Background: Early in the COVID-19 pandemic, GPs had to distinguish SARS-CoV-2 from other aetiologies in patients presenting with respiratory tract infection (RTI) symptoms on clinical grounds and adapt management accordingly. Objectives: To test the diagnostic accuracy of GPs’ clinical diagnosis of a SARS-CoV-2 infection in a period when COVID-19 was a new disease. To describe GPs’ management of patients presenting with RTI for whom no confirmed diagnosis was available. To investigate associations between patient and clinical features with a SARS-CoV-2 infection. Methods: In April 2020–March 2021, 876 patients (9 countries) were recruited when they contacted their GP with symptoms of an RTI of unknown aetiology. A swab was taken at baseline for later analysis. Aetiology (PCR), diagnostic accuracy of GPs’ clinical SARS-CoV-2 diagnosis, and patient management were explored. Factors related to SARS-CoV-2 infection were determined by logistic regression modelling. Results: GPs suspected SARS-CoV-2 in 53% of patients whereas 27% of patients tested positive for SARS-CoV-2. True-positive patients (23%) were more intensively managed for follow-up, antiviral prescribing and advice than true-negatives (42%). False negatives (5%) were under-advised, particularly for social distancing and isolation. Older age (OR: 1.02 (1.01–1.03)), male sex (OR: 1.68 (1.16–2.41)), loss of taste/smell (OR: 5.8 (3.7–9)), fever (OR: 1.9 (1.3–2.8)), muscle aches (OR: 2.1 (1.5–3)), and a known risk factor for COVID-19 (travel, health care worker, contact with proven case; OR: 2.7 (1.8–4)) were predictive of SARS-CoV-2 infection. Absence of loss of taste/smell, fever, muscle aches and a known risk factor for COVID-19 correctly excluded SARS-CoV-2 in 92.3% of patients, whereas presence of 3, or 4 of these variables correctly classified SARS-CoV-2 in 57.7% and 87.1%. Conclusion: Correct clinical diagnosis of SARS-CoV-2 infection, without POC-testing available, appeared to be complicated.