NeuroRehabilitation An InterdlaclplillllIJ oIournIII
ELSEVIER
NeuroRehabilitation 9 (1997) 179-186
Comparison of cognitive status profiles of healthy elderly persons with dementia and neurosurgical patients using the neurobehavioural cognitive status examination Noomi Katz a,*, Adina Hartman-Maeir a Pnina Weiss b , Nira Armonc a School of Occupational Therapy,
Hebrew University Jerusalem, Mount Scopus, PO Box 24026, Jerusalem 91240, Israel bBelinson Medical Center, Petach Tikua, Israel CDay Center, Talpiot, Jerusalem, Israel
Abstract
The purpose of this study was to examine and compare the cognitive profiles of three groups of Israeli individuals who exhibit cognitive deficits for diverse reasons, using the Neurobehavioural Cognitive Status Examination (Cognistat, former name NCSE). The test is a standardized cognitive screening instrument which includes general areas of alertness, attention, orientation, language (comprehension, repetition, naming), construction, memory, calculations and reasoning (similarities, judgement). Assessment results are provided as a profile of the different domains and graded on four performance levels (average, mild, moderate, severe). Subjects included 47 healthy independent elderly subjects, 47 neurosurgical patients and 42 persons suffering from dementia. Supporting the hypothesized direction, statistically significant differences were found among the three groups on raw scores of all individual subtests and on the four performance levels with the healthy elderly subjects showing the highest performance and persons with dementia the lowest. Construction subtest scores were low for all groups and seemed to detect the aging process as well as disease-related dysfunction. The mean scores and standard deviations of all groups for most subtests were similar or a little lower to those reported in the American standardization data suggesting the test's cross-cultural applicability. In conclusion, the Cognistat was found to be a useful screening test for clinical and research purposes, however, further research is required. © 1997 Elsevier Science Ireland Ltd.
Keywords: Cognitive assessment; Cognistat (NCSE); Elderly; Dementia; Neurosurgical
1. Introduction
The purpose of this study was to examine and compare the cognitive profiles of three groups of
* Corresponding author. Tel.: 5324985.
+ 972 2 5813246; fax: + 972 2
individuals who exhibit cognitive deficits for diverse reasons using the Neurobehavioral Cognitive Status Examination (Cognistat, former name NCSE) [1]. The Neurobehavioural Cognitive Status Examination (Cognistat) was developed as a screening test for assessing cognitive status across a variety of domains [2]. The Cognistat is a short battery of
1053-8135/97/$17.()() © 1997 Elsevier Science Ireland Ltd. All rights reserved. PH S1053-8135(97)00028-0
180
N. Katzetal. /NeuroRehabilitation9 (1997) 179-186
subtests in five cognitive areas: Language (Comprehension, Repitition, Naming); Construction, Memory, Calculation and reasoning (Similarities and Judgment). The level of consciousness, Orientation and Attention areas are evaluated and scored before assessing the other cognitive abilities. In each sub test an initial screen item is given and only if the subject fails, other items are given (metric) in order to establish the level of competence. The results are presented in a profile which is divided into four performance levels: average performance, mild impairment, moderate and severe impairment. The Cognistat was found to be more sensitive in the detection of cognitive dysfunction in 30 patients with known brain lesions than two known mental status examination instruments (MMSE and CCSE) [3]. The Cognistat was also found to be a rapid and sensitive measure of cognitive function, predicting functional changes as a result of inpatient stroke rehabilitation [4]. It was further found to significantly differentiate between stroke patients and orthopedic control patients [5]. A comparison between the two unilateral (right vs. left) stroke groups revealed different patterns of cognitive deficits; yet this difference did not reach statistical significance, probably because of the small sample size of each group (n = 12) [5]. Barrows [6] presents the main cognitive dysfunction of persons with Chronic Fatigue Immune Syndrome using the Cognistat profile which revealed impairments in Memory, Construction, Calculations and to a lesser degree in Attention. According to the author, the Cognistat scores corroborated clients' perceptions of their impairments. The Cognistat was also used to detect cognitive changes following neurosurgery [7]. Among 11 patients who were evaluated pre- and post-surgery, six improved but five showed a deterioration which led in four cases to a second operation. The authors suggested that the Cognistat was useful for serial testing, enabling the comparison of profiles where patients act as their own control. The same approach using the Cognistat to monitor change before and after continuous drainage to predict shunt outcome is presented by Weiss [8].
Margolin [9] showed the use of the Cognistat for documenting the pattern and temporal course of cognitive decline in dementia patients. Comparing 13 dementia patients (mean age 71.3) to normal control patients (mean age 69.8) significant differences were found on all subtests excluding Attention and Judgment. The most severe deficits were seen in Memory and Orientation sub tests which are congruent usually with the earliest and most consistent impairments in dementia. A moderate impairment was also found on the Construction sub test which is also congruent with evidence that Performance IQ is more impaired than Verbal IQ [9]. In a large study performed by Logue et al. [10] to look at the psychometric properties of the Cognistat with psychiatric inpatients. Eighthundred and sixty-six patients aged between 15 and 92 years were tested at admission and showed a poorer performance than the normative sample on all subtests. Age was negatively correlated, especially with performance on Construction ( - 0.46), Memory (- 0.31) and Similarities ( - 0.31) subtests, which was in the expected direction. However, intercorrelations among subtests did not yield the predicted pattern of high correlations between similar domains and low correlations between dissimilar ones. Therefore, they recommended that further research needs to be carried out. One major critical point that was raised by the authors relates to the screen item of the Construction subtest which is loaded with the Memory component (largest correlation Construction and Memory r = 0.48) and therefore, they recommend a change in this item. Further studies of psychiatric patients found that the Cognistat differentiated between patients with and without organic mental disorder (OMD) [11] and between OMD and elderly depressed inpatients [12]. Some Memory deficits were found in all patients, but the best predictor of OMD using multiple regression was the Construction subtest followed by the Memory subtest. In the geropsychiatric study [12] the Cognistat was found to be more sensitive to organically based cognitive impairment then the MMSE. The authors used a criterion of two or more scales below the
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
cutoff, which maximizes sensitivity, as being suggestive of cognitive deficits [12]. The issue of the accuracy of the screen vs. metric approach was recently studied in a sample of 95 male clients (mean age 66, with a mean of 11 years of education) [13]. They found that many subjects who passed the screen item failed on the metric part of the same domain, thus producing false negative results. Therefore, the authors suggest that all items (screen and metric) should be tested to reduce the possibility of missing deficit areas when they exist. It seems to us that the above findings need to be further investigated, especially checking the different screen items, as the general conclusion suggested above eliminates one of the benefits of the Cognistat. In summary, the Cognistat instrument has been utilized with a variety of patient populations where brain dysfunction was suspected and initial results showed that it was sensitive in detecting cognitive impairments, differentiating between groups as well as measuring changes over time. Some caution was expressed by Lezak [14] regarding the sensitivity of the test to mild head trauma because they found that in using the test they had to supplement the Attention and Language tasks with few more difficult items in order to assess subtle deficits in these areas. However, any instrument developed in one country when used cross-culturally needs to be further validated to ascertain its applicability. Therefore, the Cognistat Test Booklet was translated into Hebrew with minor cultural changes such as adaptations of some of the Judgment subtest situations to Israeli conditions, or few changes in the screen item of the Naming subtest (the pen parts). The subtests were administered in strict accordance with the test manual. The Hebrew translation is used in clinics by a growing number of therapists. The first study in Israel comparing 15 cerebrovascular accident (CVA) elderly patients to 24 healthy elderly showed that healthy subjects mean scores were higher than CVA patients on all subtests except for Construction. Significant statistical differences were found for Orientation and three Language subtests [15]. Based on this initial study it was recommended to
181
further investigate the test's validity with additional populations and larger samples. Hence, the specific objectives of the present study were twofold: (1) to compare the cognitive profile of three groups, healthy elderly, neurosurgical patients and persons suffering from dementia; and (2) to compare the Israeli results to the American standardization data from the Cognistat manual. It was hypothesized that significant differences will be found among the three groups on all subtests, where the healthy group will show an average profile and clients with dementia the most severe deficits. 2. Method
2.1. Subjects and procedure Three groups of subjects were tested with the Cognistat: (1) 47 healthy independent elderly subjects, 13 men and 34 women (mean age 74.75, S.D. = 5.44); (2) 47 neurosurgical patients suffering from tumors, trauma, hydrocephalus, or cerebral hemorrhage, who were tested prior to surgery, 26 men and 21 women (mean age 54.7, S.D. = 21.3); and (3) 42 clients suffering from DAT (Dementia Alzheimer Type), 20 men and 22 women (mean age 77.1, S.D. = 7.56). The healthy elderly participants were recruited from independent living settings and had no known neurological or psychiatric illness. Neurosurgical patients were tested consecutively as inpatients as part of their general evaluation prior to surgery. Clients suffering from DAT were tested at a Day Center as part of their evaluation for community intervention planning. All subjects gave their written informed consent to be evaluated as part of their admission to the hospital or day treatment. The data was analyzed by a one-way ANOVA for a general F value between the three groups on each subtest followed by a post-hoc Scheffe test for the source of significant differences among groups. Chi-square analysis was used to compare frequencies of performance levels between the two patient groups.
N. Katz et al. / NeuroRehabilitation 9 (J997) 179-186
182
3. Results 3.1. Demographics
Demographic variables of age and gender were analyzed in relation to the Cognistat score. No significant correlations with age were found among the two elderly groups (Healthy and DAT) as they were very homogenous elderly groups. Within the neurosurgical group, significant negitive correlations with age were found for Construction (r = - 0.60), Memory (r = - 0.54) and Similarities (r = -0.54) all at P < 0.0001, Comprehension (r = -0.31, P < 0.04) and repitition (r = - 0.37, P < 0.01). The relationship with age was in the expected direction as older subjects performed lower. The only gender differences were also found in the neurosurgical group. Wilcoxon non-parametric analysis showed significant differences around P < 0.05-0.01 for Orientation, Attention, Construction and Judgment where men scored higher then women and Repitition where women scored higher, P < 0.02. These results were not expected and seem unclear and inconsistent. 3.2. Cognitive status
Cognistat mean scores and standard deviations
for the three groups are presented in Table 1 along with F values and post hoc Scheffe results. Significant F values were found for all subtests. Post-hoc Scheffe comparisons of means showed significant differences between healthy subjects and subjects with dementia on all subtests. Healthy and neurosurgical patients differed significantly on six subtests, excluding three Language tests where both groups performed in the average range and Construction on which both performed low, therefore the difference did not reach a significant level. In Fig. 1, the profiles of the three group means are depicted along the four performance levels. It can be seen that persons suffering from DAT perform in the mild to severe impairment ranges with Construction, Memory, Naming and Similarities being the most severely impaired. Neurosurgical patients' performance level is in the low average to mild impairment range for most subtests and in the mild to moderate impairment range for Construction, Memory and Similarities. Healthy elderly subjects perform at the average range on all subtests, except for Construction and Memory subtests where average extends to the mild impairment range, as specified for the normal elderly population over age 65 in the test manual.
Table 1 Cognistat mean scores of three groups: healthy elderly (n = 47), DAT' (n = 42) and neurosugical patients (n = 47), and one-way ANOVA between groups Group M (S.D.) Cognistat subtest
Healthy
DAT'
Neurosurgical
F (P)
Scheffe b
Orientation Attention Comprehension Repetition Naming Construction Memory Calculation Similarities Judgment
11.91 (0.28) 6.91 (1.50) 5.53 (0.99) 10.91 (2.72) 7.19 (2.18) 3.17(2.31) 8.45 (3.57) 3.85 (0.46) 5.740.61) 5.06 (0.99)
7.57 (3.51) 4.69 (1.63) 4.48 (1.36) 8.46 (3.74) 4.0 (1.78) 2.0 (1.34) 5.05 (3.71) 3.05 (1.08) 3.15 (2.35) 3.30.62)
8.85 (2.93) 5.91 (2.47) 5.02 (1.45) 10.33 (4.13) 8.0 (2.15) 2.6 (2.26) 6.4 (4.49) 2.89 (1.32) 3.89 (2.07) 4.05 (1.60)
33.11 (0.0001) 14.77 (0.0001) 7.53 (0.0008) 5.37 (0.006) 46.13 (0.0001) 3.68 (0.03) 8.43 (0.0004) 11.99 (0.0001) 19.43 (0.0001) 17.04 (0.0001)
H-D; H-N H-D; H-N; D-N H-D H-D H-D;D-N H-D H-D;H-N H-D; H-N H-D;H-N H-D; H-N
aDA T, Dementia Alzheimer's Type. b Indicates the groups that scores were significantly different from each other: H, healthy; D, DAT; N, neurosurgical.
183
N. Katzetal. /NeuroRehabilitation9 (1997) 179-186
Average
~
-----.... ~ .....
- - - "" .... , ,
..
Mild
,,
~ \
,
..
,
Moderate
...
\ .
/
~ V.; '\ / / ,,/ "\.':" \\ V . . "',' ;'", , ,'", \ .... .. , . ,\.
,
... , . .,;
11..
o
11.. W
::&
oo
t/)
Z
o
1- Healthy - -
,,
o..J ct o
:E w :E
l-
0::
,
,
Severe ~
./
,
o
C
:J ...,
Neuro . - .. OAT I
Fig. 1. Cognistat profiles of three groups: healthy elderly; OAT and neurosurgical patients.
Mean scores of neurosurgical patients and persons with DAT differed significantly on Attention and Naming subtests only, although subjects with DAT performed lower on all subtests excluding Calculations. However, looking at mean scores does not show an accurate picture of individual subjects' performance. Therefore, frequencies for
the two groups according to the four performance levels are presented in Table 2 along with results of Chi-square analysis. Naming and Construction subtests highly differentiated between the groups followed by Judgment, Attention and Repitition. It can be seen that neurosurgical patients show higher percentages of average scores on almost
Table 2 Frequency (percentages) of four performance levels on each subtest for DATa (n chi-square analysis between groups Cognistat subtest
Orientation Attention Comprehension Repetition Naming Construction Memory Calculation Similarities Judgment
2
3
4
=
47), and
11.9 50 9.5 16.7 23.8 31 11.9 26.2 19 31
19 19 16.7 14.3 54.8 23.8 11.9 9.5 4.8 9.5
28.6 2.4 11.9 38.1 11.9 33.3 59.5 0 47.6 21.4
53.2 61.7 70.2 63.8 80.9 40.4 34 57.4 38.3 66
2
3
4
X 2 (p)
17 21.3 2.1 10.6 8.5 4.3 6.4 21.3 14.9 2.1
12.8 12.8 14.9 2.1 8.5
17 4.3 12.8 23.4 2.1 48.9 46.8 12.8 36.2 17
NS 11.3 (0.01) NS 11.3 (0.01) 45.9 (0.000) 21.9 (0.000) NS NS NS 15.7(0.001)
a OAT, Dementia Alzheimer's Type. average range; 2, mild impairment; 3, moderate impairment; 4, severe impairment.
b 1,
42) and neurosugical patients (n
Neurosurgical group performance level
OAT" group performance levels
40.5 28.6 61.9 31 9.5 11.9 16.7 64.3 28.6 38.1
=
6.4 12.8 8.5 10.6 14.9
184
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
all subtests except for Calculations. Subjects with dementia show higher percentages of severe impairments on six subtests (Orientation, Repitition, Naming, Memory, Similarities and Judgment). For persons with DAT, Calculations is the most preserved cognitive domain, it is the only subtest where the mean score was in the average range with 64% average and 26% mild impairment.
3.3. Comparison to standardization data A descriptive comparison of mean scores of the groups in this study to the American standardization data showed that on six subtests the Israeli means are a little lower than the geriatric healthy and the neurosurgical groups. The Construction subtest showed the largest difference for both groups, while Comprehension, Calculations, Similarities and Judgment showed similar means. The profile of the persons with DAT appears to be within the range shown in the Cognistat manual [1]. The three Language subtests scores are a little lower than shown in the typical profile. Memory was found to be the most impaired domain of the DAT subjects as expected, followed by Orientation and Similarities. 4. Conclusions In conclusion, the results of this study show that the cognitive profiles of the groups differed significantly in severity of deficits in the hypothesized direction, supporting the validity of the Cognistat. The instrument was found as a useful screening test applicable also to the Israeli population. Mean scores of the Israeli subjects are similar to the American standardization samples for the comparable groups. However, data to compare healthy adult groups in Israel has yet to be collected. The differences in age among this study groups, especially the younger neurosurgical group compared to the healthy elderly may have limited the gap between them. It is assumed that a healthy adult group similar in age to the patient group may have shown a larger profile difference. The gender differences found within the neurosurgical group were not expected and are not
clear. It seems an inconsistent result which may be related to the specific group studied. Further studies will have to examine if a gender relationship exists, if it does, then in which direction and in what domains. Cross-cultural validation is an essential step in using tests developed in another culture. In adapting a test, care needs to be taken to make as few changes as possible so that comparison to the original data is meaningful. On the other hand, test items should be appropriate to use in the new culture. For example, one of the problems we faced was whether some of the items from the Naming subtest are familiar, such as the Octopus or Xylophone, especially in the elderly population and in less educated populations. We have not changed them because this would mean changing actual test items which requires a more elaborate process of test validation. The higher results on the Naming subtest of the younger neurosurgical group compared to the healthy elderly group may reflect the above mentioned concern. However, the comparison of means of both groups to the American data shows similarities, thus, suggesting that the difficulties in the Naming items may be similar in both countries and do not reflect a cultural issue. One general limitation of the instrument is that it is highly verbal. The Construction subtest is the only performance subtest. Visuospatial tests are considered excellent screening devices as individuals with brain dysfunction involving frontal, parietal or occipital lesions usually fail on copying or drawing two and three-dimensional tasks [14,16]. Deficits on these tests indicate organic disturbance or decline and usually don't differentiate hemispheric involvement. According to Osmon et al. [5] the analysis of intercorrelations among Cognistat subtests and, more specifically, the Construction items showed that the screen item and the last item are too difficult, making the overall scoring of Construction too sensitive and, therefore, low for all groups. This was seen in the present study as the Construction sub test was difficult for most subjects, which indicates the sensitivity of visuospatial deficits to organic disturbance as well as in normal aging [9,12,16]. In addition, the Construction subtest involves visual
N. Kim et al. / NeuroRehabilitation 9 (1997) 179-186
Memory in the screen item making it less accurate as a constructional measure, as stated by Osmon et al. [5]. The assessment of Construction, which is critical for screening brain dysfunction and functional performance, is limited in the Cognistat. Therefore, it was recommended for example by Katz et al. [15] to use the visuomotor organization subtests of the Loewenstein Occupational Therapy Cognitive Assessment (LOTCA) [17] for further in-depth assessment when the Cognistat subtest shows constructional deficits. In general, the Cognistat appears to be a useful screening test for clinical and research use in a variety of client populations based on this research and on previous works reviewed. In the rehabilitation process, the Cognistat profile provides initial information on intact and impaired cognitive skills and may determine which cognitive areas require further, in-depth evaluation and/or intervention. In the neurosurgical group initial results before surgery are compared to a second testing following a period of recovery after surgery. Comparison of the clients' two profiles may suggest the need for a second operation, or areas which still show deficits and require cognitive intervention that may have functional implications in daily activities. It also shows the intact areas which rehabilitation professionals should build on in further intervention using the clients' abilities. This approach, taking into account intact abilities to build on, is imperative in the DAT group as they suffer from a declining process and rehabilitation is focused on maintaining function as long as possible. The results of the two studies conducted in Israel have direct implication for clinicians and researchers in Israel, suggesting that it is valid to use the Cognistat as a clinical and research tool. However, as mentioned earlier, additional data from healthy adult populations are needed to compare results to similar groups in the standardization study. Initial results imply that the Cognistat can be used as a measure of change in intervention studies, such as following surgical or other procedures, but further research needs to be carried out. It is also important to study more closely the influence of education on the performance of the test's cognitive domains. Cognitive
185
performance is known to relate to education and it may distort the interpretation of individuals results. The authors acknowledge this problem in their 'cautions section' [1], however, it would be important to have data about the relationship of education to specific subtests. Acknowledgement
The Cognistat Test Booklet was translated into Hebrew by Rina Kravitz. We would like to thank the Occupational Therapists: Shulamit Avni, Dorit Haldman, Noga Ziv and Debra Roitman for collecting the data within the healthy group. References [1] Cognistat - The Neurobehavioural Cognitive Status Examination in older psychiatric population. Fairfax CA: The Northern California Neurobehavioural Group, 1995. [2] Kiernan RJ, Mueller J, Langston JW, Van Dyke C. The neurobehavioural cognitive status examination: a brief but differentiated approach to cognitive assessment. Ann Intern Med 1987;107:481-485. [3] Schwamm LH, Van Dyke C, Kiernan RJ, Merrin EL, Mueller J. The neurobehavioural cognitive status examination: comparison with the cognitive capacity screening examination and the mini-mental state examination in a neurosurgical population. Ann Intern M ed 1987;107:486-490. [4] Mysiw WJ, Beegan JG, Gatens PF. Prospective cognitive assessment of stroke patients before inpatient rehabilitation. Am J Phys Med Rehabil 1989;68:168-171. [5] Osmon DC, Smet JC, Winegarden B, Gandhavadi B. Neurobehavioural cognitive status examination: its use with unilateral stroke patients in a rehabilitation setting. Arch Phys Med RehabilI992;73:414-418. [6] Barrows DM. Functional capacity evaluations of persons with chronic fatigue immune dysfunction syndrome. Am J Occup Ther 1995;49:327-337. [7] Cammermeyer M, Evans JE. A brief neurobehavioural exam useful for early detection of postoperative complications in neurosurgical patients. J Neurosci Nurs 1988;20:314-323. [8] Weiss P. The occupational therapist's role in assessment of patients with suspected normal pressure hydrocephalus as predictor of good shunt outcome. Isr J Occup Ther 1994;3:E33-E42. [9] Margolin DI. Cognitive neuropsychology in clinical practise. New York: Oxford University Press, 1992. [10] Logue PE, Tupler LA, D'amico C, Schmitt FA. The neurobehavioural cognitive status examination: psychometric properties in use with psychiatric inpatients. J Clin Psychol 1993;49:80-89.
186 [11]
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
Mitrushina M, Abara J, Blumenfeld A. Cognitive screening of psychiatric patients. J Psychiatr Res 1995;29:13-22. [12] Osato SS, Yang J, La Rue A. The neurobehavioural cognitive status examination in an older psychiatric population. Neuropsychiatry, Neuropsychol Behav Neurol 1993;6:98-102. [13] Oehlert ME, Hass SD, Freeman MR, Williams MD, Ryan JI, Sumeralf SW. Accuracy: the neurobehavioural cognitive status examination's screen-metric approach. Rehabil Psychol 1996;41:162 (abstarct).
[14]
Lezak M. Neuropsychological assessment. 3rd ed. New York: Oxford University Press, 1995. [15] Katz N, Elazar E, Itzkovich M. Validity of the neurobehavioural cognitive status examination (Cognistat) in assessing CVA patients and healthy elderly in Israel. Isr J Occup Ther 1996;5:EI85-198. [16] Cummings JL, Benson DF. Dementia: a clinical approach. Boston: Butterworth-Heinemann, 1992. [17] Itzkovich M, Elazar B, Averbuch S, Katz N. Loewenstein occupational therapy cognitive assessment (LOTCA) manual. New Jersey: Maddak, 1990.
ELSEVIER
NeuroRehabilitation 9 (1997) 179-186
Comparison of cognitive status profiles of healthy elderly persons with dementia and neurosurgical patients using the neurobehavioural cognitive status examination Noomi Katz a,*, Adina Hartman-Maeir a Pnina Weiss b , Nira Armonc a School of Occupational Therapy,
Hebrew University Jerusalem, Mount Scopus, PO Box 24026, Jerusalem 91240, Israel bBelinson Medical Center, Petach Tikua, Israel CDay Center, Talpiot, Jerusalem, Israel
Abstract
The purpose of this study was to examine and compare the cognitive profiles of three groups of Israeli individuals who exhibit cognitive deficits for diverse reasons, using the Neurobehavioural Cognitive Status Examination (Cognistat, former name NCSE). The test is a standardized cognitive screening instrument which includes general areas of alertness, attention, orientation, language (comprehension, repetition, naming), construction, memory, calculations and reasoning (similarities, judgement). Assessment results are provided as a profile of the different domains and graded on four performance levels (average, mild, moderate, severe). Subjects included 47 healthy independent elderly subjects, 47 neurosurgical patients and 42 persons suffering from dementia. Supporting the hypothesized direction, statistically significant differences were found among the three groups on raw scores of all individual subtests and on the four performance levels with the healthy elderly subjects showing the highest performance and persons with dementia the lowest. Construction subtest scores were low for all groups and seemed to detect the aging process as well as disease-related dysfunction. The mean scores and standard deviations of all groups for most subtests were similar or a little lower to those reported in the American standardization data suggesting the test's cross-cultural applicability. In conclusion, the Cognistat was found to be a useful screening test for clinical and research purposes, however, further research is required. © 1997 Elsevier Science Ireland Ltd.
Keywords: Cognitive assessment; Cognistat (NCSE); Elderly; Dementia; Neurosurgical
1. Introduction
The purpose of this study was to examine and compare the cognitive profiles of three groups of
* Corresponding author. Tel.: 5324985.
+ 972 2 5813246; fax: + 972 2
individuals who exhibit cognitive deficits for diverse reasons using the Neurobehavioral Cognitive Status Examination (Cognistat, former name NCSE) [1]. The Neurobehavioural Cognitive Status Examination (Cognistat) was developed as a screening test for assessing cognitive status across a variety of domains [2]. The Cognistat is a short battery of
1053-8135/97/$17.()() © 1997 Elsevier Science Ireland Ltd. All rights reserved. PH S1053-8135(97)00028-0
180
N. Katzetal. /NeuroRehabilitation9 (1997) 179-186
subtests in five cognitive areas: Language (Comprehension, Repitition, Naming); Construction, Memory, Calculation and reasoning (Similarities and Judgment). The level of consciousness, Orientation and Attention areas are evaluated and scored before assessing the other cognitive abilities. In each sub test an initial screen item is given and only if the subject fails, other items are given (metric) in order to establish the level of competence. The results are presented in a profile which is divided into four performance levels: average performance, mild impairment, moderate and severe impairment. The Cognistat was found to be more sensitive in the detection of cognitive dysfunction in 30 patients with known brain lesions than two known mental status examination instruments (MMSE and CCSE) [3]. The Cognistat was also found to be a rapid and sensitive measure of cognitive function, predicting functional changes as a result of inpatient stroke rehabilitation [4]. It was further found to significantly differentiate between stroke patients and orthopedic control patients [5]. A comparison between the two unilateral (right vs. left) stroke groups revealed different patterns of cognitive deficits; yet this difference did not reach statistical significance, probably because of the small sample size of each group (n = 12) [5]. Barrows [6] presents the main cognitive dysfunction of persons with Chronic Fatigue Immune Syndrome using the Cognistat profile which revealed impairments in Memory, Construction, Calculations and to a lesser degree in Attention. According to the author, the Cognistat scores corroborated clients' perceptions of their impairments. The Cognistat was also used to detect cognitive changes following neurosurgery [7]. Among 11 patients who were evaluated pre- and post-surgery, six improved but five showed a deterioration which led in four cases to a second operation. The authors suggested that the Cognistat was useful for serial testing, enabling the comparison of profiles where patients act as their own control. The same approach using the Cognistat to monitor change before and after continuous drainage to predict shunt outcome is presented by Weiss [8].
Margolin [9] showed the use of the Cognistat for documenting the pattern and temporal course of cognitive decline in dementia patients. Comparing 13 dementia patients (mean age 71.3) to normal control patients (mean age 69.8) significant differences were found on all subtests excluding Attention and Judgment. The most severe deficits were seen in Memory and Orientation sub tests which are congruent usually with the earliest and most consistent impairments in dementia. A moderate impairment was also found on the Construction sub test which is also congruent with evidence that Performance IQ is more impaired than Verbal IQ [9]. In a large study performed by Logue et al. [10] to look at the psychometric properties of the Cognistat with psychiatric inpatients. Eighthundred and sixty-six patients aged between 15 and 92 years were tested at admission and showed a poorer performance than the normative sample on all subtests. Age was negatively correlated, especially with performance on Construction ( - 0.46), Memory (- 0.31) and Similarities ( - 0.31) subtests, which was in the expected direction. However, intercorrelations among subtests did not yield the predicted pattern of high correlations between similar domains and low correlations between dissimilar ones. Therefore, they recommended that further research needs to be carried out. One major critical point that was raised by the authors relates to the screen item of the Construction subtest which is loaded with the Memory component (largest correlation Construction and Memory r = 0.48) and therefore, they recommend a change in this item. Further studies of psychiatric patients found that the Cognistat differentiated between patients with and without organic mental disorder (OMD) [11] and between OMD and elderly depressed inpatients [12]. Some Memory deficits were found in all patients, but the best predictor of OMD using multiple regression was the Construction subtest followed by the Memory subtest. In the geropsychiatric study [12] the Cognistat was found to be more sensitive to organically based cognitive impairment then the MMSE. The authors used a criterion of two or more scales below the
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
cutoff, which maximizes sensitivity, as being suggestive of cognitive deficits [12]. The issue of the accuracy of the screen vs. metric approach was recently studied in a sample of 95 male clients (mean age 66, with a mean of 11 years of education) [13]. They found that many subjects who passed the screen item failed on the metric part of the same domain, thus producing false negative results. Therefore, the authors suggest that all items (screen and metric) should be tested to reduce the possibility of missing deficit areas when they exist. It seems to us that the above findings need to be further investigated, especially checking the different screen items, as the general conclusion suggested above eliminates one of the benefits of the Cognistat. In summary, the Cognistat instrument has been utilized with a variety of patient populations where brain dysfunction was suspected and initial results showed that it was sensitive in detecting cognitive impairments, differentiating between groups as well as measuring changes over time. Some caution was expressed by Lezak [14] regarding the sensitivity of the test to mild head trauma because they found that in using the test they had to supplement the Attention and Language tasks with few more difficult items in order to assess subtle deficits in these areas. However, any instrument developed in one country when used cross-culturally needs to be further validated to ascertain its applicability. Therefore, the Cognistat Test Booklet was translated into Hebrew with minor cultural changes such as adaptations of some of the Judgment subtest situations to Israeli conditions, or few changes in the screen item of the Naming subtest (the pen parts). The subtests were administered in strict accordance with the test manual. The Hebrew translation is used in clinics by a growing number of therapists. The first study in Israel comparing 15 cerebrovascular accident (CVA) elderly patients to 24 healthy elderly showed that healthy subjects mean scores were higher than CVA patients on all subtests except for Construction. Significant statistical differences were found for Orientation and three Language subtests [15]. Based on this initial study it was recommended to
181
further investigate the test's validity with additional populations and larger samples. Hence, the specific objectives of the present study were twofold: (1) to compare the cognitive profile of three groups, healthy elderly, neurosurgical patients and persons suffering from dementia; and (2) to compare the Israeli results to the American standardization data from the Cognistat manual. It was hypothesized that significant differences will be found among the three groups on all subtests, where the healthy group will show an average profile and clients with dementia the most severe deficits. 2. Method
2.1. Subjects and procedure Three groups of subjects were tested with the Cognistat: (1) 47 healthy independent elderly subjects, 13 men and 34 women (mean age 74.75, S.D. = 5.44); (2) 47 neurosurgical patients suffering from tumors, trauma, hydrocephalus, or cerebral hemorrhage, who were tested prior to surgery, 26 men and 21 women (mean age 54.7, S.D. = 21.3); and (3) 42 clients suffering from DAT (Dementia Alzheimer Type), 20 men and 22 women (mean age 77.1, S.D. = 7.56). The healthy elderly participants were recruited from independent living settings and had no known neurological or psychiatric illness. Neurosurgical patients were tested consecutively as inpatients as part of their general evaluation prior to surgery. Clients suffering from DAT were tested at a Day Center as part of their evaluation for community intervention planning. All subjects gave their written informed consent to be evaluated as part of their admission to the hospital or day treatment. The data was analyzed by a one-way ANOVA for a general F value between the three groups on each subtest followed by a post-hoc Scheffe test for the source of significant differences among groups. Chi-square analysis was used to compare frequencies of performance levels between the two patient groups.
N. Katz et al. / NeuroRehabilitation 9 (J997) 179-186
182
3. Results 3.1. Demographics
Demographic variables of age and gender were analyzed in relation to the Cognistat score. No significant correlations with age were found among the two elderly groups (Healthy and DAT) as they were very homogenous elderly groups. Within the neurosurgical group, significant negitive correlations with age were found for Construction (r = - 0.60), Memory (r = - 0.54) and Similarities (r = -0.54) all at P < 0.0001, Comprehension (r = -0.31, P < 0.04) and repitition (r = - 0.37, P < 0.01). The relationship with age was in the expected direction as older subjects performed lower. The only gender differences were also found in the neurosurgical group. Wilcoxon non-parametric analysis showed significant differences around P < 0.05-0.01 for Orientation, Attention, Construction and Judgment where men scored higher then women and Repitition where women scored higher, P < 0.02. These results were not expected and seem unclear and inconsistent. 3.2. Cognitive status
Cognistat mean scores and standard deviations
for the three groups are presented in Table 1 along with F values and post hoc Scheffe results. Significant F values were found for all subtests. Post-hoc Scheffe comparisons of means showed significant differences between healthy subjects and subjects with dementia on all subtests. Healthy and neurosurgical patients differed significantly on six subtests, excluding three Language tests where both groups performed in the average range and Construction on which both performed low, therefore the difference did not reach a significant level. In Fig. 1, the profiles of the three group means are depicted along the four performance levels. It can be seen that persons suffering from DAT perform in the mild to severe impairment ranges with Construction, Memory, Naming and Similarities being the most severely impaired. Neurosurgical patients' performance level is in the low average to mild impairment range for most subtests and in the mild to moderate impairment range for Construction, Memory and Similarities. Healthy elderly subjects perform at the average range on all subtests, except for Construction and Memory subtests where average extends to the mild impairment range, as specified for the normal elderly population over age 65 in the test manual.
Table 1 Cognistat mean scores of three groups: healthy elderly (n = 47), DAT' (n = 42) and neurosugical patients (n = 47), and one-way ANOVA between groups Group M (S.D.) Cognistat subtest
Healthy
DAT'
Neurosurgical
F (P)
Scheffe b
Orientation Attention Comprehension Repetition Naming Construction Memory Calculation Similarities Judgment
11.91 (0.28) 6.91 (1.50) 5.53 (0.99) 10.91 (2.72) 7.19 (2.18) 3.17(2.31) 8.45 (3.57) 3.85 (0.46) 5.740.61) 5.06 (0.99)
7.57 (3.51) 4.69 (1.63) 4.48 (1.36) 8.46 (3.74) 4.0 (1.78) 2.0 (1.34) 5.05 (3.71) 3.05 (1.08) 3.15 (2.35) 3.30.62)
8.85 (2.93) 5.91 (2.47) 5.02 (1.45) 10.33 (4.13) 8.0 (2.15) 2.6 (2.26) 6.4 (4.49) 2.89 (1.32) 3.89 (2.07) 4.05 (1.60)
33.11 (0.0001) 14.77 (0.0001) 7.53 (0.0008) 5.37 (0.006) 46.13 (0.0001) 3.68 (0.03) 8.43 (0.0004) 11.99 (0.0001) 19.43 (0.0001) 17.04 (0.0001)
H-D; H-N H-D; H-N; D-N H-D H-D H-D;D-N H-D H-D;H-N H-D; H-N H-D;H-N H-D; H-N
aDA T, Dementia Alzheimer's Type. b Indicates the groups that scores were significantly different from each other: H, healthy; D, DAT; N, neurosurgical.
183
N. Katzetal. /NeuroRehabilitation9 (1997) 179-186
Average
~
-----.... ~ .....
- - - "" .... , ,
..
Mild
,,
~ \
,
..
,
Moderate
...
\ .
/
~ V.; '\ / / ,,/ "\.':" \\ V . . "',' ;'", , ,'", \ .... .. , . ,\.
,
... , . .,;
11..
o
11.. W
::&
oo
t/)
Z
o
1- Healthy - -
,,
o..J ct o
:E w :E
l-
0::
,
,
Severe ~
./
,
o
C
:J ...,
Neuro . - .. OAT I
Fig. 1. Cognistat profiles of three groups: healthy elderly; OAT and neurosurgical patients.
Mean scores of neurosurgical patients and persons with DAT differed significantly on Attention and Naming subtests only, although subjects with DAT performed lower on all subtests excluding Calculations. However, looking at mean scores does not show an accurate picture of individual subjects' performance. Therefore, frequencies for
the two groups according to the four performance levels are presented in Table 2 along with results of Chi-square analysis. Naming and Construction subtests highly differentiated between the groups followed by Judgment, Attention and Repitition. It can be seen that neurosurgical patients show higher percentages of average scores on almost
Table 2 Frequency (percentages) of four performance levels on each subtest for DATa (n chi-square analysis between groups Cognistat subtest
Orientation Attention Comprehension Repetition Naming Construction Memory Calculation Similarities Judgment
2
3
4
=
47), and
11.9 50 9.5 16.7 23.8 31 11.9 26.2 19 31
19 19 16.7 14.3 54.8 23.8 11.9 9.5 4.8 9.5
28.6 2.4 11.9 38.1 11.9 33.3 59.5 0 47.6 21.4
53.2 61.7 70.2 63.8 80.9 40.4 34 57.4 38.3 66
2
3
4
X 2 (p)
17 21.3 2.1 10.6 8.5 4.3 6.4 21.3 14.9 2.1
12.8 12.8 14.9 2.1 8.5
17 4.3 12.8 23.4 2.1 48.9 46.8 12.8 36.2 17
NS 11.3 (0.01) NS 11.3 (0.01) 45.9 (0.000) 21.9 (0.000) NS NS NS 15.7(0.001)
a OAT, Dementia Alzheimer's Type. average range; 2, mild impairment; 3, moderate impairment; 4, severe impairment.
b 1,
42) and neurosugical patients (n
Neurosurgical group performance level
OAT" group performance levels
40.5 28.6 61.9 31 9.5 11.9 16.7 64.3 28.6 38.1
=
6.4 12.8 8.5 10.6 14.9
184
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
all subtests except for Calculations. Subjects with dementia show higher percentages of severe impairments on six subtests (Orientation, Repitition, Naming, Memory, Similarities and Judgment). For persons with DAT, Calculations is the most preserved cognitive domain, it is the only subtest where the mean score was in the average range with 64% average and 26% mild impairment.
3.3. Comparison to standardization data A descriptive comparison of mean scores of the groups in this study to the American standardization data showed that on six subtests the Israeli means are a little lower than the geriatric healthy and the neurosurgical groups. The Construction subtest showed the largest difference for both groups, while Comprehension, Calculations, Similarities and Judgment showed similar means. The profile of the persons with DAT appears to be within the range shown in the Cognistat manual [1]. The three Language subtests scores are a little lower than shown in the typical profile. Memory was found to be the most impaired domain of the DAT subjects as expected, followed by Orientation and Similarities. 4. Conclusions In conclusion, the results of this study show that the cognitive profiles of the groups differed significantly in severity of deficits in the hypothesized direction, supporting the validity of the Cognistat. The instrument was found as a useful screening test applicable also to the Israeli population. Mean scores of the Israeli subjects are similar to the American standardization samples for the comparable groups. However, data to compare healthy adult groups in Israel has yet to be collected. The differences in age among this study groups, especially the younger neurosurgical group compared to the healthy elderly may have limited the gap between them. It is assumed that a healthy adult group similar in age to the patient group may have shown a larger profile difference. The gender differences found within the neurosurgical group were not expected and are not
clear. It seems an inconsistent result which may be related to the specific group studied. Further studies will have to examine if a gender relationship exists, if it does, then in which direction and in what domains. Cross-cultural validation is an essential step in using tests developed in another culture. In adapting a test, care needs to be taken to make as few changes as possible so that comparison to the original data is meaningful. On the other hand, test items should be appropriate to use in the new culture. For example, one of the problems we faced was whether some of the items from the Naming subtest are familiar, such as the Octopus or Xylophone, especially in the elderly population and in less educated populations. We have not changed them because this would mean changing actual test items which requires a more elaborate process of test validation. The higher results on the Naming subtest of the younger neurosurgical group compared to the healthy elderly group may reflect the above mentioned concern. However, the comparison of means of both groups to the American data shows similarities, thus, suggesting that the difficulties in the Naming items may be similar in both countries and do not reflect a cultural issue. One general limitation of the instrument is that it is highly verbal. The Construction subtest is the only performance subtest. Visuospatial tests are considered excellent screening devices as individuals with brain dysfunction involving frontal, parietal or occipital lesions usually fail on copying or drawing two and three-dimensional tasks [14,16]. Deficits on these tests indicate organic disturbance or decline and usually don't differentiate hemispheric involvement. According to Osmon et al. [5] the analysis of intercorrelations among Cognistat subtests and, more specifically, the Construction items showed that the screen item and the last item are too difficult, making the overall scoring of Construction too sensitive and, therefore, low for all groups. This was seen in the present study as the Construction sub test was difficult for most subjects, which indicates the sensitivity of visuospatial deficits to organic disturbance as well as in normal aging [9,12,16]. In addition, the Construction subtest involves visual
N. Kim et al. / NeuroRehabilitation 9 (1997) 179-186
Memory in the screen item making it less accurate as a constructional measure, as stated by Osmon et al. [5]. The assessment of Construction, which is critical for screening brain dysfunction and functional performance, is limited in the Cognistat. Therefore, it was recommended for example by Katz et al. [15] to use the visuomotor organization subtests of the Loewenstein Occupational Therapy Cognitive Assessment (LOTCA) [17] for further in-depth assessment when the Cognistat subtest shows constructional deficits. In general, the Cognistat appears to be a useful screening test for clinical and research use in a variety of client populations based on this research and on previous works reviewed. In the rehabilitation process, the Cognistat profile provides initial information on intact and impaired cognitive skills and may determine which cognitive areas require further, in-depth evaluation and/or intervention. In the neurosurgical group initial results before surgery are compared to a second testing following a period of recovery after surgery. Comparison of the clients' two profiles may suggest the need for a second operation, or areas which still show deficits and require cognitive intervention that may have functional implications in daily activities. It also shows the intact areas which rehabilitation professionals should build on in further intervention using the clients' abilities. This approach, taking into account intact abilities to build on, is imperative in the DAT group as they suffer from a declining process and rehabilitation is focused on maintaining function as long as possible. The results of the two studies conducted in Israel have direct implication for clinicians and researchers in Israel, suggesting that it is valid to use the Cognistat as a clinical and research tool. However, as mentioned earlier, additional data from healthy adult populations are needed to compare results to similar groups in the standardization study. Initial results imply that the Cognistat can be used as a measure of change in intervention studies, such as following surgical or other procedures, but further research needs to be carried out. It is also important to study more closely the influence of education on the performance of the test's cognitive domains. Cognitive
185
performance is known to relate to education and it may distort the interpretation of individuals results. The authors acknowledge this problem in their 'cautions section' [1], however, it would be important to have data about the relationship of education to specific subtests. Acknowledgement
The Cognistat Test Booklet was translated into Hebrew by Rina Kravitz. We would like to thank the Occupational Therapists: Shulamit Avni, Dorit Haldman, Noga Ziv and Debra Roitman for collecting the data within the healthy group. References [1] Cognistat - The Neurobehavioural Cognitive Status Examination in older psychiatric population. Fairfax CA: The Northern California Neurobehavioural Group, 1995. [2] Kiernan RJ, Mueller J, Langston JW, Van Dyke C. The neurobehavioural cognitive status examination: a brief but differentiated approach to cognitive assessment. Ann Intern Med 1987;107:481-485. [3] Schwamm LH, Van Dyke C, Kiernan RJ, Merrin EL, Mueller J. The neurobehavioural cognitive status examination: comparison with the cognitive capacity screening examination and the mini-mental state examination in a neurosurgical population. Ann Intern M ed 1987;107:486-490. [4] Mysiw WJ, Beegan JG, Gatens PF. Prospective cognitive assessment of stroke patients before inpatient rehabilitation. Am J Phys Med Rehabil 1989;68:168-171. [5] Osmon DC, Smet JC, Winegarden B, Gandhavadi B. Neurobehavioural cognitive status examination: its use with unilateral stroke patients in a rehabilitation setting. Arch Phys Med RehabilI992;73:414-418. [6] Barrows DM. Functional capacity evaluations of persons with chronic fatigue immune dysfunction syndrome. Am J Occup Ther 1995;49:327-337. [7] Cammermeyer M, Evans JE. A brief neurobehavioural exam useful for early detection of postoperative complications in neurosurgical patients. J Neurosci Nurs 1988;20:314-323. [8] Weiss P. The occupational therapist's role in assessment of patients with suspected normal pressure hydrocephalus as predictor of good shunt outcome. Isr J Occup Ther 1994;3:E33-E42. [9] Margolin DI. Cognitive neuropsychology in clinical practise. New York: Oxford University Press, 1992. [10] Logue PE, Tupler LA, D'amico C, Schmitt FA. The neurobehavioural cognitive status examination: psychometric properties in use with psychiatric inpatients. J Clin Psychol 1993;49:80-89.
186 [11]
N. Katz et al. / NeuroRehabilitation 9 (1997) 179-186
Mitrushina M, Abara J, Blumenfeld A. Cognitive screening of psychiatric patients. J Psychiatr Res 1995;29:13-22. [12] Osato SS, Yang J, La Rue A. The neurobehavioural cognitive status examination in an older psychiatric population. Neuropsychiatry, Neuropsychol Behav Neurol 1993;6:98-102. [13] Oehlert ME, Hass SD, Freeman MR, Williams MD, Ryan JI, Sumeralf SW. Accuracy: the neurobehavioural cognitive status examination's screen-metric approach. Rehabil Psychol 1996;41:162 (abstarct).
[14]
Lezak M. Neuropsychological assessment. 3rd ed. New York: Oxford University Press, 1995. [15] Katz N, Elazar E, Itzkovich M. Validity of the neurobehavioural cognitive status examination (Cognistat) in assessing CVA patients and healthy elderly in Israel. Isr J Occup Ther 1996;5:EI85-198. [16] Cummings JL, Benson DF. Dementia: a clinical approach. Boston: Butterworth-Heinemann, 1992. [17] Itzkovich M, Elazar B, Averbuch S, Katz N. Loewenstein occupational therapy cognitive assessment (LOTCA) manual. New Jersey: Maddak, 1990.
