"Unveiling the underlying severity of multiple pandemic indicators"

Objective: Multiple interconnected key metrics are frequently available to track the pandemic progression and one of the difficulties health planners face is determining which provides the best description of the status of the health challenge. In this study three COVID-19 indicators broadly used to monitor the evolution of the pandemic are analysed: the numbers of daily hospitalisations, ICU admissions and deaths attributable to the disease. The aim of the paper is to capture the information provided by these magnitudes in a single metric that reveals the underlying severity. Methods: Drawing on official Spanish data, we use one-sided dynamic principal components to convert a multivariate framework in a univariate scheme. The time-varying relationship between underlying severity and the number of positive cases is estimated. Results: A single component adequately explained the variability of the indicators during the analysed period (May 2020–March 2022). The severity indicator was stable up to mid-March 2021, fell sharply until October 2021, before stabilising again. The period marked by a fall coincided with the period of massive vaccination. By age group, the association between underlying severity and positive cases in those aged 80+ was almost 20 times higher than in those aged 20-49. Conclusions: Our methodology can be used in other infectious diseases to provide policy makers with a single metric that describes the severity status of the disease and enabling them to monitor the evolution. The synthetic indicator may be useful for prioritizing the vaccination of high-risk groups and evaluating the severity reduction.