JMIR Publications

ABSTRACT

Background:

The implementation of novel techniques represents an additional opportunity for the rapid analysis acting as a complement to the traditional disease surveillance systems.

Objective:

The objective of this work is to describe a web-based participatory surveillance strategy among healthcare workers (HCW) in two Swiss hospitals during the first wave of COVID-19.

Methods:

A prospective cohort of HCW was initiated in March 2020 at the Cantonal Hospital of St. Gallen and the Eastern Switzerland Children’s Hospital. For data analysis, we used a combination of the following techniques: loess regression, spearman correlation, anomaly detection and random forest.

Results:

From March 23rd to August 23rd 2020, 127,684 SMS were sent generating 90,414 valid reports among 1,004 participants, achieving a weekly average of 4.5 reports per user (SD 1.9). The symptom showing the strongest correlation with a positive PCR result was loss of taste. Symptoms like red eyes or runny nose were negatively associated with a positive test. The area under the ROC curve showed favorable performance of the classification tree, with an accuracy of 88% for the training and 89% for the test data. Nevertheless, while the prediction matrix showed good specificity (80.0%), sensitivity was low at 10.6%. Loss of taste was the symptom which paralleled best with COVID-19 activity on the population level. On the resident level, using machine-learning based random forest classification, reporting of loss of taste and limb/muscle pain, as well as absence of runny nose and red eyes were the best predictors of COVID-19.

Conclusions:

Nevertheless, we deem the presented surveillance tool highly useful in monitoring and predicting COVID-19 activity among our HCW.