- The project, the first in Spain and across the world in this area, stems from a collaboration agreement between the Community of Madrid (Department of Digitalization), Fundación 29 and Microsoft, to develop and implement innovative technological solutions that improve patient care and facilitate the work of healthcare professionals.
- As a pilot project, this web application will be accessible at Primary Care centers in the Community of Madrid from the end of September. It is based on OpenAI’s Artificial Intelligence language model, GPT-4, with Microsoft’s Azure OpenAI Service and is backed by full security and ethical development standards.
- It will enable Primary Care medical professionals in the Community of Madrid to improve their skills and diagnose patients with rare diseases more accurately and quickly, expediting patient referrals to the appropriate specialists for the right treatment.
- This agreement also opens up future routes for AI application to improve healthcare service delivery, optimizing the resources available in the Community of Madrid.
Madrid Healthcare Service (SERMAS) will be a pioneer in Spain and across the world in applying Generative Artificial Intelligence (AI) for the clinical diagnosis of rare diseases.
Through a collaboration agreement signed today by the Community of Madrid’s Department of Digitalization, Microsoft and its partner, Fundación 29, a pilot project will be launched at the end of September, providing Primary Care medical professionals in the Community of Madrid with a Generative AI-based application included into their IT platform.
On average, it takes patients with rare diseases at least five years to receive an accurate diagnosis and they have to visit up to seven specialists to reach this point. Despite this long process, more than half of these patients remain undiagnosed or are misdiagnosed. This can result in them receiving the wrong treatment that could be harmful and even aggravate their condition. With the support of the Generative AI application, DxGPT, from Fundación 29 and Microsoft, medical professionals can reduce the time this takes to minutes using an enterprise grade diagnostic process support tool. This web application does not replace physicians, but it does enhance their capabilities and helps them work faster.
Miguel López Valverde, Community of Madrid Head of Digitalization, said: «Artificial Intelligence, used with all the standards of responsibility and safety, is driving significant advances in Madrid healthcare. Projects like this demonstrate the advantages that new technologies are delivering in early detection processes, which are particularly useful for diagnosing rare diseases.»
«We are excited about this groundbreaking project using generative Artificial Intelligence to improve the diagnosis of rare diseases. This technology has great potential to support the work of healthcare professionals, reducing diagnostic times and improving accuracy, especially in complex cases. This pilot project will lay the groundwork for exploring new AI applications to improve patient care. We are grateful for the vision and support of the Department of Digitalization in implementing this initiative that will place the Community of Madrid at the forefront of digital transformation in healthcare,» stated Julián Isla, Head of Fundación 29.
«Artificial Intelligence is generating unprecedented opportunities for organizations of all sizes and from all industries. The healthcare sector is one where its application can bring the greatest benefits. We are delighted to collaborate with organizations such as the Community of Madrid’s Department of Digitalization and Fundación 29 to enable medical professionals from Madrid Health Service to provide a better service to patients,» stated Alberto Granados, President of Microsoft Spain.
IA always under medical supervision
The application, which functions as a conversational assistant, is based on AI Large Language Model developed by OpenAI with Microsoft’s Azure OpenAI Service, which helps leverage the capacity of the most advanced AI systems (including OpenAI’s GPT-4 and GPT-3.5), combined with Azure’s enterprise services and Microsoft’s AI-optimized infrastructure.
The process starts with a short clinical description of the patient that the physician enters into the tool. Based on this input and using GPT-4, DxGPT suggests a series of possible pathologies. Once the list of diseases is generated, more information can be added to fine tune the diagnosis, such as data sets and medical history or laboratory test results. This will help physicians make a more informed decision and refer patients to the appropriate specialists to expedite their treatment.