AI being your psychologist… Have you thought about it?!
Some technologies integrate into our lives. Electricity is an example. We no longer know how to live without it. In digital technologies, the internet, smartphone and social networks are already part of our daily lives. Wearables are also starting to gain space. My wife and I wear bracelets that measure some simple data like heart rate, steps and sleep quality. The first thing I do when I wake up is check my sleep, referring to the app that shows how much of my sleep was deep or light. I know this is not an accurate data, as it only measures my arm movement during the night. It makes no correlation to the brain waves that would more accurately indicate the actual sleep state. But it's already an indication that allows me to correlate that night with events from the day before, which probably affected the quality of sleep. And sleep quality has a direct influence on health.
In this article, we'll look at how AI integrates with mental and physical health. Let's start with mental health. Mens sana in corpore sano. A recent e-book, “Digital Interventions in Mental Health: Current Status and Future Directions” published by Frontiers in Psychiatry clearly shows the role of digital technology and AI in targeted mental health therapies. In fact, the first experiments took place in 1965, with a primitive bot, called Eliza, which simulated a psychotherapy session, and which generated very interesting interactions. More recently, a much more evolved bot has emerged that has been instrumental in mental health. The article “The Chatbot Therapist Will See You Now” shows that its purpose is not to replace a psychiatrist, but can complement their activity. By the way, it is curious that an informal survey of 2,000 people in the US, via Twitter, showed that 44% of them felt more comfortable talking to a computer than to a doctor about some embarrassing disease or symptom. The use of chatbots and virtual assistants is increasing and they increasingly resemble us in their ability to interact. This opens up space to increase people's communication with medicine.
A bot or virtual assistant, interacting with a person, does not have the human characteristics that a doctor has such as empathy or common sense, but it can pick up signals that go unnoticed by doctors. For example, certain patterns of voice variation can indicate signs of mental disorders. I read two thought-provoking articles published by Nature, “Digital biomarkers of mood disorders and symptom change” and “Automated analysis of free speech predicts psychosis onset in high-risk youths”, which show that Deep Learning algorithms can help identify probable cases of outbreaks psychotics. I think it's an interesting field for medicine to advance. There is a lot of talk about using AI to transform radiology or pathology, but its applicability to mental health seems to be very promising.
We can go further. Today, it is estimated that more than 1.2 trillion photos are taken via smartphones a year. A fair share of them goes to social media and a study “Instagram photos reveal predictive markers of depression” analyzed nearly 50,000 photos of 166 individuals (who gave consent to the survey), 71 of whom had a history of depression. Analyzing the characteristics of the photos, such as whether the person was present or not, whether it was taken indoors or out, day or night, special effects adopted, frequency of posts and other variables, the study was able to identify, with 70% accuracy, digital depression markers. This accuracy compares favorably with that made by physicians (“Clinical diagnosis of depression in primary care: a meta-analysis”), who on average produce more than 50% false positives in the diagnosis of depression. This study was based on health systems physicians, not psychiatrists. But considering that the vast majority of the population cannot access a psychiatrist, AI can be a powerful tool in combating mental illness. Depression is the main cause of mental disorder. In Brazil, 5.8% of the population suffers from depression, a rate above the global average, which is 4.4%. This means that almost 12 million Brazilians suffer from the disease, placing the country at the top of the ranking in the number of cases of depression in Latin America, according to the World Health Organization (WHO), in a survey released last year. While in the world there has been a reduction in suicide deaths in recent years, by about 32%, Brazil goes against this trend, having recorded, between 2006 and 2015, a 24% increase in the number of suicides committed by the population of 10 to 19 years. The subject is serious and if the AI can collaborate in this sense, I believe that we must evolve in the studies of its applicability.
Specifically in terms of suicides, in Brazil, according to official estimates by Datasus, 195,979 self-inflicted deaths were recorded between 1996 and 2017. This is equivalent to more or less two stadiums in Maracanã, completely full. A study “What the numbers say about suicide in Brazil” shows the seriousness of the problem, which, by the way, is worldwide. Estimates indicate that more than 25 million suicide attempts are made annually, all over the world. The study “Risk Factors for Suicidal Thoughts and Behaviors: A Meta-Analysis of 50 Years of Research” shows that the application of AI can greatly contribute to the prevention of suicides.
Generally speaking, medicine has focused much more on physical than mental illnesses. Physics are more visible and easier to deal with. Furthermore, mental illness tends to be stigmatized. With the use of AI we can visualize a scenario that helps to break down barriers, including psychological ones. First, it is difficult to access psychiatrists. Bots can help identify any cases of attention and direct those cases to specialists. In addition, the range of digital markers, which can range from photos to voice variation, is difficult for us humans to recognize. Algorithms can detect them quite accurately. It is an area that deserves attention.
Another transformation that digital technology and AI will bring about in health is the shift in focus from curing the disease to maintaining health. The healthier the person and their immune system, the less prone to illness. Lifestyle and eating habits are an essential part of this process. Our own genetic activity is highly dependent on our control. Our genes can predispose us to obesity, depression or type 2 diabetes, but that's like saying that a piano predisposes us to playing wrong. But if we learn to play well, we can even be virtuous pianists. We will see increasing intercessions between health techs and food techs, as in addition to exercise, healthy eating is essential. Hippocrates already said “Let your food be your medicine and let your medicine be your food”. The awareness that food and health are closely correlated is already well established. The study “Association Between Dietary Factors and Mortality From Heart Disease, Stroke, and Type 2 Diabetes in the United States” shows with reasonable evidence that about one in two deaths from cardiovascular disease or diabetes were caused by an inadequate diet.
The AI can help us a lot. Based on the analysis of our DNA, which every year tends to make its cost cheaper, we can come reasonably close to balanced diets according to our own body's metabolic response.
We still have many barriers, even the lack of knowledge of many doctors about what to do with gene indicators. For example, in 2018 I did the sequencing of my DNA to identify indicators related to my metabolism in relation to food. The analysis showed indicators of propensity for absorption or not of excess fat, metabolic consumption when at rest, lipid metabolism and so on. I took the results to two doctors, who looked at them carefully, found them very interesting, but frankly said they didn't know how to evaluate them and prescribed traditional blood tests. The explanation is simple. Medical schools cannot keep up with the rapid evolution of medicine itself, leveraged by the exponential evolution of digital technologies and the potential of AI.
Recommended diets often have no effect as each individual has a different metabolism. The study "Personalized Nutrition by Prediction of Glycemic Responses" shows that the effects of diets must be analyzed in a multidimensional way, combined with lifestyle and exercise habits, age, weight, sleep quality, gut microbiome and other markers. Therefore, the use of population averages does not make much sense for a specific individual. Using digital technologies and AI we can get closer and closer to personalized nutrition, and maybe we'll soon reach a point where we can interact with our smartphones or wearables and ask “Siri, What Should I Eat?”
The result is that the use of digital technologies and AI can help us stay healthier and consequently enable an increase in the longevity of our lives. One company, a subsidiary of Google, Calico, and several others are working on this purpose. An article “'Extraordinary' Breakthroughs In Anti-Aging Research 'Will Happen Faster Than People Think'” shows that maybe it's not that far off.
The bottom line is that AI plays an important role in our society. It is a transformative technology and as such has the potential to transform, changing and shaping society itself. In medicine, causing the shift of current paradigms of what we understand today as medicine, from disease to health; from mass therapeutics and medications to personalized; and from punctual care aimed at curing diseases to continuous health monitoring activity. We still have a lot to evolve, but I believe that each year we advance much more than the year before.
In a recent e-book I wrote, available for download and free distribution, I explored more fully the application of AI in healthcare. In the e-book I try to show the profound transformation that the health sector is already going through with artificial intelligence tools, with real examples of how professionals and companies in the segment are already being deeply positively affected by their use. The purpose of the e-book is to present an overview of the brave new world of medicine, in which AI algorithms will be more accurate in image analysis, personalized diets will be developed based on your DNA, smartphones will monitor your health in real time, and hospitals will not. beds will have. Welcome to the future of medicine. Interested? The download link for the e-book “The Future Has Already Come to Medicine” is here.
