The impact of introducing a new hospital financing system (DRGs) in Poland on hospitalisations for atherosclerosis: An interrupted time series analysis (2004–2012)

•Hospital payment based on DRGs was introduced in Poland in July 2008.•Following DRGs implementation, a sharp increase in hospitalisations for atherosclerosis was observed.•Non-public hospitals did not exhibit a rapid response to DRGs introduction.•Changes in the hospitalisation rates may be the result of DRGs introduction.•Failure to account for these changes may result in incorrect epidemiological data. Hospital payment based on diagnosis-related groups (DRGs) was introduced in Poland in July 2008. We evaluate the impact of this policy on the frequency of hospitalisation for atherosclerosis in internal medicine units of district hospitals and non-public hospitals in Poland. Data were collected between 2004 and 2012 from each district and non-public hospital participating in the General Hospital Morbidity Study (165 hospitals in total). Atherosclerosis was defined using the ICD-10 code I70. Hospitalisation patterns were examined using interrupted time series with segmented regression analysis. Results were compared between public and non-public hospitals and across patient age groups. The rate of hospitalisation for atherosclerosis rose by 27.05 per 10,000 total hospitalisations immediately following the implementation of DRGs in 2008. It then rose by 2.5 per 10,000 hospitalisations monthly between 2008–2012. The largest changes were observed for patients aged 85+ and 75–84. Rates rose by 117.5 and 54.2 per 10,000 hospitalisations in these two groups respectively following implementation of DRGs. The response to introduction of DRGs was less striking in non-public hospitals than in public hospitals. Implementation of a DRG-based system in Poland was associated with substantial increases in atherosclerosis hospitalisation rates. Failing to take into account this change in financing and not accounting for long-term trends in hospitalisation rates may result in inaccurate epidemiological data.