COVID-19 Alert and Surveillance Indicators

Abstract

Facing the arrival of the COVID-19 pandemic in France, Santé Publique France has set up an evolutionary surveillance system based on definitions of possible, probable and confirmed cases. But only cases confirmed by SARSCoV-2, RT-PCR (reverse transcriptase polymerase chain reaction) or serology, hospitalized cases and in-hospital deaths have been recorded on a daily basis. COVID-19 actual incidence has thus been estimated through additional indicators such as specific calls to emergency services (Samu) and SOS doctors, emergency rooms visits, or consultations in a sentinel network of general practitioners. Surveillance of non-hospital mortality has been impaired by delays and diagnostic inaccuracies of death certificates. Only in-hospital lethality could be reliably monitored. With a few essential statistical precautions and working hypotheses, models made it possible to anticipate the evolution of the epidemic based on two essential indicators: the reproduction ratio R, and the epidemic doubling time. In Ile-de-France region, the Greater Paris University Hospitals Group has used its data warehouse to complete this epidemic dashboard, including a fine modeling of patients’ care pathways. All these indicators have proved essential to plan the response to this unprecedented crisis.