Objectives The objectives of this study is to describe a method for monitoring unwarranted variation in procedure-based treatment in specialized healthcare focusing on high-cost areas.

Method We extracted anonymized data for surgical procedures from the Norwegian Patient Registry performed in hospitals with public reimbursement in the period 2017–2021 on episode level. Based on the Nomesco Classification of Surgical Procedures and selected radiological procedures, the procedures were classified into a four level hierarchical system. In order to estimate cost, all procedures within an episode where sorted according to diagnosis related group (DRG) weight. We selected the procedures with the highest weight for further analysis. Thus, only one procedure per episode where analyzed. We calculated standardized rates for each municipality for the procedures, and obtained central tendency measures with confidence intervals at the local health service (HSA) and regional health area (RHA) in addition to national level. We explored different methods to identify outliers to gain further understanding of appearing unwarranted variation.

Results We present utilization rates for the 20 categories of procedures associated with the highest cost. Joint replacement was the category associated with highest total cost in Norway. Only minor differences in utilization rates were observed at local, regional and national level. This may indicate standardized indications for surgery in patients with hip and knee arthritis in Norway. Other categories of procedures, such as lower endoscopic procedures and arthroscopy, revealed more profound variation in utilization rates at both local and regional levels. Our data indicate equal distribution of utilization rates for surgical procedures categorized as ‘effective care’ (according to Wennberg) like surgery for colo-rectal cancer. However, our analyses including many procedures perceived as ‘preference and supply sensitive’, revealed significant geographical variation in utility rates. Moreover, no associations between utilization rates for colonoscopy and five-years mortality for colo-rectal cancer on municipality level were found.

Conclusions Equal health services are a national goal in Norway. This study describes a tool allowing health care providers to assess geographical variation in utilization rates for various procedures. Our results show that the utilization rates for selected preference and supply sensitive procedures varies significantly both between hospital areas and regional health trusts. In addition, we found no associations between high utilization rates and outcome on population level. This tool may offer hospitals, health authorities and politicians important insight in health care distribution.