• Last updated: Wed, Feb 7, 2024Status: Ongoing
• Jonas Björk, Dominik Dietler, Tove Fall, Mattias Ohlsson, Atiye Sadat Hashemi, Thomas Eriksson, and Mirfarid Musavian Ghazani

Surveillance of disease outbreaks using health counselling data linked with registers

The overall aim is to develop a generic syndromic surveillance system, based on deep and standard machine learning methods within applied artificial intelligence, to detect unusual symptom reporting to the 1177 health counselling service. The project will use data from the recent COVID-19 pandemic as an empirical example and will have the following specific objectives:

1. To detect clusters of unusual health counselling contacts, defined by for example age, gender, country of origin, disease history, socioeconomic adversity or geographic location, that may signal the emergence of new health crises, locally or more broadly in society. This aim will be achieved by using deep and standard machine learning techniques within applied artificial intelligence to identify unusual patterns deviating from seasonal trends and detect clusters as anomalies in sparse high-dimensional time series data.
2. To rapidly identify groups in society that may be particularly vulnerable in case of a health emergency such as the COVID-19 pandemic, locally or more broadly in society. This aim will be achieved by linking initial symptom reporting, residential information, and data on socioeconomic conditions and comorbidities with subsequent health outcomes such as hospitalizations, intensive care admissions and deaths.