Back

Disrupted seasonality and association of COVID-19 with medically attended respiratory syncytial virus infections among young children in the US: January 2010-January 2023

Wang, L.; davis, P. B.; Berger, N. A.; Kaelber, D. C.; Volkow, N. D.; Xu, R.

2023-05-16 infectious diseases
10.1101/2023.05.12.23289898 medRxiv
Show abstract

Respiratory syncytial virus (RSV) infections and hospitalizations surged sharply in 2022 among young children. To assess whether COVID-19 contributed to this surge, we leveraged a real-time nation-wide US database of electronic health records (EHRs) using time series analysis from January 1, 2010 through January 31, 2023, and propensity-score matched cohort comparisons for children aged 0-5 years with or without prior COVID-19 infection. Seasonal patterns of medically attended RSV infections were significantly disrupted during the COVID-19 pandemic. The monthly incidence rate for first-time medically attended cases, most of which were severe RSV-associated diseases, reached a historical high rate of 2,182 cases per 1,0000,000 person-days in November 2022, corresponding to a related increase of 143% compared to expected peak rate (rate ratio: 2.43, 95% CI: 2.25-2.63). Among 228,940 children aged 0-5 years, the risk for first-time medically attended RSV during 10/2022-12/2022 was 6.40% for children with prior COVID-19 infection, higher than 4.30% for the matched children without COVID-19 (risk ratio or RR: 1.40, 95% CI: 1.27-1.55); and among 99,105 children aged 0-1 year, the overall risk was 7.90% for those with prior COVID-19 infection, higher than 5.64% for matched children without (RR: 1.40, 95% CI: 1.21-1.62). These data provide evidence that COVID-19 contributed to the 2022 surge of severe pediatric RSV cases.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.