Performance Invalidity is Associated with Lower Test of Premorbid Functioning IQ Estimates in Performance-Based Versus Demographics-Based Prediction Models: Implications for Neuropsychological Practice and Research

Abstract Premorbid intellectual function estimation is essential in neuropsychological evaluation to differentiate acquired neurocognitive decline from baseline functioning. Although confounding variables may impact performance-based premorbid estimates, little is known about the impact of invalid test performance on such estimates. This cross-sectional study examined the impact of performance invalidity on performance based vs. demographics-based premorbid estimation using the Test of Premorbid Functioning (TOPF). Using four freestanding performance validity tests (PVTs), 489 adults referred for comprehensive neuropsychological evaluation were divided into valid (≤ 1 failure) or invalid (≥ 2 failures) groups. Three premorbid estimates were then calculated for each patient from the TOPF (i.e., Reading-Only, Demographics-Only, and Combined). IQ estimates from all 3 TOPF models were significantly lower in the invalid group compared to the valid, with the largest effect size for the Reading-Only model and the smallest for the Demographics-Only model. Supplementary analyses revealed negligible within-group effects in the valid group and medium effects in the invalid group. TOPF prediction models including the performance-based irregular word reading score (Reading-Only and Combined Models) were significantly impacted by invalid performance. Although considered a hold test, irregular word reading scores are susceptible to invalid performance. The Demographics-Only model should be utilized in such cases.