Monday, 05 October 2020 17:27

Un estudio muestra que los modelos tradicionales no pueden predecir con certeza la evolución de una epidemia

Fit to data obtained in real time for the daily number of active cases in Spain (from March 1st to March 29th) Fit to data obtained in real time for the daily number of active cases in Spain (from March 1st to March 29th) Castro et al, PNAS
  • Los modelos epidemiológicos SIR, utilizados habitualmente para conocer el comportamiento de epidemias, solo pueden ofrecer un pronóstico probable y variable
  • La imposibilidad de predecir con certeza es una propiedad del proceso de propagación de la epidemia, más allá de la calidad de los datos o de las aproximaciones de los modelos

Un nuevo estudio muestra que los modelos epidemiológicos tradicionales, como los de tipo SIR empleados para pronosticar el comportamiento de epidemias, no pueden predecir con certeza la evolución de una epidemia, ni el pico ni el final, mientras la epidemia está teniendo lugar. “A lo más que podemos aspirar es a obtener predicciones probabilísticas, como las del tiempo, donde se nos informe de con qué probabilidad se puede alcanzar el pico antes de una fecha dada, por ejemplo”, explica la investigadora Susanna Manrubia, del Centro Nacional de Biotecnología (CNB-CSIC), que ha coordinado el estudio, publicado en Proceedings of the National Academy of Sciences (PNAS). El trabajo es fruto de la colaboración de los investigadores Susanna Manrubia y Saúl Ares, del CNB-CSIC; José A. Cuesta, de la Universidad Carlos III; y Mario Castro, de la Universidad Pontificia Comillas.

“Este problema que presentan los modelos tradicionales se puede atenuar con más y mejores datos, y con modelos testados en distintos contextos, pero no se puede resolver completamente”, añaden los autores. En los modelos tradicionales de la epidemiología se divide a la población en cierto número de clases o “compartimentos”: individuos susceptibles, infectados, recuperados, y varias otras dependiendo de cada caso particular. Estos “modelos de tipo SIR”, por las iniciales de las clases básicas anteriores, capturan las características fundamentales de la dinámica de un proceso de propagación de infecciones. “Pero en este estudio hemos mostrado que los datos empíricos no pueden predecir el curso futuro de la epidemia, cuándo llegará a su máximo, si habrá o no un repunte, cuál será el número final de fallecidos o si el confinamiento tendrá el efecto deseado”, detallan.

“Mediante los datos oficiales publicados por el Ministerio de Sanidad y las comunidades autónomas, con reportes diarios de casos confirmados, pacientes recuperados y fallecidos, obtenemos un conjunto de parámetros compatibles con las observaciones mediante métodos de ajuste estadístico”, indica la investigadora. “Si bien el ajuste es excelente para el conjunto de España y sus Comunidades Autónomas, la sensibilidad de los modelos SIR a variaciones en el valor de los parámetros, como la tasa de infección del virus, impide la predicción a medio y largo plazo”, añaden.  “Por desgracia, este resultado también implica que no podemos determinar en estos momentos la magnitud ni la duración de la segunda ola”, indican.

“Esta sensibilidad es análoga a la que se observa en los sistemas caóticos, donde dos condiciones iniciales ligeramente distintas divergen exponencialmente rápido. Este principio subyace a la imposibilidad de predecir el tiempo que hará más allá de unos pocos días vista. Es decir, el hecho de que un modelo epidemiológico de tipo SIR (que entraña un crecimiento exponencialmente rápido de las variables en la etapa de expansión de la epidemia) sea capaz de reproducir el pasado no implica que sea capaz de predecir el futuro”, argumenta los autores.

Proceedings of the National Academy of Sciences (PNAS). The turning point and end of an expanding epidemic cannot be precisely forecast. Mario Castro, Saúl Ares, José A. Cuesta y Susanna Manrubia.

DOI: 10.1073/pnas.2007868117

CSIC Comunicación

NOTE! This site uses cookies and similar technologies.

If you continue browsing or do not change browser settings, we consider your acepptance for using. Learn more

I understand

COOKIES POLICY

A cookie is a text file that is stored on your computer or mobile device via a web server and only that server will be able to retrieve or read the contents of the cookie and allow the Web site remember browser preferences and navigate efficiently. Cookies make the interaction between the user and the website faster and easier.

General information

This Website uses cookies. Cookies are small text files generated by the web pages you visit, which contain the session data that can be useful later in the website. In this way this Web remembers information about your visit, which can facilitate your next visit and make the website more useful.

How do cookies?

Cookies can only store text, usually always anonymous and encrypted. No personal information is ever stored in a cookie, or can be associated with identified or identifiable person.

The data allow this website to keep your information between the pages, and also to discuss how to interact with the website. Cookies are safe because they can only store information that is put there by the browser, which is information the user entered in the browser or included in the page request. You can not run the code and can not be used to access your computer. If a website encrypts cookie data, only the website can read the information.

What types of cookies used?

The cookies used by this website can be distinguished by the following criteria:

1. Types of cookies as the entity that manages:

Depending on who the entity operating the computer or domain where cookies are sent and treat the data obtained, we can distinguish:

- Own cookies: are those that are sent to the user's terminal equipment from a computer or domain managed by the editor itself and from which provides the service requested by the user.

- Third party cookies: these are those that are sent to the user's terminal equipment from a machine or domain that is not managed by the publisher, but by another entity data is obtained through cookies.

In the event that the cookies are installed from a computer or domain managed by the editor itself but the information collected by these is managed by a third party can not be considered as party cookies.

2. Types of cookies as the length of time that remain active:

Depending on the length of time that remain active in the terminal equipment can be distinguished:

- Session cookies: cookies are a type designed to collect and store data while the user accesses a web page. Are usually used to store information that only worth preserving for the service requested by the user at any one time (eg a list of products purchased).

- Persistent cookies: cookies are a type of data which are stored in the terminal and can be accessed and treated for a period defined by the head of the cookie, and can range from a few minutes to several years.

3. Cookies types according to their purpose:

Depending on the purpose for which the data are processed through cookies, we can distinguish between:

- Technical cookies: these are those that allow the user to navigate through a web page or application platform and the use of different options or services it exist as, for example, control traffic and data communication, identify the session, access to restricted access parts, remember the elements of an order, make the buying process an order, make an application for registration or participation in an event, use security features while browsing store content for dissemination videos or sound or share content via social networks.

- Customization cookies: these are those that allow the user to access the service with some general characteristics based on a predefined set of criteria in the user terminal would eg language, the type of browser through which you access the service, the locale from which you access the service, etc.

- Analysis cookies: they are those that allow the responsible for them, monitoring and analyzing the behavior of users of the web sites that are linked. The information gathered through such cookies are used in measuring the activity of web sites, application or platform and for the profiling of user navigation of such sites, applications and platforms, in order to make improvements function data analysis how users use the service.

Management tool cookies

This Website uses Google Analytics.

Google Analytics is a free tool from Google that primarily allows website owners know how users interact with your website. Also, enable cookies in the domain of the site in which you are and uses a set of cookies called "__utma" and "__utmz" to collect information anonymously and reporting of website trends without identifying individual users..

For statistics of use of this website use cookies in order to know the level of recurrence of our visitors and more interesting content. This way we can concentrate our efforts on improving the most visited areas and make the user more easily find what they are looking for. On this site you can use the information from your visit for statistical evaluations and calculations anonymous data and to ensure the continuity of service or to make improvements to their websites. For more details, see the link below privacy policy [http://www.google.com/intl/en/policies/privacy/]

How to manage cookies on your computer: disabling and deleting cookies

All Internet browsers allow you to limit the behavior of a cookie or disable cookies within settings or browser settings. The steps for doing so are different for each browser, you can find instructions in the help menu of your browser.

If you decline the use of cookies, since it is possible thanks to the preferences menu of your browser or settings, reject, this website will continue to function properly without the use of the same.

Can you allow, block or delete cookies installed on your computer by setting your browser options installed on your computer:

- For more information about Internet Explorer click here.
- For more information on Chrome click here.
- For more information about Safari click here.
- For more information about Firefox click here.

Through your browser, you can also view the cookies that are on your computer, and delete them as you see fit. Cookies are text files, you can open and read the contents. The data within them is almost always encrypted with a numeric key corresponding to an Internet session so often has no meaning beyond the website who wrote it.

Informed consent

The use of this website on the other hand, implies that you paid your specific consent to the use of cookies, on the terms and conditions provided in this Cookies Policy, without prejudice to the measures of deactivation and removal of cookies that you can take, and mentioned in the previous section.