Relationship between clinical symptom profiles and COVID-19 infection status during Delta-dominant period versus Omicron-dominant period-analysis of real-world data collected in Hiroshima Prefecture, Japan.

Objectives: The present study investigates the diagnosis and prediction of COVID-19 based on clinical symptoms, and corresponding difference between the Delta- and Omicron-dominant periods, using data collected at polymerase chain reaction (PCR) centers in Hiroshima Prefecture, Japan.

Methods: Data was collected using a J-SPEED-style COVID-19 standard data collection form. The analysis was done in two directions: calculating the likelihood ratio that clinical symptoms will manifest in "infected" versus "non-infected" individuals and calculating the diagnostic odds ratio (OR) of infection for those who have symptoms compared to those without symptoms.

Results: COVID-19 was more strongly associated with smell and taste disorders during the Delta-dominant period, and muscle pain during the Omicron-dominant period. An age-specific analysis of likelihood and diagnostic ORs found cold-like symptoms had the lowest ability to diagnose COVID-19, and the lowest predictability of COVID-19 with children during both periods. The likelihood and diagnostic ORs of other symptoms for COVID-19 were highest in adults and lowest in those over 65.

Conclusion: Symptoms are an important indicator of COVID-19, but the association between specific symptoms and COVID-19 is dependent on the dominant variant of the virus.