Generation of alternative models based on computational intelligence for screening and diagnosis of pulmonary Tuberculosis.

Project: Research Project

Project Details


Tuberculosis (TB) is a disease considered a global emergency by the World Health Organization (WHO), because in the last five years it has remained the largest cause of death caused by an infectious agent.

It is a disease that affects around 10 million people annually, and can cause death [1].

Infection by the My cobacterium tuberculosis bacillus is one of the leading causes of mortality worldwide, causing illness in 10.4 million people in 2016 and causing death in 1.7 million people, including 0. .4 million people with HIV co-infection [2].

It is estimated that approximately one third of the world's population has the infection and can develop the disease at any time, for example, when there is immunosuppression [1].

According to data reported by the World Health Organization and the Pan American Health Organization (PAHO) in the report on the End Tuberculosis Strategy, 18 countries in the Americas reported 14,402 deaths from TB, through their statistics. vital in 2015, many of these deaths could have been avoided [2].

In Colombia, the incidence and mortality rates due to tuberculosis are high.

For the year 2016, SIVIGILA (National Public Health Surveillance System) was notified of 13,871 reported cases of pulmonary and extrapulmonary tuberculosis, which affects other parts of the body, where 12,439 correspond to new cases with an incidence of 25.7 cases per 100,000. inhabitants[3].

Within the national protocol for the diagnosis of latent and active pulmonary TB, clinical screening is performed to rule out active TB based on specific symptoms.

If no symptoms are present, a tuberculin test (PPD) or interferongamma test (IGRA) is performed; if, on the other hand, one or more symptoms occur, it is investigated whether there is active TB or the symptoms correspond to other diseases [4].

However, throughout the national territory there are different problems related to the infrastructure necessary for these tasks due to characteristics of poverty, violence and availability of laboratories and specialized equipment for diagnosis.

Currently in the world, engineering applications in the health area are based on computational intelligence techniques, highlighting the so-called decision support systems (SAD), which have previously been used in economic systems, credit analysis systems and in marketing studies, to focus on the sale of products or services, and also the health area.

In the latter, they are useful in tasks supporting the diagnosis and prognosis of diseases, as help to professionals in the health area, where times and processes can be optimized [5].

The present proposal aims to use computational intelligence techniques to perform screening and diagnosis of active pulmonary TB, using algorithms based on neural networks and fuzzy logic applied to data that will be obtained in conjunction with professionals from the Santa Clara Health Services Unit belonging to the subnetwork. integrated health services center East.

Subsequently, the acquired variables will be processed, which will allow screening and diagnosis tasks to be carried out for low-income communities.

Initially, we will work on the Matlab software and subsequently develop a computer application to perform these tasks more automatically.

In this way, the general objective of this proposal is to generate screening models with alternative tools based on computational intelligence that allow a characterization of TB in environments with precarious situations.
Effective start/end date2/22/198/22/22

UN Sustainable Development Goals

In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all. This project contributes towards the following SDG(s):

  • SDG 3 - Good Health and Well-being

Main Funding Source

  • National


  • Bogotá D.C.


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.