Artificial Intelligence in Healthcare and Life Sciences Needs to be More Human-Centric
Healthcare and technology experts are collaborating to experiment with artificial intelligence (AI) and machine learning. Machines and the algorithms can churn colossal amounts of data, much faster and more accurately than medical professionals, to discover patterns to improve disease diagnosis and public health.
Having worked as a technologist in Healthcare and Life Sciences domain for almost eight years, I often wonder what AI, the two vowels representing so much intriguing potential when used in conjunction, can do in this field which seeks more intelligent solutions, but is groped with fear and anxiety at the same time. Also, I personally prefer using the term Machine Intelligence or Adaptive Intelligence, as opposed to Artificial Intelligence for reason so obvious that the word speaks for itself.
So what is Artificial intelligence (AI) or machine intelligence? It is the intelligence displayed by machines, especially computers systems (robots, mobile phones, etc). The processes done by AI systems include learning (the collection of information and rules for using it), reasoning, and self-correction. Particular applications of AI include expert systems, speech recognition, and machine vision, just to name a few.
AI was coined by John McCarthy, an American computer scientist, in 1956 at The Dartmouth Conference. Today, it encompasses everything from robotic process automation to actual robotics. It has gained prominence, due to the increase in speed, size, and variety of data that businesses are now collecting. AI can perform tasks such as identifying patterns in the data more efficiently than humans, empowering enterprises to gain more insight and extract maximum benefit from their data.
Though artificial intelligence (AI) has been around since 1956, it has made valuable few contributions to healthcare practice. Only recently has the hype of machine-based learning begun to influence the reality.
Conundrum surrounding AI – its applications in healthcare and even its definition – remains widespread in popular media. Today, AI is an abbreviation for any task a computer can perform just as well as, if not better than, humans. But there are different forms of machine intelligence to consider when thinking about its role in healthcare and life science.
Several software applications and computer-generated solutions are flooding the healthcare domain, which do not confide in independent computer intelligence. Rather, they rely on human-created algorithms for analyzing data and recommending treatments.
By contrast, machine learning relies on neural networks - a computer system modeled on the human brain. Such applications involve multilevel probabilistic analysis, empowering computer systems to simulate and even expand on the way the human mind processes data. Consequently, even the programmers cannot be sure of the methodology their computer programs will use to derive solutions.
With this comes another AI variant, known as deep learning, wherein software learns to recognize patterns in distinct layers. In healthcare, specifically in genomics and proteomics, this mechanism is becoming increasingly useful. Because each neural-network layer operates both independently and in collaboration – segregating aspects such as color, size, and shape before integrating the results – these nascent visual tools have the latency to transform diagnostic medicine and can even search for cancer at the individual cell level.
AI can be dissected in innumerable ways, but the best way to understand its potential use in healthcare is to break down its applications into three separate categories: solutions based on algorithms, visual pattern recognition tools, and healthcare practice.
Solutions Based on Algorithms
Today in healthcare practice, the most commonly used AI applications are algorithmic - evidence-based approaches programmed by researchers, practitioners, and clinicians. So if you embed known data into algorithms, computers can extract information and apply it to a problem to help you with possible solutions. Let’s consider cancer treatment as a use case. Using consensus algorithms from expert healthcare practitioners, along with the data that oncologists enter into a medical record, say, a patient’s age, genetics, cancer stage and associated medical problems, an application can review hundreds, sometimes thousands, even millions of established treatment alternatives and recommend the most appropriate combination of chemotherapy course or medicines for a patient.
I mustn’t forget to mention Dr. Gabriel Escobar and his team when talking about solutions based on algorithms in The Permanente Medical Group’s division of research in the field, I’m so passionate about. The team’s research focused on one of the most important populations in any hospital, patients in a medical or surgical unit who will experience a deterioration in clinical status and be transferred to the ICU (Intensive Care Unit). Though these patients get intensive care for any serious complication, and seemingly return to their prior health status, they are three to four times more likely to die than if a doctor or practitioner had intervened and averted the deterioration itself. Dr. Escobar, along with division chief Dr. Tracy Lieu and associate executive director Dr. Philip Madvig, compiled data from 650,000 hospitalized patients, 20,000 of whom required this type of shift to ICU.
The research team then built a predictive analytical model to identify which hospitalized patients today are most likely to end up in the ICU tomorrow, following which they embedded the algorithm into an application, which continuously monitors the health status of all hospitalized patients. Finally, they designed a mechanism to notify treating doctors and practitioners whenever a patient is identified to be at risk. With this information, the doctors can intervene in advance of a major complication and save thousands more lives each year. Doesn’t that call for a loud applause? Wow!
Visual Pattern Recognition Tools
To be sincerely appreciative of the potential of visual pattern recognition in healthcare practice, you must understand how often the human eye fails even the best healthcare experts or clinicians.
A couple of independent studies found that 50% - 60% of women in the United States, who get regular mammograms over 10 years will receive at least one false-positive - a test result that wrongly indicates the possibility of cancer, thereby needing additional testing and, sometimes, unwanted procedures. As much as one-third of the time, two or more radiologists looking at the same mammography will disagree on their interpretation of the outcomes and following treatments or procedures. Visual pattern recognition software, which can store and compare millions of images while using the same heuristic techniques as humans, is estimated to be 5% - 10% more accurate than the average physician.
The gap in accuracy between the human and digital eye is expected to broaden further, and soon. As machines become more powerful and deep-learning approaches gain traction, they will continue to advance such diagnostic fields as radiology, i.e., CT, MRI and mammography interpretation, pathology, i.e., microscopic and cytological diagnoses, dermatology, i.e., rash identification and pigmented lesion evaluation for potential melanoma, and ophthalmology, i.e., retinal vessel examination to predict the risk for diabetic retinopathy and cardiovascular disease, this list just goes on.
If you have watched the popular TV show House, you might have noticed that one doctor’s genius challenges the expertise of his colleagues, implying that if all physicians were as smart as Dr. Gregory House, diagnostic enigmas would all but vanish along with unexpected deaths in hospitals.
In reality, the biggest difference between physicians is not their level of intelligence, but (1) how they approach patient health issues or diseases and (2) the healthcare systems that support them. And because (1) and (2) combine to create wide variations in pre-clinical as well as clinical trials worldwide, machine learning offers great hope and extended promise for the future.
Two Artificially Intelligent approaches are currently available, that could radically improve a doctor’s performance. The first is natural-language processing, a branch of AI that helps computers comprehend and interpret human speech and writing. This software can review thousands of comprehensive electronic medical records and elucidate the best steps for evaluating and managing patients with multiple illnesses. The second approach involves using machines and applications to watch (and learn from) practitioners at work.
In San Francisco, Adrian Aoun is using his background in artificial intelligence (AI) to explore how machines can learn from professionally skilled practitioners in real-time. Rather than extracting and analyzing data retrospectively, after physicians populate their medical records, Aoun’s primary care startup Forward, is using AI to follow what doctors do, step-by-step. With touch-screen data entry and voice recognition, Forward’s applications and machines record as well as analyze how the best physicians achieve superior results. The outcomes benefit their colleagues and patients.If all physicians matched the performance of the top 20% globally, patient deaths from cancer, infection and cardiovascular disease would decrease by the hundreds of thousands annually.
Unfortunately, the biggest hurdle to artificial intelligence in healthcare and life sciences isn’t mathematics. Rather, it’s a healthcare culture or practice that values doctor intuition over evidence-based solutions. Physicians latch on hard and fast to their independence and despise being told what is the demand of the hour. Getting them comfortable with the idea of a machine watching over their shoulder as they practice, may prove very tough in days to come.
Understanding the Deep Impact of AI
Startups and tech firms have leaped all aboard the AI pomp-n-show, promising a host of sophisticated new solutions from nurse-bots to AI Insurance (insurance powered by AI) to AI wearables, to name a few. In general, they are algorithmic and not true machine-learning approaches. Nearly all have failed to move the needle on quality outcomes or life expectancy.
For every entrepreneur promising AI as the next big thing in healthcare, there are many who fear machines will replace or even turn on humans. I sincerely believe these fears are grounded more in science-fiction than reality. It’s true that machine intelligence is advancing faster than human intelligence. But this development offers far more opportunities than risks. If we see computer speeds double another five times over the next 10 years, machine-learning tools and inexpensive diagnostic software apps could soon become as essential to physicians as the stethoscope was in the past.
At the same time, we need to accept a harsh reality. If technology is going to improve the quality of healthcare practice and lower costs, some healthcare jobs will disappear. According to one study, Artificial Intelligence is set to take over 47% of the US employment market within 20 years. Unfortunately, that’s the nature of progress. What improves lives and lowers prices for many will negatively affect those who benefited from the old model of success. Without a second thought, the role of the physician will change in the future. Fortunately for doctors, however, machines and adaptive intelligence have yet to demonstrate the kind of empathy and compassion that millions of patients depend on in their medical care.
The Promising Future of AI In Medicine
Business enterprises are expected to continue to invest in AI applications. Indeed, machine learning has the potential to take pharmaceutical as well as healthcare industry far beyond what it’s capable of today. Evidence of this fact can be found in an ancient Chinese game invented more than 2,500 years ago. Go, a two-player board game in which opponents try to claim the most territory, is incredibly complex and abstract, with a seemingly infinite set of possible moves. Its degree of difficulty left few observers believing that a machine could ever best a competent human. That myth was shattered in 2015 when AlphaGo, a program created by the Google Deepmind division, bested Lee Se-dol, one of the world’s top players.
Unlike IBM’s Deep Blue, which defeated chess champion Garry Kasparov in 1997, AlphaGo did not learn by studying humans and replaying prior matches. According to an article in Nature, humans may have taught AlphaGo the rules, but the program mastered the game by playing against itself, so amazingly powerful.
This kind of deep learning could be the very thing that creates propelling force for American healthcare into the future, helping to clarify the best care approaches, creating new approaches for diagnosing and treating hundreds of health issues, and measuring practitioner's compliance without the faulty biases of the human mind.
These types of advances will come sooner to healthcare organizations that are integrated, capitated and technology empowered. These organizations are expected to embrace algorithmic solutions on smartphones or tablets first, followed by pattern recognition software and, finally, machine-generated best practices for individual patients.
As mankind makes intelligent leaps, patients will be able to use a variety of AI apps to care for themselves, just as they manage so many other aspects of their lives today. Artificial intelligence and machine learning in healthcare will continue to get better and impact disease prevention and diagnosis, extract more meaning from data across various preclinical as well as clinical trials, help develop customized drugs based on an individual’s unique DNA and indicate treatment options among other things. However, no machine can have the human touch and empathy that still serve as the best healer. Doctors, nurses, physiotherapists, technicians, clinicians jobs are to stay no matter how much penetration machine intelligence makes in healthcare industry. Won’t you trust your physician more than any damn machine algorithm or inference? Even the world-renowned cardiologist Dr. Devi Shetty practices the power of human touch in treating patients.