Introduction Several cross-sectional studies indicate that occupations such as physicians, nurses, dentists, physiotherapists, bus- and taxi drivers, bartenders, waiters, chefs and shop assistants either have higher risks of dying or be tested positive due to COVID-19.

Objectives The aim of this study was to analyse if employees in occupations characterized by working close to other people and/or being more exposed to diseases or infections are treated for inpatient care due to COVID-19 more often than other employees.

Methods Survey data regarding proximity to other people and to what extent individuals are exposed to diseases and infections were retreived from https://onet.rti.org/about.cfm. Analyses werebased on a prospective cohort comprising 1 037 811 residentsin Region Stockholm March 1, 2020, 18–64 years old and having a valid occupational code. The follow-up period was nine months. Quartiles were created, based on averages for each occupation regarding these questions. Adjusting factors were age, country of birth, daily smoking, overcrowding, household size and percentages in demographic areas being test positive to COVID-19. We calculated both IRR’s and SMR’s.

Results During the period, 2984 individuals received inpatient care for COVID-19. Higher risks were observed in the highest exposed quartiles after adjusting for confounders. These associations were found among both sexes, in all age groups and in each 3-month calender period with IRR’s ranging from 1.68 to 2.58 compared to the least exposed quartile. SMR’s calculated for each occupation revealed higher risks among e.g. physicians, nurses, assistant nurses, care assistants and traffic instructors with SMR’s ranging from 138 to 272.

Conclusions A strength of this study is the use of inpatient care as an outcome, and adjustment for individual confounders. The observation that the risk from close occupational contacts was not reduced over the observation period highlights the importance of further improvement of risk assessment and protective measures.