A doctor explains how artificial intelligence could improve the patient-doctor relationship



[ad_1]

Despite all the debate about replacing doctors with artificial intelligence algorithms, Eric Topol is not worried. Topol is a cardiologist at Scripps Research Institute, a geneticist and author of several books on the future of health care.

His last book is called Deep medicine: how artificial intelligence can make human health again (released from Basic Books). Topol states that human beings will always seek the desire to be cared for by other human beings and that AI can help strengthen that link and restore it – if doctors are willing to defend their business interests.

The edge spoke to Topol about the current workings of health care, privacy issues related to AI health and the importance of physician activism.

This interview has been slightly modified for clarity.

Before talking about how the AI ​​will affect health care and the patient-doctor relationship, can you tell me what this relationship looks like now?


Photo courtesy of Eric Topol

The relationship has deteriorated. It's really terrible. The patient is badly taken because you have so little time with your doctor and he does not even have eye contact during this time. And it's not just time. It's the distraction of doctors who work as data clerks. In this case, you will not be a good listener. Doctors are terribly disenchanted and disillusioned, exhausted and depressed.

Plus, we have all this data – gathered from genomes and sensors – and everything is dressed with nowhere to go. I open the book with one of the stories that reinforce this for me, namely that this information was not treated by the doctors and the people who took care of me, which I hurt and healed me. This reinforced the fact that if all of this data was provided to people, we could make things safer and get better results and make them more effective, but clinicians are too busy. They can not necessarily put their arms around each person's data. I think that the transformative potential of artificial intelligence has the power to improve the human aspect of medicine, something we have lost.

The human aspect of medicine is also why you think AI will never replace doctors, right?

We not only need human monitoring, because you can not always trust an algorithm, even if it's validated – it can be hacked – but I think we'll always be looking for that link, this intimacy of communication. We used to have it. It was precious, and I remember it. Over time, the activity of medicine has taken over and all these forces have eroded this relationship. We can recover it.

Let's talk first about the diagnosis of AI. It seems that every week, a study shows that AI can better diagnose a disease than doctors. How is it going to be played?

AI can see things that humans can not see. In-depth learning causes the machines to see things much better than what a human will ever see and we begin to realize all of these things that we would never have guessed before. There are so many examples now. You can determine the level of potassium in your blood on your watch without some blood. You can analyze the retina to see if it is a man or a woman with great precision. You can analyze a colonoscopy and the artificial vision will detect the polyps forgotten by the gastrointestinal doctors. The list is lengthened increasingly.

The missing piece, of course, is the careful and rigorous prospective studies with validation and replication. We have the promise now. We've seen enough data and they're as exciting as anything I've seen in my 40 years of medicine, but we also need to go from excitement and hyperbole to reality and irrefutable evidence. .

You envision a world in which we will have help in artificial intelligence for diagnostics as well as algorithms integrating all these data sources. What would it look like in the clinic?


Image provided by Eric Topol

Then we have a different world. When you see patients, you are not trying to navigate these different pages and data sources. You are at: "Ok, I'm going to contextualize this for my patient. I will have a meaningful relationship and understand the presence of the person in order to express my human wisdom and empathy. This is a completely different look from what we have now.

Everyone benefits if you are more efficient and the doctors get up and say, "We are going to give this back to our patients." The data is more loaded and more powerful through algorithms, while at the same time unburden the burden of clinicians. And then, clinicians improve their performance and effectiveness and remember why they started taking medication. This creates a flying effect. You get both sides to get this performance improvement, and that fundamentally changes all clinicians' perspectives.

What about privacy?

It's really important. It is essential that each person has their data and collects it so that it is not only what is in your medical records, but that it is scattered across many places, hospitals and sensors. Right now, no one has all their data, although you want it from the moment you're in the womb until you get an assessment.

The biggest thing we can do is give users ownership of the data. We need to evaluate data security and privacy, but it also implies a different ownership model.

How do you know that efficiency will be given back to patients, instead of forcing doctors to simply see more patients in a shorter time frame?

I spent a few years at the request of the British government to help review and evaluate the national health system. Economists were working on it and it was striking that every minute saved in these speech recognition environments [where doctors aren’t sitting at a keyboard inputting information] translates into a huge amount of time to release for the doctors. The exponential impact is pretty incredible.

AI can improve efficiency, productivity, and workflow, but if we go that route, we must be willing to defend our patients. This has not happened in the past. If we continue to live as is, the medical community will be more and more in a hurry and there will be more burnout, more depression and suicide. The real test is whether the medical community can withstand commercial interests. We have sometimes been passive, and we can not afford that anymore. Activism will be needed to make humans more humane, machines to improve and to improve human effects.

What types of activism can doctors do?

You did not have the habit of seeing doctors standing up. It's only in recent years that you've seen the National Rifle Association say "stay in your way" about gun policy and you've seen doctors get up. These tend to be the youngest, not the old dogs, which is too much for us in medicine. You are starting to see doctors talking, and we can do it on a large scale for the most important of all, namely the restoration of health care. I am convinced that there is a way to do it.

How far is it?

We have evidence that people are interested. It's very fast, but over the years I've learned that no matter how much time it takes, I should probably multiply it by four or five. I learn that even when you have something so exciting, it takes a lot more time.

There is not enough investment in the high quality research needed. Companies currently developing a radiology algorithm or a dermatology or voice recognition algorithm at the clinic are among the best in this field. They do not necessarily have the necessary resources, unless they have been acquired by Google or an Amazon (which has no experience in rigorous medical research either).

It is difficult to show evidence. The medical community and, in this case, the patients, will probably not accept this reconfigured health care without evidence. Of course, this should be rigorous, because if you have a faulty algorithm, you can hurt a lot of people very quickly, so there should be very strict criteria and requirements for research that is in large numbers on people and various places. We need evidence that no one can argue, and then it will go much faster and we will start to have the momentum we need.

[ad_2]

Source link