Interobserver variation in the AO/OTA fracture classification system for pilon fractures: is there a problem?

Area of Science:

• Orthopaedic surgery
• Radiology
• Trauma care

Background:

• The AO/OTA fracture classification system is widely used for pilon fractures.
• Accurate classification is crucial for treatment planning and research.
• Interobserver reliability of classification systems can vary.

Purpose of the Study:

• To assess the interobserver variation in applying the AO/OTA fracture classification system to pilon fractures (region 43).
• To determine the reliability of classifying fracture type, group, and subgroup.

Main Methods:

• Eighty-four sets of patient radiographs were independently reviewed by a senior attending, two fellows, a resident, and a research coordinator.
• Evaluators were blinded to treatment and outcomes.
• Statistical analysis included the kappa statistic, Williams index, and SAV statistic to measure agreement.

Main Results:

• The average chance-adjusted agreement (SAV statistic) was moderate: 0.57 for fracture type, 0.43 for group, and 0.41 for subgroup.
• Kappa statistics for fracture types A, B, and C were 0.49, 0.58, and 0.57, respectively, indicating adequate agreement.
• Agreement was generally moderate across different classification levels.

Conclusions:

• Interobserver agreement for AO/OTA pilon fracture classification is comparable to other systems.
• The classification system's reliance on subjective judgment for continuous fracture patterns is noted.
• Reporting only fracture types (A, B, C) may be adequate for clinical research purposes.