Limitations of fracture classification systems

1
Limitations of fracture classification systems

• Reliable and valid
• Reliability
• Interobserver reliability
• Same case vs. Multiple observers
• Intraobserver reproducibility
• Same observer vs. Multiple occasions
• Validity
• Accuracy
• Gold standard

2
Accuracy vs. Agreement

• Accuracy
• True or false
• Agreement
• Optimal answer

3

• Albert Burstein
• Do thy work and are they useful?
• Substantial interobserver variability
• Kappa statistic compare 2 groups
• 1.0 perfect agreement
• 0.0 chance agreement
• -1.0 perfect disagreement
• Weighted kappa statistic weighting scheme

4
Kappa value categorization

• Landis and Koch
• 0.00 poor reliability
• 0.010.20 slight
• 0.210.40 fiar
• 0.410.60 moderate
• 0.610.80 substantial
• 0.811.00 neearly perfect agreement
• Svanholm
• lt0.50 poor
• 0.510.74 moderate
• gt0.75 excellent

5
Observer variability

• Limitation to many fracture classification
  systems
• Fair to poor intraobserver reliability
• Neer
• Garden
• Ruedi and Allgower and AO
• Schatzker and AO
• Lauge-Hansen and Weber
• Gustilo-Anderson
• Other injuries
• Unknown etiology

6
Methodological issues

• Review 44 published studies assessing reliability
• Fracutre radiographs represented the spectrum and
  frequency of injury? 61
• Justify the size of the study group? 100
• Inadequate statistical analysis? 61
• Methodology of validating highly detailed

7
Outcome validation of the AO/OTA fracture
classification system

• Prospective multicenter study, 200 patients with
  unilateral isolated lower-extremity
• Functional outcome measurement at 6 and 12 months
• AO/oTA code
• Not a good predictor

8
Quality of fracture radiographs

• X-ray
• Not a significant source of intraobserver
  variability
• Tibia-plafond
• CT
• No improvement for intraobserver variability
• Neer
• Tibia-plateau
• Useful adjunt for severe articular fracture
• Tibia-plateau
• Distal radius
• Change classification in 57
• MRI improvement of classification in 3 studies

9
Difficulty identifying fracture lines on
radiographs

• Overlapping osseous fragement or densities
• Osteopenia
• Difficult reading
• Not mentioned in classification
• Periarticular fractures small fragment
• Observer difficulty

10
Variability making measurement on radiographs

• Amount of displacement/angulation/articular
  fracture
• Substantial unpredictability
• CT for tibia plateau
• /- 12mm for maximal articular depression
• /- 9mm for maximum condylar widening
• Less tolerance limits when reduced
• CT more accurate than X-ray