Interobserver Variability in Interpreting Radiographic Lucencies About Total Hip Reconstructions

Abstract

Radiographs are commonly used to identify the loosened total hip reconstruction (THR), to plan revision surgery, and to evaluate long-term follow-up patients for study purposes. Interobserver variability was studied in an attempt to interpret 63 random follow-up films in a group of 93 patients eight to 12 years post-THR. Nine observations were made on the acetabular and femoral sides of the reconstructions on all films for a total of 567 observations. There was considerable disagreement in observations of the acetabular lucencies; a consensus of all three observers was obtained only 46% of the time. There was more uniform agreement among observers (87%) in analyzing the femoral side, although agreement was most frequent when there was no lucency. There was no difference in observer variables when greater lucencies were reported by one or more observers than when there were smaller lucencies. It was concluded that a single observer can reliably record a femoral observation of no lucency and, making allowances for the Mach effect, can reliably distinguish the presence or absence of a lucency. However, a single observer may not be able to make observations of acetabular lucencies with great reliability and, in contrast to femoral lucencies, a single observer may be less reliable in recording acetabular lucencies over 2 mm as compared with those under 2 mm.