Detecting significant change in neuropsychological test performance: A comparison of four models

A major use of neuropsychological assessment is to measure changes in funct ioning over time; that is, to determine whether a difference in test perfor mance indicates a real change in the individual or just chance variation. U sing 7 illustrative test measures and retest data from 384 neurologically s table adults, this paper compares different methods of predicting retest sc ores, and of determining whether observed changes in performance are unusua l. The methods include the Reliable Change Index, with and without correcti on for practice effect, and models based upon simple and multiple regressio n. For all test variables, the most powerful predictor of follow-up perform ance was initial performance. Adding demographic variables and overall neur opsychological competence at baseline significantly but slightly improved p rediction of all follow-up scores. The simple Reliable Change Index without correction for practice performed least well, with high error rates and la rge prediction intervals (confidence intervals). Overall prediction accurac y was similar for the other three methods; however, different models produc e large differences in predicted scores for some individuals, especially th ose with extremes of initial test performance, overall competency, or demog raphics. All 5 measures from the Halstead-Reitan Battery had residual (obse rved - predicted score) variability that increased with poorer initial perf ormance. Two variables showed significant nonnormality in the distribution of residuals. For accurate prediction with smallest prediction-confidence i ntervals, we recommend multiple regression models with attention to differe ntial variability and nonnormality of residuals.