To the Editor—A recent large study have shown that only 3% of healthcare workers (HCWs) infected with the novel severe acute respiratory coronavirus virus 2 (SARS-CoV-2) have been exposed to an inpatient with coronavirus disease 2019 (COVID-19) prior to the onset of symptoms, suggesting that most HCWs acquire infection in the community, perhaps due to contact with presymptomatic or asymptomatic carriers, rather than in hospital settings.^{Reference Kluytmans, Buiting and Pas1} The high prevalence of infection (11%–20%) among HCWs supports this assertion,^{Reference Tostmann, Bradley and Bousema2,Reference Lo3} despite the reportedly low risk of nosocomial infection associated with SARS-CoV-2.^{Reference Kluytmans, Buiting and Pas1,Reference Cheng, Wong and Chen4,Reference Ng, Poon and Kiat Puar5}

Wee et al^{Reference Wee, Sim and Conceicao6} recently reported their experience with syndromic surveillance of HCWs for COVID-19. The surveillance was based on symptoms of acute respiratory illness (ARI) and fever. Pointing to community as well as in-hospital secondary transmission, these researchers detected 14 cases of COVID-19 among HCWs with 4 clusters; most were linked via transmission outside of hospital.^{Reference Wee, Sim and Conceicao6} However, although syndromic surveillance, based on fever and ARI symptoms, was shown to be effective during the outbreak of severe acute respiratory syndrome (SARS),^{Reference Tan7} it may not be as effective in containment of COVID-19 infection because of different patterns of clinical presentations. Indeed, in the study by Wee et al,^{Reference Wee, Sim and Conceicao6} heat maps did not pick up the cluster with suspected intrahospital spread. Moreover, in a study by Tostmann et al^{Reference Tostmann, Bradley and Bousema2} employing a low-threshold screening for SARS-CoV-2, most HCWs with SARS-CoV-2 infection had mild clinical presentations, frequently not including fever or respiratory symptoms. Using the same clinical data, these researchers developed a model, excluding fever and cough, to predict COVID-19 among HCWs with a fair discriminative ability.^{Reference Tostmann, Bradley and Bousema2}

Early identification and control of COVID-19 among HCWs is of paramount importance particularly in the postoutbreak period to prevent in-hospital secondary transmission to other HCWs and inpatients. The concern of transmitting infection into the healthcare system has been highlighted in a recent study in which HCWs were linked to transmission of COVID-19 into long-term healthcare facilities.^{Reference McMichael, Currie and Clark8} Considering that as many as 50% of all SARS-CoV-2 infections are asymptomatic,^{Reference Graham, Maldonado, Tompkins, Wald, Chawla and Hawn9} it would seem appropriate, when resources are available, to perform routine SARS-CoV-2 nasopharyngeal screening for all HCWs. The prediction model described by Tostmann et al,^{Reference Tostmann, Bradley and Bousema2} rather than fever and ARI symptoms, can be used to guide a targeted screening strategy in settings with limited availability of testing materials.