Data quality in the Norwegian National Quality Registry for Pain Management

The Norwegian National Quality Registry for Pain Management (SmerteReg) is used for quality improvement, research and healthcare management. High data quality is therefore essential. This study investigates the accuracy and inter-rater reliability (IRR) of a selection of key registry variables.

Accuracy was assessed by reviewing the medical records of a random sample of patient cases from 10 pain management teams in the period 2023-2024. Observed accuracy was calculated using the Wilson score interval. IRR was assessed using four fictitious case histories registered by 15 data abstractors during the registry's 2022 annual meeting, and analysed using Gwet's AC.

Of the 300 patient cases selected, 217 were included in the analysis. Overall observed accuracy was 69 % (95 % CI 68-70 %), ranging from 29 % to 99 % across the different variables. Thirteen of 22 variables relating to quality indicators demonstrated moderate to very good accuracy. The IRR analysis showed that 22 of the 27 variables had an agreement coefficient (Gwet's AC) of 0.60 or higher. Five variables had an agreement coefficient below 0.60. These related to maximum and minimum pain at rest and during movement at the final assessment, and whether an opioid tapering plan was documented in the medical record.

The Norwegian National Quality Registry for Pain Management demonstrated moderate to high data quality for most of the variables analysed. Data quality was particularly high for variables that could be directly extracted from the medical records, whereas variables requiring interpretation showed lower quality. The study has already led to improvements in the registry's set of variables and user guidance.