In this interview, given to uah.esnoticia, Raquel tells us about the findings of this research work that by developing an algorithm detects allergies early. The clinical study of this research was carried out at the Universitary Hospital of Guadalajara, under the supervision of Dr. Arantza Vega Castro, Head of the Alergology Service.

- How is drug intolerance detected by this method? What has been the process of creating this new allergy prediction method?

The allergy provocation tests (both food and medication) are to provide the patient with ever-increasing doses of the allergen suspected of, leaving a certaing time between each new dose. To confirm the allergy, the patient suffers some kind reaction. Reactions are very variable in terms of how they manifest of their severity, and it´s not possible to establish a pattern depending on the characteristics of patients, or the type of allergen.

At the Irish University of Cork (UCC), with which we collaborated, a doctoral thesis (Niall Twomey, supervised by Professor Liam Marnane) was developed in which, with data from 24 pediatric patients (under 12 years of age) performing food allergy provocation tests, a link was detected between certain variations in heart rate and the subsequent onset of allergic reactions.

My doctoral thesis, entitled  'Real-Time Early Detection of Allergic Reactions based on the Heart Rate Variability', aimed to determine the specific relationship between variations in heart rate with the subsequent occurence of allergic reactions and express it in a mathematical way, in order to try to detect such reactions through the electrocardiographic signal.

From the database available to us, and by linking the (positive-negative) results of provocative tests to the characteristics of electrocardiograms, we were able to establish the pattern that followed heart rate variability in most positive patients and did not occur for non-allergic patients. 

Raquel Gutiérrez Rivas

- First of all, can you explain to us how you have carried out the process of collaboration and development of the project? How has the collaboration been with the Guadalajara Hospital?

The database obtained in Ireland helped us establish a first pattern, but it was not extensive enough to continue the study, so we decided to contact Dr. Arantza Vega Castro, Head of the Allergy Service at the University Hospital of Guadalajara.

After listening to our proposal, he was not only willing to collaborate, but he proposed to extend the study on patients and the type of allergen, adding adults and drug allergies to the study.

From there, with the help of the allergy nurses, we began a data acquisition phase. For two years, we obtained more than 170 new electrocardiographic signals, wich adds more than 630 hours of measurement. During this time, and as we acquired new data, we were able to improve the detection and analyze in which cases it was an improvement and the magnitude of that improvement.

Described that in many of the tests the algorithm detected allergic reactions even at lower doses (in the early stages of the test), than those necessary for allergic reactions to manifest. That is, in such cases, both the test time and the dose of allergen provided to the patient could have been reduced. This would have reduced the risk of severe reactions, increasing the safety of provocations and the quality of life of patients undergoing them.

- How do mathematical algorithms play a role in this process? Why use algorithms for this purpose?

Once the pattern defining a possible allergic reaction is established, the next step is to detect it on new patients and generate some type of alarm or warning.

An algorithm is used to perform this search automatically, so that there is no need for a person to continuosly observe each patient´s heart rhythm during these tests. It should be noted that several patients perform these tests simultaneously, and may last up to five hours.

The defined algorithm can be programmed to a portable wireless device, so that it automatically and continuously analyzes the electrocardiogram, obtains the heart rate and variability, and finds the set pattern. In case of detection, the device itself can alert the medical personnel. This process would be complementary to the current one, as it does not interfere at all with the procedure currently used in hospitals.

Our proposal improves patient safety, as the study demonstrates, these detections can be made earlier than with traditional means. It also facilitates the work of medical staff, who have an additional and objective source of information and the possibility of receiving a warning before the patient reacts to the allergen.

- How can mathematics help predict human body reactions? Could this application detect drug allergies help detect other types of health problems?

All the devices we use today to measure physiological parameters (blood pressure meters, thermometers, cardiographs, etc.) require signal processing based on mathematical algorithms. The comparisons we make between parameters of two different individuals or from two different states of the same individual are made mathematically.

The electrocardiogram is used to detect numerous diseases (not just heart disease) or certain physical conditions. Because it´s controlled by the central newvous system, and this control is perfomed according to body needs, it´s possible to extract a large amount of information. As an example, it can be detected that a person is feeling pain, which is tremendously useful in patients who cannot express themselves (people in coma, neonates).

In our case, the algorithm we have developed is specific for the detection of allergic reactions, and it´s necessary to increase the database and continue the study to make this detection as accurate as possible.

Publicado en: Inglés