60 Are all Embedded Measures Created Equal? A look at Embedded PVTs in Major Neurocognitive Disorder

Abstract

Objective:

Although performance validity is critical in determining the quality and accuracy of test data, research suggests not all neuropsychologists incorporate performance validity tests (PVTs) in dementia evaluations (McGuire et al., 2019). Furthermore, well-validated embedded measures, such as Reliable Digit Span (RDS) from the Wechsler Adult Intelligence Scale - Fourth Edition (WAIS-IV), have evidenced an unusually high number of failures in a dementia population when utilizing typical clinical cut-offs (Zenisek et al., 2016). The objective of this study is to explore performance on embedded PVTs among older adults who have a major neurocognitive disorder (MND), specifically among Alzheimer disease (AD) and non-AD patients.

Participants and Methods:

Archival data from outpatient neuropsychological evaluations were analyzed. All participants were at least 65 years of age, diagnosed with a MND, and completed Digit Span from the WAIS-IV, Brief Visuospatial Memory Test- Revised (BVMT-R), and Hopkins Verbal Learning Test-Revised (HVLT-R). In total, 84 participants, aged 67-96 (M=78.44, SD=6.11) with 6-20 years of education (M=13.47, SD=3.30), were included. The sample predominantly identified as female (n=60) and White (n=61). More individuals were diagnosed with AD (n=50) than non-AD dementia (n=34). Common non-AD diagnoses included Vascular (n=44), Lewy bodies (n=8), and Parkinson’s (n=2) dementias. Fisher’s Exact Test of Independence was used to account for the smaller sample to determine if there was a nonrandom association between diagnosis (AD vs non-AD) and embedded PVT performance: RDS< 7, BVMT-R Hits<4, BVMT-R Recognition Discrimination (RD) < 4, and HVLT-R RD < 5 (Bailey et al., 2018).

Results:

The Fisher’s Exact Test of Independence revealed a statistically significant association between neurocognitive diagnosis and RDS (p= .008), BVMT-R RD (p<.001), and HVLT-R RD (p<.001). BVMT-R Hits were not significantly associated with diagnosis (p = 0.10). These measures evidenced opposite patterns with RDS demonstrating a higher percentage of fails for the non-AD (63%) versus AD (20%) group. The AD group had a higher percentage of fails for BVMT-R RD (58% for AD and 13% for non-AD groups) and HLVT-R RD (66% for AD and 29% for non-AD group).

Conclusions:

The current study suggests performance on embedded PVTs vary across MND diagnoses. Individuals with a non-AD diagnosis were more likely to fail RDS than those with AD. This is likely secondary to attention and working memory demands that are mediated by the frontal-subcortical networks, which are less impacted by AD pathology (Bonelli & Cummings, 2022; Loring et al., 2016). In contrast, AD patients were more likely to fail embedded PVTs within memory measures, which are largely mediated by the mesial temporal cortex associated with AD (Pluta, 2022). These results suggest embedded measures operate differently based on diagnosis and neuroanatomical systems affected. The clinical relevance of these findings includes potentially using alternative PVTs or different cut-offs based on diagnosis. Future research should attempt to better delineate more appropriate, as well as time efficient, PVTs among the dementia population.