Correct probability estimates for risk of anaesthetist contracting severe acute respiratory virus coronavirus 2 after aerosol-generating procedures: In reply

Professor Pandit has misunderstood three aspects of my model.

First, my model¹ does not apply to coronavirus disease 2019 (COVID-19)–negative patients, but rather to ‘patients from the general population with no indication for testing’. In my home state (Queensland), asymptomatic patients currently do not isolate for 14 days, nor do they undergo polymerase chain reaction antigen testing (to look for the presence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in swabs obtained from them) prior to surgery. Such patients might be considered COVID-19 unknown according to Pandit’s model. ² However, the corresponding calculations would not apply because they rely on data from an entirely different population. At the time of writing, in the past week, my home state of Queensland (population around five million) has processed 101,301 tests and recorded one new case, ³ whereas the UK (population around 67 million) has processed 1,193,413 tests and recorded 6141 new cases. ⁴

Second, my ‘ballpark estimate’ for transmissibility of 22% (around one in five) in the COVID-19-unknown population cannot be extrapolated to the COVID-19-positive population. Patients from the intubateCovid registry were known (or suspected) to have COVID-19, and therefore intubated using airborne precautions—often with personal protective equipment (PPE) that exceeded the World Health Organization recommended minimum standards (e.g. FFP3/N100 masks, powered air-purifying respirators). ⁵ My one-in-five transmissibility estimate refers to the protective effect of ‘any respiratory PPE (versus no PPE) during the SARS-CoV-1 outbreak of 2003’, where often the only respiratory PPE worn was a flat surgical mask. ⁶

My one-in-five transmissibility estimate applies to an (undiagnosed) asymptomatic or pre-symptomatic carrier of SARS-CoV-2 being cared for by an anaesthetist wearing a flat surgical mask (i.e. contact precautions or droplet precautions). That is the PPE recommendation currently endorsed by the Australian and New Zealand College of Anaesthetists when patients are deemed to be at low risk of SARS-CoV-2 and when the risk of community transmission is low or moderate. ⁷ That is why we need a model that can estimate the risk of this approach at the local level, modify this estimate as community transmission fluctuates and communicate these concepts effectively.

Third, prevalence counts known carriers of SARS-CoV-2. If such patients were to undergo anaesthesia, their known COVID-19-positive status would trigger the use of airborne precautions, so my model would not apply. The risk of transmissibility in my model is only posed by asymptomatic or pre-symptomatic carriers of SARS-CoV-2 who are undergoing anaesthesia without first being tested. That is why my calculation for incidence is more accurate than prevalence for estimating the probability that an asymptomatic patient (outside of the known COVID-19-positive population) will be a carrier of SARS-CoV-2.