Purpose
The purpose of this paper is to analyze how specialization in hospitals affects operational performance, measured by the length of stay and readmission rate. The authors assess a public policy change in the Danish healthcare sector from 2011 which required that some hospital services had to be centralized leading to specialization within the merged departments.

Design/methodology/approach
Taking an institutional theory perspective, the authors conduct a natural experiment. The data include 24,694 observations of urological patient treatments from 2010 to 2012.

Findings
The econometric difference-in-difference analysis finds that the readmission rate decreases by approximately four percentage points in the departments affected by the policy change. Contrary to expectations, the length of stay increases by 0.38 days. The authors complement the natural experiment with a mixed-methods approach that includes proprietary data from the management control system of the hospital, public documentation on the policy change, as well as interviews with key informants. These data suggest that operational deficiency is related to the fact that specialization was externally enforced through the public policy change. The authors illustrate how the hospital staff struggle for legitimacy after this policy change, and how cost savings obstructed the specialized department in achieving its goals.

Originality/value
The authors conclude that the usual economies-of-scales-based logic of (higher)volume-(better)outcome studies cannot easily be transferred to specialization in hospitals, unless one accounts for the institutional reason of the specialization.