Predictive factors for hospitalization in a cohort of primary healthcare patients with suspected COVID-19

BACKGROUND: The objective of the study was to identify clinical and demographic factors predictive of hospitalization in primary healthcare patients diagnosed with suspected COVID-19 at the beginning of the pandemic. METHODS: A retrospective cohort study design was used. Patients attended in Casanova primary healthcare centre (CAP) (Barcelona, Spain) for symptoms compatible with possible or confirmed SARS-CoV-2 infection between February 24 and May 30, 2020, were included. Data was collected through the electronic medical record and by telephone interview. RESULTS: 518 patients were included, of whom 283 (54.6%) were female. The median age was 50.2 years and 19.3% were aged ≥ 65 years: 79% were followed on an outpatient basis while the rest were hospitalized. Predictive factors for hospital admission were male sex, older age, a history of ischemic heart disease and the presence of dyspnoea, haemoptysis, nausea and vomiting, with a sensitivity of 48% and a specificity of 95.4%. Odynophagia and nasal congestion were predictors of a good prognosis. Mortality was 2.3% and 25% of deaths did not occur in hospital. CONCLUSIONS: Nearly 80% of primary healthcare patients received only outpatient care. Male sex, older age, a history of ischemic heart disease and symptoms like dyspnoea, haemoptysis, nausea and vomiting could lead to a greater risk of an unfavorable evolution during COVID-19. Patients with at least one of the above factors, which correlate with a higher hospital admission rate, should receive a closer follow-up to early detect when they can benefit from a hospital evaluation based on their clinical evolution.