By Nitin Goyal, MD, Orthopaedic Surgeon, Founder & CEO at Pulse Platform. Originally published on Quora.
Physicians are busy. So are patients. We’re all moving faster than ever.
We’ve made it easier to communicate. To stream shows. To buy a pair of shoes. People look around and they see so many aspects of their lives becoming automated, becoming more convenient. And they want that same ease and convenience when it comes to medical care.
But that can be a difficult proposition. Because when people believe they’re ill, they generally have to go to their doctor’s office or hospital. Even if they call their doctor, there’s no guarantee that they’re making the call at a time when their doctor is available.
As a doctor, I want to be there for my patients. I want to provide that round the clock service. But I’m limited by time and the sheer number of patients I see each year. There are only so many hours in the day doctors can be available.
Chatbots As A Solution
Digital health chatbots could be part of the solution to this problem. If you’re wondering how a chatbot can take the place of a doctor, well, it won’t. But chatbots can do some of the triage work associated with a patient’s initial complaint. To be clear, the chatbot won’t be making an actual diagnosis. Instead, it works by guiding the patient and helping to identify if something is seriously wrong with them. From there, patients could move on to medical care from their doctor or provider.
And in some cases, patients are actually more forthcoming with the chatbot than they are with their doctor. This is especially true for mental health issues. Those are issues that patients can be uncomfortable with. There is still a stigma there, and a lot of people don’t like talking about those problems. We’ve known about the human tendency to speak freely to computer programs since the 1960s, but now we might have a way to harness it and improve patient outcomes.
AI And Machine Learning
So how is this even possible? Artificial intelligence is everywhere right now, and a lot of people throw the term around casually. But there’s really much more to creating an intelligent system than just AI, and chatbots are actually a great lesson in what makes an intelligent system.
When you teach AI to do something, you’re telling it what to do. You create the rules and teach it to perform tasks. That means if you want to design a program that’s great at playing chess, you have to be great at playing chess yourself. You have to be able to tell it what to do.
Machine learning is different. It’s more of a subset of AI, but it’s vital for systems like chatbots.
With machine learning, you don’t need to know how to play chess. You give the program examples to learn from. The only rules you create for it tell it how to learn and create algorithms. Machine learning creates better chatbots over time, because they can learn from experiences. But in order to be useful, the chatbots need large datasets to work with. They need plenty of old chess games to learn from. Or in our case, they need lots of patient data to get better over time.
Thoughts For The Future
It’s important to remember that all of this isn’t meant to replace doctors. There are still some fundamental concerns with chatbots. For one, I would of course want the chatbot to err on the side of caution. So if there’s any question about what’s going on with a patient, I would want it to defer and tell them to see a medical provider. Most people would agree that’s the reasonable thing to do. We wouldn’t want a chatbot making a decision that it isn’t certain about.
But then we have to know how many patients it will defer. Will it be so many that the chatbot isn’t useful? If the program is too sensitive and continually pushes patients to doctors, then it hasn’t really fulfilled its purpose. But on the other hand, what will happen if there’s a mistake and the chatbot guides a patient away from seeking immediate medical treatment?
The other thing we have to keep in mind is that no matter how good chatbots get, they can’t replace doctors. Chatbots don’t have intuition. They don’t have empathy. They can’t develop a relationship with the patient. People want ease and convenience, but they also don’t love the idea of having their actual treatment automated. They still want the personalized care that a doctor provides.
Is there tremendous potential in using chatbots to make it easier for some patients to get answers? Sure. There’s definitely value in using them to separate the issues from the non-issues that patients have. But there are still plenty of questions about the role that these chatbots can and will play in medicine.
By Nitin Goyal, MD, Orthopaedic Surgeon, Founder & CEO at Pulse Platform. Originally published on Quora.
For more trending tech answers from Quora, visit HackerNoon.com/quora.