Regression-based norms in Clinical Neuropsychology

In clinical neuropsychology, regression-based norms (RBNs) are statistical methods used to interpret test scores by accounting for individual differences in demographics and other relevant variables. They provide a more personalised and precise way of evaluating cognitive performance compared to traditional age-only normative data.

Key features of regression-based norms: #

1. Adjusting for individual differences:
   • RBNs use statistical regression models to predict an individual’s expected test performance based on variables such as:
     • Age
     • Education
     • Gender
     • Cultural background
   • These models account for how these variables influence cognitive test performance, offering a personalised baseline for comparison.
2. Predicting expected scores:
   • Instead of relying on static average scores from a normative sample, RBNs calculate an individual’s expected score by plugging their demographic and background information into the regression equation.
3. Comparing observed and expected scores:
   • After the individual’s “expected score” is computed, their actual test score (observed score) is compared to the expected score. This comparison helps identify whether their performance is above, below, or within the expected range for someone of their demographic profile.
4. Adjusting for Covariates:
   • RBNs can also consider other covariates (e.g., medical history, mental health conditions) that may influence performance, offering a more refined evaluation.

Why use regression-based norms in clinical neuropsychology? #

• Improved accuracy: They allow for a nuanced interpretation of test results, accounting for factors that can impact cognitive performance.
• Avoiding misclassification: By considering demographic and individual factors, RBNs reduce the chances of falsely identifying someone as impaired or underperforming.
• Dynamic adjustments: They adapt to the diversity of test-takers, offering more tailored assessments than traditional norms derived from a homogeneous sample.

Clinical practice example #

Consider a memory test for a patient:

• Traditional norms might label a person as impaired if their score is below a fixed raw cutoff number (e.g., scoring <25 might be considered impairment).
• Using regression-based norms, the test score is evaluated based on the person’s specific age, education level, and other variables. A lower score might appear normal for individuals with less education, but the same score could indicate potential impairment for someone with a higher education level.

Guidelines for demographic normative adjustments in clinical neuropsychological assessment #

• Neurodiagnostic assessment where the goal is detecting cognitive decline relative to presumed premorbid functioning
• Minimising confounding demographic variables (education, sex, ethnicity) in determination of acquired cognitive dysfunction

• Psychoeducational evaluations
• Intellectual disability determinations
• Vocational assessments requiring absolute functional levels
• Cases where general population comparisons are needed

Use age-based norms when: #

• Evaluating functioning relative to general population standards
• Assessing capacity for independent living
• Determining treatment consent capacity
• Diagnosing learning or intellectual disabilities
• Assessing IQ

Use demographic adjustments when: #

• Primary purpose is detecting acquired cognitive decline
• Evaluating return-to-work capacity in specialised occupations
• Determining presence of neuropathology

Professional standards #

• Maintain consistent application within assessment
• Document clear clinical reasoning for adjustment choices
• Consider test-specific normative properties
• Apply adjustments that align with referral question

Summary #

In summary, regression-based norms are a statistically robust method for interpreting neuropsychological test results, often offering a more individualised and accurate approach compared to traditional normative techniques, but they do not set out to replace traditional age-based norms as the former are intended for certain circumstances only.

Useful references #

Kiselica, A. M., Karr, J. E., Mikula, C. M., Ranum, R. M., Benge, J. F., Medina, L. D., & Woods, S. P. (2024). Recent advances in neuropsychological test interpretation for clinical practice. Neuropsychology review, 34(2), 637-667.

Podcast: Introduction to Neuropsychological Test Interpretation and Regression-Based Norms #

Dr. Andrew Kiselica talks about regression-based norms in this Neuropsychology Podcast from Navigating Neuropsychology.