Background The aims of this study were: (1) to explore the frequency of discrepancies in work accommodations reported by workers and their supervisors, and (2) to investigate whether these discrepancies are associated with full return to work (RTW). Methods We used data from a longitudinal survey study of long-term sick-listed workers and their supervisors (n = 406). Discrepancies in reports on implementing eight types of work accommodations were explored. Logistic regression analyses were conducted to test associations between discrepancies in reported work accommodations and odds of full RTW 27 months after the sick-leave onset. Results Discrepancies were the lowest for the work accommodation therapeutic RTW (53%) and the highest (85%) for job training or education and reimbursement of therapy or treatment. Four out of eight types of work accommodations were more often reported by workers than by their supervisors. Only a discrepancy on a job reassignment within the organization was associated with lower odds of full RTW (OR 0.56, 95%-CI 0.36–0.88). Conclusion We found substantial discrepancies in the reported implementation of work accommodations between workers and their supervisors. Future research should focus on disentangling mechanisms that lead to discrepancies to avoid inefficiencies in the RTW process.