Technology helps us to be more human

Paco Miralles

"Learning to read is the most important thing that has happened to me in life," acknowledged Vargas Llosa in his speech at the reception of the Nobel Prize. I share his discernment, although I have had some more important achievements. This task is something inherent to the doctor, one of the weapons that help the indispensable humanistic aspect of medicine.

When you read Daniel Kahneman in Thinking, Fast and Slow you realize how human reasoning works. There is a system 1 that operates quickly and automatically: it involves little effort and offers little sense of control; thanks to it, we automatically answer the result of 3 + 3. And there is also system 2, which needs more attention and concentration, involves more reflection and helps us, for example, to understand (or not) the de-escalation phases of the COVID-19 epidemic.

Faced with a clinical decision, an experienced doctor recognizes the problem and the fast track starts working with a high probability of success with little effort. If the problem is not recognized, we collect information in other ways and, using a more analytical and complex method, we make a judgment that has cost us more effort and energy and that produces a slower response. The balance between the two is a must. This involves continuous wearisome and tedious vigilance. Also, it’s difficult to be persistent when the sources of information are many and changing, we move in scenarios of uncertainty, we accumulate fatigue and we rush into difficult decisions. The ideal would be to have the integration and analysis capacity of system 2 with the speed of the first system, something that to a great extent already exists, that is, the different artificial intelligence systems from which we can benefit so much in medicine.

During this epidemic, we have been learning from within the hurricane. Different scenarios have been developed with the availability of different diagnostic tests. On many occasions, the increase in the arsenal of tests gave doctors less certainty, not only because of their interpretation but also because of the dubious quality of the material. The literature is providing multiple articles and reviews to help clinicians understand its operation and interpret its results. It’s clear that certain specialities are more familiar with these tests, but it’s also true that their interpretation is going to be globalized in all specialities.

Eric Topol exposes the luddites - those artisans of the early nineteenth century who destroyed machines to avoid the unemployment they would cause - in medicine (see post). In his book Deep Medicine, he relates how artificial intelligence should help doctors stop doing jobs that a machine can develop. And we could use the extra time to recover the lost humanism.

This approach seemed like a perfect scenario to try to develop a system that would help us in the diagnosis of patients infected by COVID-19. Until now, some tools help discern if the clinical picture was suggestive of infection, most of them focused on patients as instruments for their initial triage.

Some weeks later, and thanks to the collaboration with Presentyis, we developed a chatbot (virtual assistant) that directs the doctor through a series of questions about the case and indicates the diagnosis and the stage of transmissibility in which the patient is. The robot is called IMPAI and it’s a machine that collects all possible options combining the type of symptoms with their evolution and the availability or not of diagnostic tests. It also offers us options to assess asymptomatic patients who contacted confirmed patients.

We have carried out its development in open source with free accessibility for all healthcare personnel who need it. This type of code allows other people to improve the application so that it becomes a participatory project. We hope that this tool is the seed of a larger project and many others to come. The results matrix is based on clinical experience and the published document by the Ministerio de Sanidad y la Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica and is the soul of a system based on artificial intelligence, the chatbot.

So far it has been widely accepted by professionals. Proof of this is the high number of entries and the endorsement granted by the Sociedad Española de Medicina Interna, the Colegio de Médicos de Málaga and the Consejo Andaluz de Colegios de Médicos. Therefore, the initial objective has been fulfilled: to provide help to our colleagues in the nightmare that we are living. Through this link, you can access the use of IMPAI.

They say that what is priceless is not valued. People who have read some of my previous posts know that I defend value medicine. For this reason, I don’t offer this tool altruistically but, in return, I ask that the time that this application can save us, we dedicate it to humanize medicine, to read and enjoying time with our families.

Link to blog Doctor Miralles