BACKGROUND: The SARS-CoV-2 pandemic has caused an unprecedented strain on healthcare systems worldwide, including the United Kingdom National Health Service (NHS). We conducted an observational cohort study of SARS-CoV-2 infection in frontline healthcare workers (HCW) working in an acute NHS Trust during the first wave of the pandemic, to answer emerging questions surrounding SARS-CoV-2 infection, diagnosis, transmission and control. METHODS: Using self-collected weekly saliva and twice weekly combined oropharyngeal/nasopharyngeal (OP/NP) samples, in addition to self-assessed symptom profiles and isolation behaviours, we retrospectively compared SARS-CoV-2 detection by RT-qPCR of saliva and OP/NP samples. We report the association with contemporaneous symptoms and isolation behaviour. RESULTS: Over a 12-week period from 30th March 2020, 40·0% (n = 34/85, 95% confidence interval 31·3-51·8%) HCW had evidence of SARS-CoV-2 infection by surveillance OP/NP swab and/or saliva sample. Symptoms were reported by 47·1% (n = 40) and self-isolation by 25·9% (n = 22) participants. Only 44.1% (n = 15/34) participants with SARS-CoV-2 infection reported any symptoms within 14 days of a positive result and only 29·4% (n = 10/34) reported self-isolation periods. Overall agreement between paired saliva and OP/NP swabs was 93·4% (n = 211/226 pairs) but rates of positive concordance were low. In paired samples with at least one positive result, 35·0% (n = 7/20) were positive exclusively by OP/NP swab, 40·0% (n = 8/20) exclusively by saliva and in only 25·0% (n = 5/20) were the OP/NP and saliva result both positive. CONCLUSIONS: HCW are a potential source of SARS-CoV-2 transmission in hospitals and symptom screening will identify the minority of infections. Without routine asymptomatic SARS-CoV-2 screening, it is likely that HCW with SARS-CoV-2 infection would continue to attend work. Saliva, in addition to OP/NP swab testing, facilitated ascertainment of symptomatic and asymptomatic SARS-CoV-2 infections. Combined saliva and OP/NP swab sampling would improve detection of SARS-CoV-2 for surveillance and is recommended for a high sensitivity strategy.