Measurement Error Versus Repeated Measurements: A Guide Describing Two Methods for Computing Bias and Precision of Migration Measurements From Double Examinations Using Radiostereometric Analysis.

Abstract

Radiostereometric analysis is a method to measure implant migration where an ISO standard recommends double examinations (i.e., acquisition of two independent sets of biplanar images on the same day) to compute bias (i.e., mean) and precision (i.e., standard deviation) of differences in repeated migration measurements (termed repeated measurement statistics). However, repeated measurement statistics do not provide information regarding trueness of the measurements. Double examinations also can be used to compute measurement error statistics (i.e., mean and standard deviation of migration measurements relative to trueness). Our objectives were to derive measurement error and repeated measurement population parameters in six degrees of freedom (6DOF) and in maximum total point motion (MTPM), demonstrate quantitative differences by computing measurement error and repeated measurement statistics from a clinical study for an example implant, and demonstrate the importance of determining mean measurement error in MTPM. Three key findings were: (1) in 6DOF, measurement error and repeated measurement statistics were nearly identical; (2) for MTPM, measurement error and repeated measurement statistics had different means of 0.21 mm and 0.00 mm, respectively, but similar standard deviations; and (3) mean measurement error in MTPM is important for drawing conclusions about early implant stability. Because measurement error statistics are the same as repeated measurement statistics in 6DOF but provide additional information in the form of mean measurement error in MTPM, researchers should report measurement error instead of repeated measurement statistics. Furthermore, the ISO standard should be revised to include measurement error statistics.