Related Experiment Videos
Interobserver variation using the AO/ASIF classification of long bone fractures
D J Johnstone1, W J Radford, E J Parnell
1Charing Cross Hospital, London, UK.
Injury
|March 1, 1993
Summary
The AO/ASIF long bone fracture classification system showed low agreement among surgeons. Even with training, coding accuracy was poor, suggesting consensus is needed for research and audits.
Area of Science:
- Orthopaedic Surgery
- Medical Classification Systems
Background:
- The AO/ASIF Foundation developed a comprehensive classification system for long bone fractures.
- Accurate fracture classification is crucial for research, treatment planning, and auditing.
Purpose of the Study:
- To evaluate the reliability and accuracy of the AO/ASIF long bone fracture classification system among orthopaedic surgeons.
- To assess the impact of prior experience on classification accuracy.
Main Methods:
- 18 orthopaedic surgeons classified 10 long bone fractures using the AO system after training.
- Individual classifications were compared to a derived consensus classification.
- Agreement rates between individual and consensus classifications were analyzed.
Main Results:
- Only 32% of all individual fracture classifications agreed with the final consensus.
- Surgeons with prior experience showed no significant improvement in accuracy compared to novices.
- High rates of coding inaccuracy were observed regardless of experience level.
Conclusions:
- The AO/ASIF long bone fracture classification system demonstrates significant inter-observer variability.
- Consensus-based classification is recommended for research and computer-based audits utilizing the AO system.
- Further training or system refinement may be necessary to improve classification consistency.