Normative data and validation of a regression based summary score for assessing meaningful neuropsychological change

Abstract

Reliable detection and quantification of longitudinal cognitive change are of considerable importance in many neurological disorders, particularly to monitor central nervous system effects of disease progression and treatment. In the current study, we developed normative data for repeated neuropsychological (NP) assessments (6 testings) using a modified Standard Regression-Based (SRB) approach in a sample that includes both HIV-uninfected (HIV−, N=172) and neuromedically stable HIV-infected (HIV+, N=124) individuals. Prior analyzes indicated no differences in NP change between the infected and uninfected participants. The norms for change included correction for factors found to significantly affect follow-up performance, using hierarchical regression. The most robust and consistent predictors of follow-up performance were the prior performance on the same test (which contributed in all models) and a measure of prior overall NP competence (predictor in 97% of all models). Demographic variables were predictors in 10%-46% of all models and in small amounts; while test retest interval contributed in only 6% of all models. Based on the regression equations, standardized change scores (z-scores) were computed for each test measure at each interval; these z scores were then averaged to create a total battery change score. An independent sample of HIV− participants who had completed 8 of the 15 tests was used to validate an abridged summary change score. The normative data are available in an electronic format by email request to the first author. Correction for practice effects based on normative data improved the consistency of NP impairment classification in a clinically stable longitudinal cohort after baseline.

The Wide Range Achievement Test–4 Reading subtest “holds” in HIV-infected individuals

“…For those HIV+ individuals followed longitudinally, global neurocognitive change was examined across the two study visits using a standard regression-based change score (sRCS) that corrects for test reliability, regression to the mean, practice effects, and other factors that may influence follow-up test results in people who are neurocognitively stable (see Cysique et al, 2011 for detailed algorithm and validation of this method in HIV-infected individuals). sRCS were calculated based on neuropsychological performance of a previously identified reference sample of both seronegative adults and neuromedically stable HIV+ individuals (Cysique et al, 2011).…”

“…sRCS were calculated based on neuropsychological performance of a previously identified reference sample of both seronegative adults and neuromedically stable HIV+ individuals (Cysique et al, 2011). In brief, the follow-up neuropsychological performance of the HIV+ participants included in the current study was compared to the predicted follow-up performance based on the reference group established in the Cysique et al (2011) study. The difference between predicted and actual neuropsychological (NP) scores was converted into a standard, z-score (by dividing the difference by the error term of the regression equation) that can be used as a continuous or categorical score.…”

A Longitudinal Analysis of the Impact of Physical Activity on Neurocognitive Functioning Among HIV-Infected Adults

“…The domains and tests included verbal fluency, working memory, speed of information processing, learning, memory, executive function, and motor function [45]. Due to the repeated neuropsychological testing of participants, we used published norms for change to compute practice-adjusted scaled scores for each neuropsychological test [45].…”

Randomized Trial of Central Nervous System–Targeted Antiretrovirals for HIV-Associated Neurocognitive Disorder

“…An intent-to-treat (ITT) analysis included all participants who provided GDS outcome data at week 16. Normative adjustments for practice effects due to repeated NP test exposures became available after the study began [12] and were used in secondary analyses. An as-treated analysis was performed on all participants who remained on the regimen to which they had been randomized through week 16.…”