Loyola University predicts the evolution of the coronavirus in Andalusia through artificial intelligence
Loyola University through its Institute of Science and Technology (LOYOLATECH) has carried out a study based on the use of Artificial Intelligence (AI) techniques and a mathematical model applied to epidemiology that has been able to predict that the peak of cases of coronavirus in Andalusia is yet to come. The study offers an estimate of the hospital response to the pandemic in the coming weeks in the different provinces of Andalusia, using epidemiological models commonly used in these situations. The scientists have developed a website (https://www.uloyola.es/loyolatech/proyectos/covid-19) where they detail the studies carried out and will post the updated data and their forecasts day after day for the duration of the pandemic.
One of the results of the study is that, in case of maintaining the 9000 Plan, approved on March 29 by the Andalusian Government to attend to possible 9000 infected by coronavirus, it would probably be necessary to reinforce the ICUs available in some provinces. At this time, as the new Plan 15,000 was approved on April 8, the researchers, based on the mathematical model developed, have verified that the Andalusian Health System (SAS) will be able to cover all the cases of infected patients foreseen with enough anticipation and guarantees.
The research led by the team of Professor Pedro Rodríguez, in collaboration with the German research center Mosaiques Diagnostics & Therapeutics, points out that, if Plan 9000 were maintained and according to the data produced by mathematical models, the ICUs of some Andalusian provinces wouldn’t have been enough. However, it also indicates that with the activation of the new plan, the available ICUs would be enough to cover the demands indicated by the predictions in all the provinces.
Graphics extracted from the website prepared by the researchers https://www.uloyola.es/loyolatech/proyectos/covid-19
The study is described in detail in the article entitled “COVID-19 Impact Estimation on ICU Capacity at Andalusia, Spain, by using Artificial Intelligence”, which is the evolution of a study carried out in previous days entitled “Monte Carlo Deep Neural Network Model for Spread and Peak Prediction of COVID-19 in Spain ”, in which, using epidemiological models adjusted by AI techniques, the peak of COVID-19 infections was predicted to occur when 74 to 94 days have elapsed from the first case of contagion registered nationwide.
Mathematical models that determine the probability of infection
To develop the study, the researchers have used a mathematical model commonly used in epidemiological studies called SIR (Susceptible-Infected-Recovered). In short, this model describes the evolution of susceptible, infected and recovered people in an epidemic based on the probability that individuals have to be infected and to recover in case of infection. The question is to determine the value of the probabilities of infection and recovery in a specific population and for that, scientists have used AI techniques.
Specifically, official data from the provinces of Andalusia in the months of March and April have been used to apply them to this mathematical model and thus determine the probability of infection and recovery of individuals by province, considering 550,000 possible scenarios. In this way, it has been possible to estimate the future evolution of the pandemic in each province, concluding that the peak of infections in the different provinces will be reached between April 14 and 20.
In reality, this study initially aimed to know how the COVID-19 crisis would modify our energy consumption and how this would affect our pockets. However, once it was known how the pandemic could evolve, it was considered interesting to predict if the SAS ICUs could eventually collapse. For this, the relationship between infected individuals and individuals admitted to ICUs in the public system in each province was statistically analyzed. According to the main researcher Pedro Rodríguez: "The study tries to predict if the ICU beds, respirators and normal hospital beds in each province would be sufficient to cover potential patients according to an epidemiological mathematical model based on Artificial Intelligence.”
In the absence of other conclusive data, this type of study offers the competent authorities useful estimates for decision-making sufficiently in advance and with an adequate magnitude in each province, facing the possible peak of patients that the SIR epidemiological model indicates that it is yet to come. This study is an example of how science and its tools can help to program hospital resources and other measures in the face of an increase in the number of infected people, thus helping public officials to develop policies aimed at minimizing the effects of the COVID-19 crisis on the population.
DISSEMINATION OF THE NEWS IN THE MEDIA:
Novaciencia, ABC Sevilla, ABC Andalucía, La Vanguardia, La Razón, Europa Press, Diario de Sevilla, Diario de Cádiz, La voz de Cádiz, Diario de Córdoba, Málaga Hoy,
Granada Hoy, Europa Sur, Diario de Almería, Huelva Información, Diario de Jerez, OK Diario, Cibersur, El Mira, El Día de Córdoba, Gente Digital, Cordópolis, PaperBlog,
InfoSJ, Religión Digital, Revista IDescubre
Radio Programme “Por tu Salud” Canal Sur Radio. April 13. Minute 55
Radio Programme “Las mañanas de Andalucía” Canal Sur Radio. April 17. Minute 49
ABRIL 13, 2020
