“Diseases desperate grown, By desperate appliance are relieved, Or not at all”
William Shakespeare – Hamlet
The ability that’s appreciated the most in a doctor is the art of diagnosis. With the advent of fMRI, we are getting insights into the human decision-making process. In fact, we have narrowed down the process of diagnosis to a set algorithm for any and every particular disease of known etiology. Now combining computer technology and medical science, we are in the process of developing advanced algorithms that can possibly replace most human doctors when concerning its efficacy and affordability.
Time and time again, humanity has spent billions of dollars and thousands of hours to keep humankind at a good state of health. We started with religious and supernatural approaches to health classifying “demons” as the cause of diseases which later gave rise to several theories and philosophies of medicine. For instance, the philosophy of ‘Yin’ and ‘Yang’ forms the basis of traditional Chinese medicine, the philosophy of ‘Vata’, ‘Pitta’ and ‘Kapha’ forms the basis of Indian Ayurveda and similarly, the philosophies of Greek medicine gave rise to ‘Hygeia and Panacea’ (Goddess of Health and Goddess of Medicine respectively).
With the explosion of knowledge of biology and chemicals in the 20th Century, a revolution was brought about in how doctors or clinicians approach medicine. By the early 2000s, a medical doctor had radiographs, ultrasounds, advanced CT & MRI, and whatnot. With time slipping by, the healthcare sector was merging with AI, and machines were becoming more and more efficient in understanding humans. In an era, where digitization is becoming extremely omnipresent, why should healthcare be left behind!
Our cognitive skills – a foreplay of neurotransmitters
From the day we evolved from apes to Homo sapiens, we have had two main abilities: Physical & Cognitive. However, with the advent of the Industrial Revolution, machines have overpowered humans in physical work (e.g. weight lifting, cutting, pulling, etc.).
We always used to think that cognitive work requires critical thinking, a calculative and intuitive mind and intellectual skills. However with the advances in biotechnology, genetics, and fMRI, we realized that all our critical thinking and analytical skills are nothing but a fine game of axons, neurotransmitters and synapses.
What separates an average doctor from a good doctor are skills of ‘pattern recognition and analytics’. Once we have understood the depths of human cognitive skills, we will be at a stage of making machines that will have advanced cognitive functions, for instance, a digital doctor.
We know that advanced jobs like being a clinician require a high level of skill sets – right from analytical skills and problem-solving abilities to critical and parallel thinking (Differential Diagnosis is all about parallel thinking!). And we thought that as humans, we are ‘special’ in our ability to acquire all these skills, but it turns out that we are not. IBM’s Chess Program – Deep Blue, beat the world chess champion, Garry Kasparov, in 1997, IBM’s Watson computer program has beaten every human on the earth on the USA TV Quiz show ‘Jeopardy’  and Google’s Alpha ZERO program could perform 70 million calculations in a second. If all of these computer algorithms have beaten the best that humans have to offer, the day isn’t far from where we are going to have advanced machines replacing human doctors.
The Drama of Decision Making (and how the art of diagnosis is just an algorithm)!
Here I want to make a point about how every decision, right from diagnosis to treatment of a disease is nothing but a set algorithm. Let’s take the example of Tuberculosis:
As we can see here in this example, we have a set algorithm in approaching a disease.
Let us understand how this art of diagnosis can be incorporated into a machine.
Apple Inc. developed face-scanning technology that uses advanced hardware and software systems projecting more than 30,000 infrared dots onto the user’s face and then reading the pattern. Apple claims that the probability that a random person in the population could look at your apple device and unlock it using Face ID is one in a million. And if we further develop this technology to detect skin diseases by scanning the full body and then prescribe medications or to provide laser treatment directly (which is technically easily possible and is currently being researched), then it can act as a full-time dermatologist. Any normal student would need several years of hard work in a medical school to achieve the same.
Once such an algorithm is developed, it’s possible to replicate it quite easily and prepare multiple expert robotic dermatologists directly without much hassle. All the data available in the world about rare skin disease cases in history and several exceptions could be incorporated into this program and hence improving its algorithm of diagnosing, making it more accurate with handling rare cases. A good doctor would still take years and years of study to learn such rare cases.
Let us go back to the IBM Watson computer system again. Scientists working on IBM Watson stated that within a matter of a few hours, Watson learned more than 605,000 pieces of medical evidence, more than two million pages from medical journals and the further ability to search through up to 1.5 million patient records for further information giving it a breadth of knowledge no human doctor can match.
As per the report by Sloan-Kettering, human doctors use only 20 % of their knowledge while diagnosing and treating patients and that too relies on trial-based evidence. It would take at least 160 hours of reading a week just to keep up with new medical knowledge as it’s published, let alone consider its relevance or applying it practically.
Watson’s ability to absorb this information faster than any human should, in theory, fix a flaw in the current healthcare model. IBM conducted experiments for checking Watson’s ability to diagnose cancer at Memorial Sloan Kettering Cancer Centre in New York and has claimed that, in tests, Watson’s successful diagnosis rate for lung cancer is 90 percent, compared to 50 percent for human oncologists.
The making of a doctor is a costly affair
Dr. Harsh Vardhan from AIIMS Delhi stated that on average, government medical institutions like AIIMS in Delhi spends around £180,000 to £200,000 on an undergraduate to make a doctor. And this one doctor may serve on average 40,000 – 60,000 patients over 30 years of practice. Even after spending hundreds of thousands of pounds on a medical student, it doesn’t guarantee him/her being a good quality clinician and if a student decides to change his career after graduation, all the expenditures on medical education by the government would go in vain.
Now, we might think that developing a digital doctor is going to be no cheap affair and will cost billions of pounds. But this isn’t the case as technology works in a different way.
Mapping the first human genome required 15 years and over £2.5 billion. In contrast, today we can map a person’s DNA within a week at a cost of just a few hundred pounds. Hence once the technology for a digital doctor is developed, it can be set up all across the world and the cost is going to reduce without a doubt.
Why humans are not so great!
- Ever since the time of Hippocrates, physicians have had a fundamental problem with internal injuries and disease — they can’t directly see inside the human body.
- Also, humans cannot know everything in medical science. And no matter how much a human learns, they are eventually going to die; thereby can only a small percentage of the patients of the world.
- Misdiagnosis is a huge problem in the case of human doctors, mainly because of incomplete knowledge. A study estimated that as many as 1 in 20 U.S. adults are misdiagnosed by their human doctors each year, so it’s an area ripe for improvement and competition.
- Human doctors, especially consultants have limited working hours, with an average of 6-7 hours per day. Whereas the digital doctor can be made available 24/7, hence improving the overall needs of a patient.
- A particular human doctor is available only at a given geographical area and that limits their reach, treatability and accessibility.
Perks of a digital doctor!
- It’s based on all available medical knowledge: Human doctors can’t possibly hold this much information in their heads, or cope up as it changes over time. A digital doctor knows it all and never overlooks or forgets anything.
- Accuracy: With all the available knowledge, skill set, critical and analytical thinking, the AI doctor will be an excellent diagnostician indeed.
- Consistency: Given the same inputs, a digital doctor will always output the same diagnosis. Inconsistency is a surprisingly large and common flaw among human medical professionals, even experienced ones. And our digital doctor is always available and never annoyed, sick, nervous, hungover, upset, in the middle of a divorce, sleep-deprived, and so on.
- Low-cost treatment: It’ll be very expensive to build and train a digital doctor, but once it’s up and running the cost of doing one more diagnosis with it is essentially zero, unless it orders tests.
- Accessible from anywhere in the world: If a person has access to a computer or mobile phone, our digital doctor is on call for them.
- It can be easily updated: If WHO comes up with a new treatment or a drug regime for a disease; it usually takes several months to years in order to reach a physician in some developing countries. And when we think of new surgical procedures discovered in a hospital in one country, it could take several decades to reach a surgeon in another corner of the globe. On the contrary, even if you have 10 billion AI doctors in the world – each monitoring the health of a single human being – one can still update all of them within a split second. The potential advantages of connectivity and upgradeability of a digital doctor makes great sense in replacing most ground-level human doctors – if not specialists.
What future might look like:
One might argue that by switching from human doctors to digital ones, we will lose the advantages of individuality. For example, if a human doctor makes a wrong decision; a few individuals may die or suffer. But if a digital AI system makes a wrong judgment, it would become a potential disaster. In reality, an algorithm can maximize the benefits of connectivity without losing the advantages of individuality. It’s quite possible to run many alternative algorithms on the same network. Hence we might have a day when a patient sitting in some remote part of the world can select if he wants to consult one AI doctor or some other AI doctor (in the current context, maybe a Google doctor or a Microsoft doctor).
The future looks fascinating and surprising beyond a doubt.
“The woods are lovely, dark and deep, But I have promises to keep, And miles to go before I sleep”
This idea of continuing human endeavors no matter what, still holds true in our time. We are not sure what will result from all of the technological and biological advances that we are making – artificial heaven or an artificial catastrophe.
For example, if we develop gene-specific medications that are capable of altering defective genes and improving the physical and mental abilities of a person, it might so happen that only a certain class of people could afford it and that would create a bifurcation of classes of people at a biological level. And that would be no less than a disaster.
In the current time, we are clouded by the risk of breach of our privacy by the tech companies. In the context of AI doctors who monitor our health all the time, for instance, when such AI doctors will store our medical information, there will always be the fear of health-care companies misusing the information of our health and breaching our privacy.
Now imagine that you are just diagnosed with HIV. You would much prefer a human being who understands human emotions and feelings to give you this news rather than an AI doctor giving you a diagnosis. Hence, digital doctors will need to understand human emotions and empathy before dealing with patients. Such problems or questions will also arise with digital doctors and we aren’t sure if we will ever have all the answers.
However, all we can do right now is to keep growing and exploring as fast as we can and as much as we can; with the ultimate idea of taking humanity a step forward.
It might sound very simple in this article, but it’s a lot easier said than done. We will have to face innumerable questions on social, political, emotional and cultural grounds. And we aren’t sure if we are ready to answer them.
- K. Park – Preventive and Social Medicine (Man and Medicine: Towards health for all – 2017)
- Daniel Kahneman – Thinking, Fast and Slow (New York, Farrar; Stratus & Giroux 2011)
- “complete genomics” from http://www.completegenomics.com/
About The Author
sneh sonaiya is a future doctor, geek, a writer and with all the romance playing violin. He commenced his astrobiology course a month back all hand in hand with medicine. Can’t deny, he can drive you high with words..