Background: The objective of this study was to investigate whether work unit-levels of psychosocial working conditions modify the effect of depressive symptoms on risk of long-term sickness absence (LTSA). Methods: A total of 5,416 Danish female eldercare workers from 309 work units were surveyed using questionnaires assessing depressive symptoms and psychosocial working conditions. LTSA was derived from a national register. We aggregated scores of psychosocial working conditions to the work unit-level and conducted multi-level Poisson regression analyses. Results: Depressive symptoms, but not psychosocial working conditions, predicted LTSA. Psychosocial working conditions did not statistically significantly modify the effect of depressive symptoms on LTSA. Conclusions: Psychosocial working conditions did not modify the effect of depressive symptoms on LTSA. The results, however, need to be interpreted with caution, as we cannot rule out lack of exposure contrast and non-differential misclassification of the exposure. Am. J. Ind. Med. 56:1329-1340, 2013.