Table 1. Combined data for SARS-CoV-2 RNA positive HCWs by role and screening arm, from the present study and our previous study¹. Difference in proportions of HCWs testing positive through the symptomatic screening arm was analysed using Pearson’s chi-square test.

Overall, 360 individuals underwent repeat testing, either as part of the asymptomatic screening programme, or for other reasons as previously described.¹ Median turnaround time from sample arrival in the laboratory to final verification was 18 hours 45 mins. Positive results were called out on the same day, with negative results emailed within 24 hours.

Between 25^th April 2020 and 24^th May 2020, a total of 34 new positive tests were reported. In the HCW symptomatic and HCW symptomatic household contact screening armscombined (reflecting all individuals with self-reported symptoms at the time of testing), 13/771 (1.7%) tests were positive, which was significantly lower than 30/221(13%) in the original study period (Fisher’s exact test p<0.0001). In the HCW asymptomatic screening arm, 21/2,611 (0.8%) tests were positive, which again was significantly lower than 31/1,032 (3%) in the original study period (Fisher’s exact test p<0.0001). As we previously observed¹, individuals captured in the HCW asymptomatic screening arm were generally asymptomatic at the time of screening, however these individuals could be divided into sub-groups. In the first subgroup, 8/21 (38%) HCWs had no symptoms at all. Of these, 5/8 (63%) remained entirely asymptomatic 5-7 weeks after their positive test, whereas 2/8 (25%) developed symptoms 24 – 48 h after testing. One HCW could not be contacted to obtain further history. In the second subgroup, 6/21 (29%) had retrospectively experienced some symptoms prior to screening. Of these, 5/6 (83%) had symptoms with a high pre-test probability of COVID-19¹ commencing >7 days prior to screening, of whom 3/5 had appropriately self-isolated then returned to work, and 1/5 was tested shortly after developing symptoms. 1/6 (17%) had symptoms with a low pre-test probability of COVID-19¹ commencing <7 days prior to screening and had not self-isolated. In the third subgroup, 7/21 (33%) were detected through repeat sampling of HCW who previously tested positive. Of these, 4/7 (57%) were tested to determine their suitability to return to work with severely immunocompromised / immunosuppressed patients, as dictated by UK national guidance.⁶ The remaining 3/7 (43%) were from HCWs tested incidentally for a second time in the asymptomatic HCW screening programme. The median interval between serial positive tests was 16.5 days (IQR 9.5-19.5). All cases were attributable to prolonged SARS-CoV-2 RNA detection from a single infection, rather than re-infection. Our approach to patients with repeatedly positive SARS-CoV-2 PCR tests is described in the Methods.

The fraction of positive tests amongst the HCW asymptomatic, and HCW symptomatic and household contact screening groups combined varied over time (Figure 1A, Table 2). In particular, during the last two weeks of the study period (11^th May to 24^th May) we identified only 4 positive SARS-CoV-2 samples from 2,016 tests performed, 2 from the HCW asymptomaticand 2 from the HCW symptomatic / symptomatic household contact arms . This fall in positive HCW tests mirrored the decline in both patients testing positive at CUHNFT and those tested throughout the wider region (Figure 1B). Similar trends were observed in a smaller cohort study of HCWs in London.⁷