home
  Validation of the Computer Administered Neuropsychological Screen for Mild Cognitive Impairment (CANS-MCI)


Jane B. Tornatore, PhD1, Jo Anne Laboff1 Emory Hill, PhD1,
& Bruce Center2

1. Screen, Inc. Seattle WA, 2. University of Minnesota, Minneapolis, MN

Poster: Presented at the 16th American Association for Geriatric Psychiatry, Hawaii, 2003

ABSTRACT

Background: As treatments for Alzheimer's disease (AD) emerge, it becomes important to identify people who have the earliest signs of the cognitive impairments most likely to become AD. People with Mild Cognitive Impairment (MCI) appear to develop AD at a rate of 10-15% a year. Since most new treatments for dementia focus upon slowing the progression of AD, it is critical to identify the markers of future cognitive decline at the earliest stage.

Objective: The Computer-Administered Neuropsychological Screen for Mild Cognitive Impairment (CANS-MCI), a fully self-administered touch screen battery, was designed to recognize the earliest predictive signs of AD. The CANS-MCI incorporates tests of all cognitive dimensions known to predict AD: executive inhibitory functions, memory, spatial relations, and language recognition/retrieval. This study presents analyses to establish the factor analytic properties of the items and the ability of the measures to differentiate between MCI and normal cognitive functioning.

Methods: 265 elderly volunteers enrolled in a 3-year longitudinal NIA-funded study.
We conducted exploratory (N=132) and confirmatory (N=116) factor analysis. ROC curve analysis (N=256) was used to examine the sensitivity and specificity of the individual tests. A logistic regression was run on a subset of participants (N=42) who received a full neuropsychological exam.

Results: Confirmatory factor analysis supported the highly correlated 3-factor model (Memory, Language/Spatial Fluency, and Executive Function/Mental Control) found in the exploratory analysis, indicating that the tests measure the intended cognitive dimensions. The factor loadings in the confirmatory analysis were all significant and ranged from .55-.96; goodness of fit indicators were strong (Bentler Comparative Fit Index = .99; Root Mean Square Error of Approximation=.038). For sensitivity percentages ranging between 70-80%, specificity scores ranged from 39%-82%. Logistic regression analyses indicated that a model incorporating 2 CANS-MCI tests and age correctly classified 88% of the cases

Conclusions: The CANS-MCI is an easily administered, robust screening tool measuring all cognitive dimensions that predict whether professional testing for pre-AD cognitive impairments is warranted. Analyses to date indicate respectable levels of validity, a clear representation of the 3 primary factors predictive of AD, and the ability to distinguish between MCI and normal functioning.


INTRODUCTION

People with Mild Cognitive Impairment (MCI) appear to develop AD at a higher rate than the general elderly population. Instruments focused upon MCI measurement would provide useful screening information for decisions concerning full diagnostic evaluations for AD. Brief or automated neuropsychological tests may be the preliminary step most suited to determining the need for evaluations, which require costly neuropsychological or neuroimaging techniques. Current methods of detection are costly & often deferred until later in the disease process when interventions to delay AD are likely to be less effective.

Previous studies found that tests sampling different cognitive domains, when combined, significantly enhance the predictive validity of a test battery because of variations in the initial cognitive deficits associated with AD.


METHODS

Subjects
A total of 265 elderly people were recruited through senior centers & retirement homes in Washington State for a 3-year longitudinal NIA-funded study.

STATISTICAL ANALYSES

We conducted exploratory (N=132) and confirmatory (N=116) factor analysis. ROC curve analysis (N=256) was used to examine the sensitivity and specificity of the individual tests using the lowest 10th percentile of cognitive functioning versus the highest 90th percentile of the Weschler Memory Scale Logical Memory II score (LMS-II) as the criterion standard. A preliminary logistic regression was run on a subset of participants (N=42) who received a full neuropsychological exam. In addition to the CANS-MCI tests, age, gender, and education were included in the analysis.


RESULTS

Results: Confirmatory factor analysis supported the highly correlated 3-factor model (Memory, Language/Spatial Fluency, and Executive Function/Mental Control) found in the exploratory analysis, indicating that the tests measure the intended cognitive dimensions. The factor loadings in the confirmatory analysis were all significant and ranged from .55-.96; goodness of fit indicators were strong ( 2 70 = 97.2, p=.13.; Bentler Comparative Fit Index = .99; Root Mean Square Error of Approximation=.038). Because the three factors were highly correlated (.84, .82, .71), we tested for the presence of one global factor. The goodness of fit statistics were not as strong ( 2 3 = 63.8, p<.001) and the factor loadings were weaker. For sensitivity percentages ranging between 70-80%, specificity scores ranged from 39%-82% with LMS-II used as the criterion standard. Logistic regression analyses indicated that a model incorporating 2 CANS-MCI tests and age correctly classified 88% of the cases (Nagelkerke pseudo-R2 =.48)

DISCUSSION

Factor Analysis: The Memory, Language/Spatial Fluency, and Executive Function/Mental Control factors were highly correlated, as would be expected of cognitive functions all associated with MCI and/or AD. Even though highly correlated, they appear to be distinct factors given the poorer fit of the model when we tested just one factor.

Sensitivity/Specificity: The tests demonstrate a moderate ability to discriminate between cognitive impairment and normal cognitive functioning. This ability is expected to improve when the criterion standard available is a more extensive and more detailed set of measures (full evaluation). The CANS-MCI has multiple measures for each cognitive domain and combined scoring is likely to increase the sensitivity and specificity in future analyses using full neuropsychological evaluations as the criterion standard.

Logistic Regression: The preliminary evaluation of differences between normal and impaired cognitive ability groups, as determined by full neuropsychological evaluations (N= 42) indicates a high degree of prediction. The small number of MCI's (n=12) could account for the fact that only two tests, Clock Hand Placement and Word-to-Picture Matching (Latency) were included in the final equation.

CONCLUSION

As effective treatments for AD emerge, it becomes important to identify people who have the earliest signs of the cognitive impairments most likely to become AD. The CANS-MCI tests appear to measure 3 different cognitive domains & differentiate memory impaired from normal elderly, as determined by the WMS-R LMS II. The CANS-MCI is an easily self-administered, computer scored and interpreted touch screen battery that ascertains whether more intensive testing for early cognitive impairment is warranted.


Table 1: Confirmatory Factor Analysis

Executive Function/Mental Control
General Reaction Time (Latency) .60
Design Matching -.76
Word-to-Picture Matching (Latency) .75
Stroop (Discordant Latency) .72
WAIS Digit Symbol Test -.81

Language/ Spatial Fluency
Clock Hand Placement .69
Picture Naming .82
Picture Naming (Latency) -.96
Mattis Dementia Rating Scale - Initiation Scale .63

Memory
Free Recognition I .66
Guided Recognition I Errors -.58
Free Recognition II .61
WMS-R LMS-I .75
WMS-R LMS-II .80
Mattis Dementia Rating Scale - Memory Scale .55


Table 2: Specificity Ranges for 70-80% Sensitivity (LMS II Delayed Memory)

CANS-MCI Score % Specificity
General Reaction Time 59-50
Design Matching 71-57
Word-to-Picture Matching (Latency) 74-69
Clock Hand Placement 66-55
Stroop (Discordant Latency) 57-39
Free Recognition I 82-73
Guided Recognition I Errors 80-63
Free Recognition II 66-47
Free Recognition I & II 78-71
Picture Naming 75-58
Picture Naming (Latency) 69-60

 
 

   
home