Week 5 - Psychometric Assessment (neurology) Flashcards
(25 cards)
What are examples of conditions requiring a neuropsychological assessment?
- TBI - traumatic brain injury
- Stroke
- Dementia - Alzheimer’s, Parkinson’s
- Epilepsy
- Psychiatric disorders
What are 7 cognitive abilities that are assessed in neuropsychology?
- Orientation - are they confused/disoriented?
- Intellectual functioning
- Memory
- Language
- Attention and concentration
- Visual perception
- Executive - frontal lobe
Rationale of deficit measurement - comparing to population norms
Lezak argues that we can only quantify an individual’s acquired impairment compared with an “optimal” performance
Neuropsych assessment approach 1: Halstead-Reitan Test Battery
Standardised test for brain-behaviour relationships in kids/adolescents.
Neuropsych assessment approach 2: Boston process approach
To analyse HOW someone performs a task/test, rather than the final outcome.
What is a fixed battery?
- Quantitative/acturial
- Halstead-Reitan Neuropsychological Test Battery
What is a flexible battery?
- Qualitative/process oriented
- Boston Process Approach
What are four things to compare a patient’s performance with?
- Premorbid performance
- Peers
- Cognitive ability
- Different subscales
What is CTT Reliability, and what is the equation?
Classical Test Theory: understand how measurement errors affect test scores.
Observed score = true score + error score
What are four sources of error?
- Environmental: distractions
- Individual: fatigue, boredom
- Test administrator
- Test itself
What are reliability and validity?
- reliability = consistency of results
- validity = does the test measure what it says? Generalisation to wider population
Reliability - test-retest
How reliable are test scores over time?
- High correlation = great!
Reliability - Inter-rater
Agreement between raters for rating scales (Intraclass Correlation Coefficient [ICC]).
- Do we both look at the colour blue and both see blue?
- Do two different teachers rate the same student the same on the test?
Reliability - Internal consistency
How well do items collectively measure the same underlying construct?
- Split-half: 1 half of test is summed and correlated with other half for Pearson’s r.
- Cronbach’s Alpha: above 0.7 is acceptable, 0.8 is good and 0.9 is excellent!
- Lower when co-variance of items differ.
Reliability - Standard Error of Measurement (SEM)
The Error bar around a test score
- smaller suggests more precise
- bigger suggests score is affected by random error
Validity - Face validity
How does the measure appear to respondents on the surface level?
Validity - Content validity
Is the test comprehensive and relevant to everything it aims to assess?
- eg. shapes flavours, but one is missing = bad content validity
Validity - Criterion validity (C + P)
Does the measure correlate well with another measure of a gold standard?
- Concurrent: predicts outcome at one time point (cross-sectional)
- Predictive: predicts variable at at a future time (longitudinal)
WAIS
Wechsler Adult Intelligence Scale - measures cognitive abilities
Validity - Construct validity (D + C)
Does the test measure things theoretical constructs? eg. IQ, happiness.
- Discriminant: doesn’t align with things that are different
- Convergent: does align with things that are similar
What is utility?
Utility is the practical aspects of the test eg. time, expenses, readability.
What is sensitivity?
How often test predicts a positive.
Good sensitivity: true positive
Poor sensitivity: false negative
What is specificity?
How often a test predicts a negative.
Good specificity: true negative
Poor specificity: false positive
What is PPV?
Positive predictive value = probability of a true positive
