Reliability and validity of plain radiographs to assess angulation of small finger metacarpal neck fractures: Human cadaveric study

To quantify reliability and validity of plain radiographs for assessing the degree of small finger metacarpal neck fracture angulation, we created typical two-fragments fractures in 30 adult cadaveric specimens. Reliability and validity of different radiographic measurement methods were determined by the intraclass correlation coefficient (ICC) and the Bland and Altman graphical approach. Intraobserver and interobserver reliability was high with any radiographic measurement method. Mean ICCs values (95% confidence intervals) varied from 0.76 (0.56-0.88) to 1.00 (0.99-1.00). The graphical approach confirmed good agreement. Validity was substantial when the fracture angle was measured between the line along the longitudinal axis of the metacarpal shaft and the line from the center of the metacarpal head to the fracture site on lateral radiographs. Mean ICCs values varied from 0.70 (0.36-0.86) to 0.79 (0.5-0.90). The graphical analysis also indicated good agreement. In contrast, considerable lack of validity was observed when the angle was measured on oblique radiographs. Although the mean ICCs values varied from 0.68 (0.12-0.88) to 0.74 (0.05-0.90), suggesting substantial correlation, the graphical analysis provided evidence for poor validity. There was systematic bias with oblique radiographs consistently producing higher readings (up to 35 degrees ). In summary, reliability and validity are good only when the degree of small finger metacarpal neck fracture angulation is measured after drawing lines on lateral radiographs. Oblique radiograph measurements consistently produce higher readings.