Objective: To analyze the associations between favorable physical and psychosocial work factors and health behavior among healthcare employees (nurses and care assistants) with health complaints.

Methods: The study was based on seven iterations (2001-2013) of a biennial Swedish work environment survey linked with data from public registers. In all, 7,180 healthcare employees, aged 16-64 years, who had reported health complaints, were included. Health behavior was operationalized through four combinations of sickness absence (SA) and sickness presence (SP): 'good health behavior' (Low SP/Low SA), 'recovery behavior' (Low SP/High SA), 'risk behavior' (High SP/Low SA), and 'poor health behavior' (High SP/High SA). Odds ratios (OR) were calculated by multinomial logistic regression with 95% confidence intervals (CI).

Results: After adjusting for socio-demographic factors, those who rarely worked in strenuous postures had an increased probability of having 'good health behavior' (OR range: nurses 1.72-2.02; care assistants 1.46-1.75). Those who rarely experienced high job demands had increased odds for having 'good health behavior' (OR: nurses 1.81; OR range: care assistants 1.67-2.13), while having good job control was found to be related to 'good health behavior' only among care assistants (OR range 1.30-1.68). In the full model, after also considering differences in health, none of the work environment indicators affected 'good health behavior' among nursing professionals. Among care assistants, rarely having heavy physical work and having low psychosocial demands remained significantly associated with 'good health behavior' (OR range: 1.24-1.58) and 'recovery behavior' (OR range: 1.33-1.70). No associations were found between favorable work environment factors and 'risk behavior' among the two groups of employees. However, positive assessments of the work situation were associated with 'good health behavior,' even after controlling for all confounders for both groups (OR range: 1.43-2.69).

Conclusions: 'Good health behavior' and 'recovery behavior' among care assistants were associated with favorable physical and psychosocial working conditions even when health was considered. This implies that reduced sickness presence and sickness absence among care assistants can be achieved through improved physical and psychosocial working conditions.