There is currently no standardised complication grading classification routinely used for paediatric orthopaedic surgical procedures. The Clavien-Dindo classification used in general surgery was modified and validated in 2011 by Sink et al. and has been used regularly to classify complications following hip preservation surgery. The aim of this study was to adapt and validate Sink et al.'s modification of the Clavien-Dindo classification system for grading complications following surgical interventions of the upper and lower extremities and spine in paediatric orthopaedic patients.

Sink et al.'s modification of the Clavien-Dindo classification system was further modified for paediatric orthopaedic procedures. The modified grading scheme was based on the treatment required to treat the complication and the long term morbidity of the complication. Grade I complications do not require deviation from standard treatment. Grade II complications deviate from the normal post-operative course and require outpatient treatment. Grade III complications require investigations, re-admission or re-operation. Grade IV complications are limb or life threatening or have a potential for permanent disability (IVa: with no long term disability and IVb: with long-term disability). Grade V complications result in death. Forty-five complication scenarios were developed. Seven paediatric orthopaedic surgeons were trained to use the modified system and they each graded the scenarios on two occasions. The scenarios were presented in a different random order each time they were graded. Fleiss' and Cohen's k statistics were performed to test for inter-rater and intra-rater reliabilities, respectively.

The overall Fleiss' k value for inter-rater reliability was 0.772 (95% CI, 0.744–0.799). The weighted k was 0.765 (95% CI, 0.703–0.826) for Grade I, 0.692 (95% CI, 0.630–0.753) for Grade II, 0.733 (95% CI, 0.671–0.795) for Grade III, 0.657(95% CI, 0.595–0.719) for Grade IVa, 0.769 (95% CI, 0.707–0.83) for Grade IVb and 1.000 for Grade V (p value <0.001). The Cohen's k value for intra-rater reliability was 0.918 (95% CI, 0.887–0.947). These tests show that the adapted classification system has high inter- and intra-rater reliabilities for grading complications following paediatric orthopaedic surgery.

Given the high intra- and inter-rater reliability and simplicity of this system, adoption of this grading scheme as a standard of reporting complications in paediatric orthopaedic surgery could be considered. Since the evaluation of surgical outcomes should include the ability to reliably grade surgical complications, this reproducible, reliable system to assess paediatric surgical complications will be a valuable tool for improving surgical practices and patient outcomes.