Diagnostic reliability of the FinSpine nationwide spine surgery registry: Comparison between clinical registry diagnoses and blinded imaging assessments

Introduction

The Finnish national spine surgery registry (FinSpine) collects data on patients, operative procedures, and outcomes.

Research question

Is the reliability of diagnosis collection for the FinSpine registry across various institutions sufficient?

Methods and methods

A random sample of 110 spine surgeries performed in Finland since 2017 was selected from the FinSpine registry. Details of the operative procedures, including characteristics of the hospitals, primary and secondary diagnoses, the anatomical level of the spine surgery, and the date of the surgery, were recorded. Three experienced spine surgeons assessed radiographic imaging independently on two occasions with a one-month interval. Cohen and Fleiss kappa values were calculated for inter- and intra-rater reliability with registry diagnoses as the standard.

Results

Altogether, 85 patients were included in the final analysis. The intra-rater reliability for primary diagnosis was moderate to almost perfect among all three evaluators, with Cohen's kappa values ranging from 0.765 to 0.847. In the inter-rater reliability analysis, there was moderate agreement between evaluators 1 and 2 (Cohen's κ = 0.592, SD 0.061) and evaluators 1 and 3 (κ = 0.595, SD 0.061). The agreement was substantial between evaluators 2 and 3 (κ = 0.676, SD 0.057). The Fleiss' kappa value was 0.622 (95 % CI from 0.570 to 0.673). The inter-rater reliability ranged from 0.596 to 0.652 when the evaluators' diagnoses were compared with those in the registry.

Discussion and conclusion

The primary diagnoses recorded in the FinSpine registry are reproducible and reliable, making the registry data valuable for further research and clinical decision-making.