Technology companies prepare the assault on primary care

“My head and belly hurt.

I have had diarrhea,” she writes on ChatGPT and Gemini Raquel Jiménez, an 18-year-old student who is in her last year of high school at a high school in Seville.

She believes that the exams and the proximity of the university entrance exams are causing her discomfort, but the closest medical appointment they offer her is in 15 days.

“I'm sorry you don't feel well,” both platforms respond before summarizing possible causes, proposing remedies and recommending going to the doctor if the symptoms persist.

Both platforms, in their most basic versions, offer some empathy, possible diagnoses, and treatment suggestions.

Big technology companies are perfecting these systems to attack consultations and bring their conversational machines to the field of primary care, but always, they say, as a support tool, the doctor's co-pilot and possible solution to the collapse of primary care.

Artificial intelligence has proven extremely effective in specific medical practices.

“LucidWave [an artificial intelligence (AI) program] makes it possible to have ultrasound medical images available anywhere, with portable and wireless technology, process the data and provide a diagnosis prior to the doctor's diagnosis,” explains Belgian Bart van Duffel, member of the project and director of innovation at the Catholic University of Leuven.

At the last international congress of the Spanish Society of Preventive Medicine, Public Health and Health Management (Sempspgs), several hospitals presented applications already underway for the management of notifiable diseases (Reina Sofía de Córdoba), surveillance of hospital infections (Universitario of Ourense) and hip prosthetics (Jiménez Díaz) or tumor registry.

But these applications are advanced uses in specific areas.

Big technology companies want to go further and reach the beginning of the health system and establish themselves in primary care, which 81.4% of women and 72.3% of men attend at least once a year, according to the Statistics National Institute.

More information

Doctor Google

A review of 53 research on the impact of AI highlights the value of decision support tools to help improve error detection and medication management.

IBM, which has Watson Health, highlights that resources such as conversational robots are available at any time, collect basic information and alert of possible alterations that require additional attention.

In the professional field, Microsoft develops Azure Health Bot, a conversational system based on medical information, classification protocols and language models trained to understand clinical terminology.

Google has also entered this market with the family of models integrated into MedLM.

Greg Corrado, head of Artificial Intelligence for Health at the multinational, highlights the radiographic image analysis tools and AMIE, an application “optimized for diagnostic reasoning and conversation that emulates interactions between the patient and the caregiver.”

Miguel Ángel Mayer, specialist in Family and Community Medicine and author of an article published in

Primary Care

,

considers that, “although it is true that AI offers multiple opportunities and its future seems promising, it presents various challenges in its application, especially in medicine".

Advantages and challenges

Among the advantages, Mayer highlights its ability to automate administrative processes and facilitate decision-making by combining different data sources, from the patient's characteristics, their symptoms and the medications they take to test results or genetic data.

“They can help generate diagnostic and therapeutic suggestions, as well as play a very important role in the management of chronic diseases,” she writes.

But the specialist points out some limitations.

The first is the quality of the data that feeds the system.

Sharavya Shetty, a researcher at Google, has observed this deficiency during the development of AI for dermatological solutions: “Many data sets are collected in clinical settings, which means they could include skin cancers and more serious conditions, but they lack representation of more common concerns such as an allergic reaction.”

To solve this deficiency, they have turned to Stanford University in order to build a bank of images that include hair, nails, skin tones, age or sex of the greatest population diversity possible.

Mayer also warns of the dehumanization of medicine if the weight of technology is excessive.

Corrado also claims to have considered this limitation in his developments: “Empathy is a fundamental part of medicine and, therefore, we designed this model [AMIE] to communicate with respect, explain things clearly and support the individual in their process.” decision-making”.

Another challenge is hallucinations, false answers with the appearance of truth that AI systems can provide due to the deficiency of the data on which they are based or due to errors in generating the solution.

José Ruipérez is a primary care doctor and gladly accepts the proposals for conversational robots in his office, where he can dedicate between six and 10 minutes to fifty patients a day: “The robot can collect information before the in-person consultation and incorporate analysis data, electrocardiograms or relevant patient information.

That would allow us to optimize the time we can dedicate to them.

But he is not going to do the examination and he is not going to look the patient in the eyes either.

Some patients who are more familiar with technology will be able to use these tools, but the older ones want and need personal treatment.”

In this sense, technology companies insist time and again that their developments are not substitutes but complementary.

Corrado is blunt: “We are not talking about giving advice, we are not talking about making a decision or sharing a result or anything like that.

We're talking about the part of the conversation where the doctor gathers information and asks you about what's happening to you.

“We can explore how we do that in terms of being helpful, empathetic and helpful to people.”

You can follow

EL PAÍS Tecnología

on

Facebook

and

X

or sign up here to receive our

weekly newsletter

.

Subscribe to continue reading

Read without limits

Keep reading

I am already a subscriber

_