Improved allergological diagnosis and treatment by machine learningImproved allergological diagnosis and treatment by machine learning

Scientific Research Project Number: MA 2021 02
Place: Department of Dermato-Allergology, Amsterdam UMC, location AMC


Allergological diseases are frequently increasing in the western world with an annual rise of 15%. Currently, more than 150 million citizens of the European union are suffering from allergic diseases. The diagnosis of an allergic disease is based on medical history and allergologicaltesting. In order to confirm the clinical relevance of positive test reactions consecutive consultations are conducted. However, this whole diagnostic procedure is time consuming, expensive and highly depends on the medical skills and efforts of the doctors. Thus, the Department of Dermato-Allergology, one of the leading mondial departments, uses since 2004 a standardized form for taking the medical history. These data are stored in the European Surveillance System of Contact Allergies (ESSCA) database together will the performed allergy tests, the outcomes and the clinical relevance of the test reactions.

Description of the SRP Project/Problem

The current diagnostic procedures in clinical allergy testing are too dependent on medical knowledge of individual doctors. The risk of making medical mistakes such as missing the right diagnosis is unacceptably high. Until now, at the department of dermato-Allergology many efforts have been made to minimize these mistakes as far as possible. Hereto, many redundant checkings and controlling procedures have ben introduced. However it is still dependent on human skills and commitments.

The goal of this project is the improvement of the diagnostic procedure with an interactive form of the medical history. This form can be programmed based on the available data from more than 30.000 patients. Using deep learning, an self-improving algorithm has to be developed with the continuous input from new patients and their evaluation of relevance of new allergic reactions.

Research goals

Improvement of medical diagnosing of allergies. Hereto, a self-learning form for taking medical histories of patients suspected from allergies has to be developed.

Expected results

Development of a deep-learning form for the medical history of patients with suspected allergies.

The research project can be conducted, in part, in European collaborating departments.

Time period (usually 7 months)

7 months


Prof. dr. dr. Thomas Rustemeyer, email



Prof. dr. dr. Thomas Rustemeyer, Department of Dermato-Allergology, Amsterdam UMC, location AMC,