Image: K-Health Facebook page
Anyone who knows me for more than an hour knows that I am a professional hypochondriac. Every mole, every pain, every little redness I discover - straight to Google. And of course, no matter which symptoms we look up, the result will always always be: cancer.
The distance from there to hysterically arriving at the ER, while running through every possible disaster scenario in my head, is very short. The thing is, statistically, for someone out there, the cancer result was actually accurate. Even a broken clock is accurate twice a day.
The problem: Google
The problem begins when people like me arrive at the urgent care centers unnecessarily, as they could have waited a day or two to see their family doctor and thus the centers would have dealt with fewer false positive visits. But when a person who grew up all his life on the assumption that every question can get a reliable and correct answer through a Google search, receives the most horrible answer from Google - he doesn't wait a day or two and certainly not an hour or two.
This is the space in which the K-Health app operates. In short, K-Health is a bot that allows you to understand, through targeted questions, what you suffer from and how other people in your situation have been treated. The system actually guides the user through a flow whose goal is to narrow down the options and reach 2-3 diseases or phenomena that may, with a high probability, be the disease the user suffers from. Think of an episode of 'House', only in an app and without the cynicism we've come to love.
Basically, K-Health offers its users to test themselves using the app before they run to the emergency room, providing a reliable and high-quality buffer that reduces the load on emergency rooms and emergency medicine centers.
The solution: humanity, data and a perfect world
The challenge of K-Health as an application that deals with personal medicine, is to create a system that will acquire the trust of the user. One that the user can trust and know that, despite what Google said, there is no need to run to the emergency room right now, and it is enough to make an appointment with the family doctor for tomorrow morning. So how does K-Health do it? How does a bot, which is fundamentally non-human, manage to evoke in the human user feelings such as trust and confidence? Let's dive into it:
Femininity
Yes, femininity. We all came from the womb, and some of us want to go back there at any given moment in out adult lives. The feminine-maternal temperament prevalent in human culture throughout the generations is a calm, contained temperament, one that overflows with a subconscious feeling of a mother holding and cradling you when you were a baby.
The K-Health bot is trying to immitate that character, somehow, as much as it can be done with artificial intelligence. The bot that the app provides to users has been deeply and comprehensively personified, so that the character that is reflected to the user is a female character, calm and reassuring, which makes you believe that everything will be fine, there are other people who suffer from the symptoms you suffer from, you are not alone in the world. Here, take a pie chart and see for yourself.
When the user talks to K-Health, the motherly and inclusive bot, the feeling is like being treated by the best doctor in the world. She sees you, listens to every complaint, doesn't rush you out the door, treats every symptom you mention in depth. It digs deeper and reaches high resolutions regarding each pain - the design developed to make this possible is really impressive. When the bot asks you about another symptom and you answer in the negative ("I don't have headaches"), she even encourages you: "I'm glad to hear." In general, you feel that someone hears you and the knowledge that this is a machine that cannot forget or skip any detail, reassures you.
Data
But it doesn't end here. The female character of K-Health leads you through many diagnostic questions, at the end of which she provides a sort of answer. Of course, there is a clear disclaimer before that which says that the answer given is not absolute and does not replace a visit to a real doctor, but the bot also conveys this in a pleasant and non-threatening manner, as are many other disclaimers.
The moment of truth arrives, you ask K-Health to view the answers it found about your medical condition, and then some kind of magic happens. The contained and maternal figure also becomes smart, knowledgeable, data-based. The application presents the user with 2-3 options that can explain what they are suffering from, along with the probability of each symptom. Clicking on each phenomenon shows the user the prevalence of this phenomenon in the general population, and also what to do next, what are the relevant treatments, and what is the recovery timeframe. All this data is presented to the user via a clean and clear infographic, which instates an additional sense of trust in the user. Okay, this Kay knows what's going on, she has data, she's accurate.
K-Health's data collection system is dynamic and a few days after the user receives a reply from the application, the Kay bot returns to them asking how they are doing, whether they were diagnosed by a doctor, which doctor it was and what the treatment was. This way, K-Health maintains the freshness and authenticity of its data, which instills in the user a feeling of trust and closeness.
A world without cancer
This is where I get into speculation, because the guys at K-Health don't exactly reveal the method to everyone, but of all the countless tests and scenarios I've run through the app (which is probably currently diagnosing me with Munchausen's disease), I haven't yet received a result of any cancer. I assume that in some cases the app does show some kind of softened version of suspected cancer or suspected tumor, but certainly not as often as Google gives the same answer.
Furthermore, when the user starts chatting with Kay, and she asks them what they suffer from or what bothers them, they choose from a closed list of symptoms that does not contain diagnoses. I mean, you can't go into the app and search for: "lymphoma" or "inflammation", because inflammation is already a diagnosis, and self-diagnosis is exactly what K-Health wants to prevent.
So while the world of medicine is changing, medical information is flooding us online, and patients are coming to the doctor armed with data from Google and their personal interpretations, K-Health stands with the almighty bot Kay, and tries to calm things down.
Comments