Introduction Sickness absenteeism is a complex and multifactorial phenomenon. The present paper studies the association between variables of different levels of determination and short/long sickness absenteeism reported by nursing staff.

Methods Sectional study with 1509 workers (89.4% of all workers) from three public hospitals in Rio de Janeiro, Brazil. Sickness absence was assessed by means of the self-reported question: ‘‘How many whole days have you been off work due to health problems (diseases or healthcare or for examination) during the past year (12 months)?” and we defined short sickness absenteeism (1–9 days) and long (≥10 days). The logistic regression analysis used a conceptual model that consider distal (socio-economic conditions), intermediate I (occupational conditions), intermediate II (lifestyle characteristics) and proximal determinants (diseases and health conditions).

Results The prevalence of short and long absenteeism were respectively 20.3% and 16.6%. In multivariate analysis, both types of absenteeism were associated with the role of practical nurse, musculoskeletal diseases, self-perceived poor/regular health and presence of minor psychological distress. In addition, short absenteeism was also associated with age and having more than one job, while long absenteeism was related to marital status and being a public servant.

Conclusion The results indicated variables of different levels of determination were linked to sickness absenteeism. Sickness absenteeism seems to indicate a multiple and complex determination that cannot be explained exclusively by health conditions.