When Is It Appropriate to Infer Cognitive Impairment on the Basis of Premorbid IQ Estimates? A Simulation Study

Whether an individual meets psychometric criteria for cognitive impairment is dictated by the comparison criterion, which is typically either a normative mean or a known or estimated previous level of ability. This study investigated the conditions under which adjusting normative expectations based on estimated premorbid intelligence would be appropriate. A simulated data set was derived and several parameters were systematically varied: the correlation between premorbid intelligence and the cognitive test score, the cutoff used to classify a score as “normal” or “abnormal”, and the population base rate of cognitive impairment. Simulation results demonstrated that the correlation between premorbid intelligence and the cognitive score was the only parameter to substantially influence the trade-off between the two normative approaches, with correlations above ρ =.35 signifying greater advantage to adjusting normative expectations by premorbid intelligence. These findings inform common neuropsychological practices regarding the application of premorbid intelligence estimates to the detection of cognitive impairment.