Newly Published Free
Correspondence  |   May 2020
Considerations for Assessing Risk of Provider Exposure to SARS-CoV-2 after a Negative Test
Author Notes
  • University of Washington, Seattle, Washington (W.V.C.). vancleve@uw.edu
  • (Accepted for publication May 1, 2020.)
    (Accepted for publication May 1, 2020.)×
Article Information
Correspondence
Correspondence   |   May 2020
Considerations for Assessing Risk of Provider Exposure to SARS-CoV-2 after a Negative Test
Anesthesiology Newly Published on May 20, 2020. doi:https://doi.org/10.1097/ALN.0000000000003392
Anesthesiology Newly Published on May 20, 2020. doi:https://doi.org/10.1097/ALN.0000000000003392
To the Editor:
Coronavirus disease 2019 (COVID-19) is caused by infection with the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). The transmission of coronaviruses occurs via direct contact, droplets, and aerosols. Healthcare professionals involved in airway management of patients infected with SARS-CoV-2 are at high risk of exposure and subsequent infection, as has been observed in previous coronavirus epidemics.1  This risk is most pronounced in aerosol-generating procedures such as intubation. On March 22, 2020, the American Society of Anesthesiologists (Schaumburg, Illinois), in partnership with other professional organizations, offered guidance for the use of personal protective equipment that have been interpreted by some providers as recommending the use of airborne precautions for all aerosol-generating procedures during the pandemic.2 
Healthcare systems operating under pandemic conditions may need to balance the protection of staff with the allocation of scarce resources, including personal protective equipment. One strategy to address this problem relies on preprocedural testing of asymptomatic individuals. Recent publication of data suggesting imperfect clinical sensitivity of reverse transcription polymerase chain reaction assays for SARS-CoV-23  could lead healthcare providers to intuitively question the wisdom of a strategy that relies on a negative SARS-CoV-2 test, particularly when planning high-risk procedures such as endotracheal intubation. Knowledge of test characteristics, however, is insufficient to guide decision making: the prevalence of the disease in the population for which the test is performed has a critical bearing on the information provided by the test. Prevalence estimates are complicated by the fact that they will differ (sometimes substantially) between different locations, may be unavailable or poorly measured, and will be inherently dynamic during a pandemic. These uncertainties may substantially affect the safety of both patients and providers and may impact the utilization of scarce resources such as personal protective equipment.
To help providers and clinical leaders grapple with this dynamic uncertainty, we have developed an online tool (https://covid-airway-npv.info) that enables the user to examine the impact of different assumptions regarding SARS-CoV-2 reverse transcription polymerase chain reaction test characteristics and disease prevalence on the potential risk of provider exposure during airway management. Uncertainty is modeled by asking the user to provide the most likely, minimum, and maximum value of the parameter (here, SARS-CoV-2 testing characteristics and COVID-19 community prevalence), using a Project Evaluation and Review Techniques distribution.4  The Project Evaluation and Review Techniques distribution was initially developed by the U.S. Navy in an effort to add mathematical rigor to the process of complex project planning, and requires users to provide input uncertainty to enable modeling of output uncertainty.5 
To inform an example calculation, we use publicly published data for analytic specificity of the Quest Diagnostics reverse transcription polymerase chain reaction assay (likely 100%, minimum 95%, maximum 100%) and an informed but pessimistic assumption regarding the clinical sensitivity of the reverse transcription polymerase chain reaction assay (likely 90%, minimum 65%, maximum 99%). Estimation of population prevalence is challenging: the minimum in this scenario is based on a recent measurement of the prevalence of reverse transcription polymerase chain reaction positivity among asymptomatic individuals in Iceland (0.6%), while our maximum is based on a recently published estimate among asymptomatic parturients at a major academic center (13.8%).6,7  As is the case with nearly all measurements of disease prevalence, both of these estimates were measured in unique populations at specific points in time. We chose a “most likely” prevalence estimate of 1.0% based on preliminary, unpublished data emerging from various screening programs within our own health system. A screenshot from the calculator’s analysis under these assumptions is depicted in figure 1. A 90% credible interval for negative predictive value is bounded by 0.06% and 1.12%, giving posttest probabilities of disease ranging from 1 in 89 to 1 in 1,636, and centered at 1 in 338. It is worth noting that a provider in Iceland and another in New York City might have very different assumptions regarding the most likely, minimum, and maximum estimates in their hospitals.
Fig. 1.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
Fig. 1.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
×
We believe that in the setting of a pandemic, airway providers and clinical leaders should engage in this type of quantitative risk assessment—even when that assessment is bounded by significant uncertainty—because it yields four important insights. First, when SARS-CoV-2 is demonstrably uncommon in the population presenting for care, the negative predictive value of a SARS-CoV-2 test should provide reassurance to the individual clinician regarding their case-by-case risk of exposure from an asymptomatic and test-negative patient. Second, in a setting with high surgical volumes, the same calculation could paradoxically provide an opposing perspective: a provider exposure risk that potentially ranges from 1 in 89 to 1 in 1,636 indicates that a busy hospital utilizing droplet precautions for intubation of asymptomatic, test-negative patients could be tolerating the exposure of staff members to aerosolized SARS-CoV-2 on a regular basis. The third insight involves periods or regions in which active disease prevalence is known to be high: full airborne precautions should be considered even for test-negative patients, given the risk of exposure during airway management. Fourth, and perhaps most challenging to grapple with: the uncertainty in our estimates of the components that define negative predictive value generates a wide interval of possible risk that must be acknowledged and thoughtfully considered by clinicians and healthcare leaders.
In an emergency, decisions will be made in the absence of definitive data, and these decisions may be harshly judged in the future through the lens of hindsight. A rational, quantitative approach to decision making has the potential to provide a shared understanding of actions taken in the face of uncertainty. In the case of airway management, we believe that decision makers could specify a lower threshold of negative predictive value that would justify the use of universal airborne precautions, irrespective of preoperative test results: given the relatively fixed characteristics of the reverse transcription polymerase chain reaction test, any such change would be driven by estimates of population prevalence. As a specialty, we should engage in a continuous, transparent process of adapting these policies in collaboration with other leaders and stakeholders in the context of new information.
Research Support
Support was provided solely from institutional and/or departmental sources.
Competing Interests
The authors declare no competing interests.
References
Raboud, J, Shigayeva, A, McGeer, A, Bontovics, E, Chapman, M, Gravel, D, Henry, B, Lapinsky, S, Loeb, M, McDonald, LC, Ofner, M, Paton, S, Reynolds, D, Scales, D, Shen, S, Simor, A, Stewart, T, Vearncombe, M, Zoutman, D, Green, K . Risk factors for SARS transmission from patients requiring intubation: A multicentre investigation in Toronto, Canada. PLoS One 2010; 5:e10717 [Article] [PubMed]
UPDATE: The Use of Personal Protective Equipment by Anesthesia Professionals during the COVID-19 Pandemic. Available at: https://www.asahq.org/about-asa/newsroom/news-releases/2020/03/update-the-use-of-personal-protective-equipment-by-anesthesia-professionals-during-the-covid-19-pandemic.
Wang, W, Xu, Y, Gao, R, Lu, R, Han, K, Wu, G, Tan, W. Detection of SARS-CoV-2 in different types of clinical specimens. JAMA 2020doi: 10.1001/jama.2020.3786. [Epub ahead of print]
Pouillot, R. Tools for Two-Dimensional Monte-Carlo Simulations [R package mc2d version 0.1-18]. Available at: https://CRAN.R-project.org/package=mc2d.
Clark, CE . Letter to the editor—The PERT model for the distribution of an activity time. Oper Res 1962; 10:405–6 [Article]
Gudbjartsson, DF, Helgason, A, Jonsson, H, Magnusson, OT, Melsted, P, Norddahl, GL, Saemundsdottir, J, Sigurdsson, A, Sulem, P, Agustsdottir, AB, Eiriksdottir, B, Fridriksdottir, R, Gardarsdottir, EE, Georgsson, G, Gretarsdottir, OS, Gudmundsson, KR, Gunnarsdottir, TR, Gylfason, A, Holm, H, Jensson, BO, Jonasdottir, A, Jonsson, F, Josefsdottir, KS, Kristjansson, T, Magnusdottir, DN, le Roux, L, Sigmundsdottir, G, Sveinbjornsson, G, Sveinsdottir, KE, Sveinsdottir, M, Thorarensen, EA, Thorbjornsson, B, Löve, A, Masson, G, Jonsdottir, I, Möller, AD, Gudnason, T, Kristinsson, KG, Thorsteinsdottir, U, Stefansson, K. Spread of SARS-CoV-2 in the Icelandic population. N Engl J Med 2020 doi: 10.1056/NEJMoa2006100. [Epub ahead of print]
Sutton, D, Fuchs, K, D’Alton, M, Goffman, D. Universal screening for SARS-CoV-2 in women admitted for delivery. N Engl J Med 2020 doi: 10.1056/NEJMc2009316. [Epub ahead of print]
Fig. 1.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
Fig. 1.
Screenshot from https://covid-airway-npv.info calculator. Negative predictive values and posttest probability of SARS-CoV-2 infection under specific assumptions of sensitivity, specificity, and prevalence. IQR, interquartile range.
×