

More to come...

No comments:

Post a Comment