Skip to main content
SearchLoginLogin or Signup

Review 2: "The Prevalence of SARS-CoV-2 Infection and Uptake of COVID-19 Antiviral Treatments During the BA.2/BA.2.12.1 Surge, New York City, April-May 2022"

Overall, reviewers note that the manuscript doesn't report details such as response rates or seem to control for potential confounders, which limits the manuscript's believability.

Published onJul 06, 2022
Review 2: "The Prevalence of SARS-CoV-2 Infection and Uptake of COVID-19 Antiviral Treatments During the BA.2/BA.2.12.1 Surge, New York City, April-May 2022"
1 of 2
key-enterThis Pub is a Review of
The prevalence of SARS-CoV-2 infection and uptake of COVID-19 antiviral treatments during the BA.2/BA.2.12.1 surge, New York City, April-May 2022
The prevalence of SARS-CoV-2 infection and uptake of COVID-19 antiviral treatments during the BA.2/BA.2.12.1 surge, New York City, April-May 2022

Abstract Importance Routine case surveillance data for SARS-CoV-2 are incomplete, biased, missing key variables of interest, and may be unreliable for both timely surge detection and understanding the burden of infection.Objective To determine the prevalence of SARS-CoV-2 infection during the Omicron BA.2/BA.2.12.1 surge in relation to official case counts, and to assess the epidemiology of infection and uptake of SARS-CoV-2 antivirals.Design Cross-sectional survey of a representative sample of New York City (NYC) adult residents, conducted May 7-8, 2022.Setting NYC, April 23-May 8, 2022, during which the official SARS-CoV-2 case count was 49,253 and BA.2.12.2 comprised 20% of reported cases.Participants A representative sample of 1,030 NYC adult residents >18 years.Exposure(s) Vulnerability to severe COVID-19, including vaccination/booster status, prior COVID, age, and presence of comorbidities.Main Outcome(s) and Measure(s) Prevalence of SARS-CoV-2 infection during a 14-day period, weighted to represent the NYC adult population. Respondents self-reported on SARS-CoV-2 testing (including at-home rapid antigen tests), testing outcomes, COVID-like symptoms, and contact with confirmed/probable cases. Individuals with SARS-CoV-2 were asked about awareness/use of antiviral medications.Results An estimated 22.1% (95%CI 17.9%-26.2%) of respondents had SARS-CoV-2 infection during the study period, corresponding to ∼1.5 million adults (95%CI 1.3-1.8 million). Prevalence was estimated at 34.9% (95%CI 26.9%-42.8%) among individuals with co-morbidities, 14.9% (95% CI 11.0%-18.8%) among those 65+ years, and 18.9% (95%CI 10.2%-27.5%) among unvaccinated persons. Hybrid protection against severe disease (i.e., from both vaccination and prior infection) was 66.2% (95%CI 55.7%-76.7%) among those with COVID and 46.3% (95%CI 40.2-52.2) among those without. Among individuals with COVID, 55.9% (95%CI 44.9%-67.0%) were not aware of the antiviral nirmatrelvir/ritonavir (Paxlovid™), and 15.1% (95%CI 7.1%-23.1%) reported receiving it.Conclusions and Relevance The true magnitude of NYC’s BA.2/BA.2.12.1 surge was vastly underestimated by routine SARS-CoV-2 surveillance. Until there is more certainty that the impact of future pandemic surges on severe population health outcomes will be diminished, representative surveys are needed for timely surge detection, and to estimate the true burden of infection, hybrid protection, and uptake of time-sensitive treatments.

RR:C19 Evidence Scale rating by reviewer:

  • Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.



This study uses a representative survey design to estimate the prevalence of, risk factors for, and vaccine effectiveness against SARS-CoV-2 infection during a 2-week period in New York City. It also included questions about the awareness and uptake of outpatient treatment among those reported as having COVID-19. The authors make two claims – 1) routine SARS-CoV-2 surveillance “vastly underestimated” the true magnitude of infection; and 2) surveys are an important approach to understanding “true burden," "hybrid protection” and treatment uptake going forward.

While the evidence to support their first claim is reliable, at least in terms of the direction of the finding, I would consider the conclusion about the approach to be potentially informative and worth further consideration and investigation.

The argument for the first statement is laid out clearly in their background and include limitations of passive surveillance, namely the potential for someone with SARS-CoV-2 infection to not be detected because the person wasn’t tested, was tested at home, or just wasn’t reported. Seroprevalence studies done throughout the pandemic have long shown that there is a gap between the number of infections that were likely to have occurred (based on prevalence of anti-N antibody detection), with infection to case ratio being a key output of those studies. The magnitude of the difference reported here, however, is unique.

The rapid, representative survey approach used here has advantages and limitations worth considering. As they note, this approach appears to be quite timely. This survey about the April-May time period was completed in May and already in June this information is being made available to the public. Whether or not it estimates true burden is an assertion that it is much more challenging to verify. Within their analysis of risk factors and estimation of vaccine effectiveness, there were a number of data points that did not seem consistent with prior studies and/or scientific plausibility, which also raises questions about reliability.

Whereas seroprevalence studies tend to estimate an infection to case ratio of 2-3, the study investigators here estimate that there were 31 infections for every case reported. While it’s possible seroprevalence studies underestimate true infections (because of decreased seroconversion in infections after vaccination or challenges identifying reinfection) or that this ratio is changing drastically over time (as they propose), there is also the possibility this approach is overestimating that difference. It’s not clear how many people had access to or received this survey, to understand the response rate, or how subject this may have been to selection bias. A survey about COVID-19 may have been of greater interest to those who recently had it, and although the introductory line doesn’t mention COVID-19, it is already included in the second question, at which point people may have made the decision to participate or not. This may have driven a number of differences, including the estimated burden and some of the unusual findings such as “hybrid immunity” and vaccination with boosting being associated with an increased risk of infection. Findings about risk factors for infection were somewhat difficult to assess, as the layout of Table 1 was unclear (what testers and non-testers refer to) and it doesn’t appear that there was any attempt to control for confounders in the data presented.

As the authors note, the rapid survey approach used here can be a very important complement to other surveillance/research methods, as it provides key details on testing, health care seeking behaviors, and awareness/uptake of treatment, that are often not available. It’s not clear is the results are truly representative, and neither the methods nor their discussion fully explore the limitations in a way that allow the reader to draw a strong conclusion about the findings. For this reason, I would rate this conclusion as “potentially informative” and suggest that further analysis be done to try to validate and/or explore this data.

No comments here
Why not start the discussion?