E Process Med, the company located in the Central Building of the Alava Technology Park and associated with the Basque Health Cluster began the D&S-AI project in January 2022 together with Alberto Romero Palacio and José Antonio Girón González, INiBICA researchers and heads of services in hospitals in Cadiz. This project sought to improve the identification of COVID-19 negative patients using artificial intelligence. After two years of research, they developed a tool with a sensitivity of 94.6%, facilitating the management of patients suspected of SARS-CoV-2 infection. This breakthrough is a great help in the diagnosis and management of the disease.

The D&S-AI project, led by researchers from INiBICA and E Process Med, sought to optimise the diagnosis of COVID-19 in overburdened hospitals. The hypothesis was that a predictive algorithm, together with an antigen test and an optimised questionnaire, could improve the detection of COVID-19 negative patients. Initially, PCR diagnostics were slow and antigen tests, although fast, were not as accurate. D&S-AI combined both methods with artificial intelligence and machine learning to create a fast and accurate diagnostic tool, optimising healthcare resources and improving patient care.

This is a multicentre prospective clinical study, meaning that it collects data directly from patients over time and involves multiple health centres. This study looked at 1,500 patients from two hospitals in Cadiz, collecting information on their symptoms and evolution. The clinical approach uses real patient data, the prospective approach follows patients over a period of time to record events in real time, and the multicentre approach involves collaboration between several hospitals.

RESULTS IN THE DIAGNOSIS OF COVID-19

The D&S-AI project has achieved remarkable results in the diagnosis of COVID-19. Combining an antigen test with a 6-question questionnaire, an accuracy of 87% was achieved, with a sensitivity of 94.6% and a specificity of 83%. Without the antigen test, a 10-question questionnaire achieved an accuracy of 87%, with a sensitivity of 92% and specificity of 82%. This method is effective even in the absence of symptoms. In critical settings, accuracy is vital, but in primary care, the tool significantly improves diagnosis, optimising resources and time.

The D&S-AI project has used artificial intelligence (AI) because of its ability to handle and analyse large volumes of data quickly, which is essential in biomedicine. AI enables more accurate diagnoses, personalised treatments and optimisation in research, reducing time and costs. In this project, AI helped identify complex patterns and keep maintenance and diagnostic costs low. The success of the project was due to a multidisciplinary team of clinicians, scientists, bioinformaticians and deep learning experts.

The Institute for Innovation and Biomedical Research of Cadiz (INiBICA) is applying AI in several innovative projects. The PIBICRA Project develops a platform to improve cancer screening, the PEOPLE Project supports clinical decisions in patients with silicosis, the PERSUADE Project stratifies patients for atrial fibrillation surgery, and the FARMASTOCK Project optimises drug management in hospitals, saving significant costs. These projects reflect INiBICA’s growth and enthusiasm in applying AI to improve healthcare.

FUTURE TOGETHER

The next step for D&S-AI is its implementation in the Spanish healthcare system, evaluating the pace and scope together with E Process Med. Given the change in the COVID context, international markets in developing countries where the tool may still be relevant are considered. In addition, they plan to develop PathogenIA to diagnose respiratory infections, differentiating between bacteria and viruses, avoiding unnecessary use of antibiotics. The collaboration with E Process Med has been essential, combining the researchers’ clinical expertise with advanced AI technology to optimise diagnosis and treatment.

ACCURACY AS A RESULT

The most satisfying aspect of the D&S-AI project has been the final result. The creation of a diagnostic tool that combines a simple questionnaire and an antigen test, achieving a sensitivity of 94.6%, is a significant achievement. In addition, the 10-question questionnaire, without invasive testing, can diagnose with an accuracy of 92%. This tool is effective regardless of the patient’s symptoms, greatly improving the diagnosis of COVID-19 and justifying the effort and investment made.

In addition, E Process Med works in different areas of health of vital importance such as odontology, urology and general surgery. In the field of urology, thanks to their experience with experts such as their CEO Sebastían Armijos, a urologist with more than 10 years of experience, they are taking firm steps forward. An example of this is the online event they have organised for September, which will be the first of three planned.