Hundreds of articles have been written on Artificial Intelligence and how it will have a deep impact on every aspect of our life. Few, however, have been addressing how AI can and will actually keep us healthy.
In healthcare there’s historically a big focus on curing those who have already been struck by illness. This by itself is quite logical as throughout the history of medicine few tools and resources were available to continuously follow up an individual’s health closely and prevent a disease from happening at all.
Preventive medicine is nothing new however. Even “The father of medicine”, Hippocrates and his followers were very concerned about preserving health through proper diet and activities, such as exercise and getting enough rest. Limited by what could be measured through taking irregular snapshots, the feedback of the body was mostly hardly understood.
The times they are a-changin’ though, to quote a recent Nobel prize winner. Every day, new technological breakthroughs are made and more and more we are able to actually “listen” to our own body in real-time. Even though not all technology provides proper insights yet, it is clear this is just a matter of time.
While many people still get excited by the use of activity trackers to measure how many steps they take every day, a world of much more advanced insights is looming. Ready to radically change healthcare and push towards a more empowered population focusing on staying healthy and prevent illness with the need to cure as the last resort.
Technology is just one piece of the puzzle though. This shift in healthcare will need policy changes as well, as in Europe for example, only 3% of the healthcare budget is spent on prevention. All the rest is still being spent on treatment while much treatment wouldn’t be necessary if a disease would be prevented in the first place.
So what role does AI play in this shift towards a more preventive mindset for healthcare? A big role is the answer. Capturing real-time physiological and wellbeing data results in vast amounts of information. Information needing to be processed and from which lessons should be drawn in order to be meaningful.
What’s the impact of an evening out on your health? What’s the impact of a lack of activity or too much sugar? How does stress affect your health? There is an infinite amount of questions to be asked and an equally infinite amount of lessons to be learned. Lessons to adapt your life in a healthier way, or at least to raise the awareness of one’s health status and the impact of a certain lifestyle on it.
AI can do just that.
AI can look for patterns and make suggestions on lifestyle improvements. It can spot health problems early on and enable you to take action before they turn into a disease needing a cure.
It’s not the intention to give a deep technological insight in how AI is used but rather to peek into the possibilities of AI for our health.
So where are we now in this journey? Let’s have a look how some Digital Health startups are already using AI today to keep us healthy tomorrow.
Digital Health Startups Using AI
Few Digital Health startups in that list are focusing on empowering the still healthy people to stay healthy though.
To have a good understanding of the companies that do focus on preventive care, let’s look at the data they are using to accomplish their mission. As AI is nothing without data, we can distinguish a number of data sources used by these companies to provide advice on how to preserve or improve our health.
Packed with sensors and being used heavily by people all around the globe, the smartphone is an enormous source of behavioral data. Touchkin, an Indian startup, uses passive sensor data from the smartphone to track changes in behavior indicating ill-health. Its prediction engine learns patterns in behavior that may mean mental or physical ill-health, and connects the user with direct support.
The advantage of the smartphone is that it can act as gateway for any type of information. Think about easily integrating data about your nutrition as done in one snap by apps such as Lose It and Ava. Or using the tone of your voice to detect the onset of illness as done by Sonde. There’s already so much user generated data on the smartphone waiting to be transformed into valuable health advice.
In order to also continuously measure physiological data or more advanced activity and posture patterns we need something more than the smartphone alone.
Swedish startup Magnea has a Machine Learning based platform using on-body sensors to detect traumatic falls, what side of the body you are lying and the more common movements. A more sport-oriented wearable solution, Boltt, coins itself as the wearable of the future and uses an AI-based health assistant. This type of virtual assistants are popping up more and more in all kinds of domains.
The biggest predictive power of our future health most probably lies in genomics. By understanding our own genetic code we can get more personalized and actionable insights leading to better health. Clinical genetic tests can tell you whether you are at risk of a certain particular disease for example.
There are few companies however directly selling a genomics based product directly to consumers with the most well-known example being 23andme. Now how does a genomic company like 23andme use AI? It uses AI (Machine Learning) to create genetics-based risk prediction algorithms based on the immense genomic data sets it has. In the DNA report you get from 23andme you’ll get a risk indication for a number of diseases. If you have an increased genetic risk for heart disease you can adapt your lifestyle with more exercise and a healthy diet to decrease the risk.
The vast amount of data available inside of social networks is another data source expected to play a role in preventing illness. One example of a company using social media to prevent illness is Sickweather who scans social media for signs of illness in your neighborhood. It goes without saying that the quality of the predictions is totally dependent on the social media usage in your neighborhood. Compare it to low quality sensors in wearable devices. If the input data quality is low you won’t get any accurate predictions either.
Social media can be particular interesting to detect early signs of mental health issues. Next to simply tracking the social media usage frequency more advanced patterns can be discovered through AI techniques such as sentiment analysis techniques.
Electronic Health Records
What about Electronic Health Records? Oh wait, wasn’t the goal to look at AI technology preventing us to get ill at all?
That’s right. We’re still far off of having no need for a medical history however and therefore medical records still provide an invaluable source of information. In fact, the aforementioned sources can be expected to all become part of your medical record of the future.
As more and more data is digitized in your current encounters with healthcare, typically in the hospital or with your doctor, this data and even data of other patients becomes increasingly important to limit the amount of future hospital admissions.
Currently most focus goes to AI-based Clinical Decision Support to support the clinician in obtaining an as accurate diagnosis as possible, based on all data at hand.
One company that’s already using your medical history today as part of its mobile consultations and advice offering is Babylon. An AI-based chatbot is combining the person’s medical history and a database of medical symptoms to perform a first screening and even making a direct contact with a doctor unnecessary in some cases. While this is currently mainly used when signs of illness surface, and hence you get medical assistance, it’s not hard to envisioning the same system will in the future help you detect early-signs of illness before you notice them yourself.
Artificial Health Intelligence
We’ve seen some examples of how A.I. is already used today to minimize the illness time of people. It will become even more interesting when we can combine all the data from the different data sources and have fully holistic view to prevent disease. Something which still requires some hurdles to be taken including the non-trivial interoperability hurdle.
The use of A.I. techniques will be pivotal in supporting the shift towards healthcare rather than sick care as we know it today. Provide people with the tools to stay healthy. Tools giving you a comprehensive view on your health by mining through all available data and continuously learning to provide you the best advice possible.
Are you ready to embrace Artificial Health Intelligence?