Age and Ageing Advance Access originally published online on June 28, 2006
Age and Ageing 2006 35(5):503-507; doi:10.1093/ageing/afl058
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Alzheimers Quick Test cognitive screening criteria for West African speakers of Krio
1 Department of Psychiatry, Hvidovre Hospital, Vesterbrogade 31, Copenhagen, Denmark
2 Knowledge Research Institute, Inc., 7101 Lake Powell Drive, Arlington, TX 776016-3517, USA
Address correspondence to: Elisabeth H. Wiig. Tel: (+817) 572 6254. Fax: (+817) 478 1048. Email: ehwiig{at}krii.com
| Abstract |
|---|
|
|
|---|
Objectives: to obtain normative data for Alzheimers Quick Test (AQT) measures of perceptual and cognitive speed from West African speakers of Krio.
Subjects: normal adults, who were functionally independent, from Sierra Leone (n = 164) aged 2579 years.
Methods: perceptual and cognitive speed were measured with AQT single- and dual-dimension colournumber (CN) and colouranimal (CA) naming tasks. Tests were administered individually in the participants communities.
Results: men and women performed similarly (P>0.05), whereas literate speakers used significantly less time than preliterate peers (P<0.01). Correlations between age and colour naming were low (P<0.01) and speed decreased by <0.1 s per year. Dual-dimension naming remained stable across ages. Correlations with years of education were low for dual-dimension naming (P<0.01) and speed increased
0.4 s per added year. Correlations between age and education and AQT naming were non-significant for literate participants. Criterion time cut-offs (seconds) for screening were developed for preliterate and literate speakers of Krio for typical (<+1 SD), slower-than-typical (between +1 and +2 SD) and atypical (>+2 SD) performance.
Conclusion: AQT CN and CA naming are time efficient (35 min each), objective and reliable and can be administered in Krio to West African adults in Africa, Europe or North America to screen for cognitive impairments and facilitate referral for medical workup.
Keywords: cognitive processing speed, attention/executive, set shifting, screening criteria, elderly
| Introduction |
|---|
|
|
|---|
With rapid global growth in elderly populations, it is important to find behavioural measures to differentiate normal ageing from cognitive impairments caused by disease processes. This is challenging, as normal performance varies because of hereditary, developmental, educational or environmental factors [1, 2]. There is limited information about normal cognitive ageing [37] and none for African adults. This is in spite of the estimates that the African population over the age of 65 years would grow from 17.7 to 37.9 million from 1997 to 2025 [8]. This study provides measures of normal cognitive ageing in West African Krio speakers, obtained with simple, objective and reliable processing-speed tests with broad cultural/linguistic application [9].
Tests of content are typically used to screen for cognitive impairment and decline associated with neurological conditions. However, they introduce cultural, linguistic and/or educational biases, when used outside Western cultures for which they were developed. The common measure, Mini-Mental State Examination (MMSE) [10, 11], shows limitations in differentiating normal function from Alzheimers disease (AD) [12, 13]. Thus, MMSE showed specificity of 100% and sensitivity of 84%, whereas Alzheimers Quick Test (AQT) colourform (CF) naming [9] showed specificity of 97% and sensitivity of 97%. Educated Westerners readily perform the MMSE tasks, but abilities required for subtraction, reading and writing might not be developed in societies where literacy is not generally established. It is, therefore, of importance for the developing societies to identify cognitive-screening measures that are quick, reliable, objective and cross-culturally applicable, and can be administered with minimal training. These considerations prompted the collection of normative data for AQT colournumber (CN) and colouranimal (CA) naming [9] from functionally intact West African speakers of Krio.
AQT assesses processing speed (i.e. time to complete a controlled-input task) with rapid naming. Processing-speed tests are sensitive to small changes in time, and older adults typically use longer time than younger adults [1416]. Similarly, AQT dual-dimension (CN and CA) naming shows significant, but small, time increases with age for American adults below the age of 60 years (1 s per decade) and slightly larger increases after age 60 (1 s per seven years) [5]. Naming times are not affected by gender [9, 17].
The AQT naming tasks are criterion-referenced for normal adults in the US, Sweden and Greece, and cut-offs (seconds) for typical, slower-than-typical and atypical naming speeds are reported [9, 17, 18]. Testretest reliability is high, and effects of age on naming are similar for American and Swedish speakers [9, 17]. Speakers of Greek require a higher cut-off point (100 s) for atypical CF naming than American and Swedish speakers (
+30 s) [18].
Sustained naming causes redistribution of cerebral blood flow. During AQT CF naming, normal adults show increased blood flow to temporalparietal brain regions bilaterally and suppressed flow to frontal regions compared with the rest [9, 17, 19, 20]. CF naming shows high sensitivity (93.198.7%) and specificity (96.699.9%) and differentiates Swedish and Greek adults with mild-to-moderate AD from cognitively normal peers [17, 18], and in differentiating AD from dementia with Lewy bodies [21]. AQT dual-dimension naming assesses attention/executive functions mediated by temporalparietal activation [2224].
This study was descriptive and objectives were pragmatic. They were as follows: (i) to obtain normative data for AQT CN and CA naming from West African speakers of Krio, (ii) to compare naming times and effects of ageing for Krio and American speakers and (iii) to develop culturally/linguistically appropriate criterion cut-offs (seconds) for typical, slower-than-typical and atypical performance.
| Methods |
|---|
|
|
|---|
Participants
There were 59 women and 105 men (mean age 44.6 years, SD 9.5). All gave informed consent in accordance with the Declaration of Helsinki. A well-being questionnaire about patterns of eating, sleeping, working and mood and absence of fever, neurological disorders, diabetes and HIV screened participants. Adults with neurological disorders or HIV, or who did not meet the criteria for well-being, were excluded. Participants were recruited from the members of the Lutheran community, and each was paid $5.00 by the authors. All were natives of Sierra Leone, equally distributed between urban (Freetown) and rural areas. Participants completed between 2 and 17 years of education (M 10 years, SD 4 years). A total of 45 participants had recurring malaria and 119 were malaria free. Participants included housewives, day workers, traders and professionals, who led functionally independent lives in which they shopped, held jobs and performed societal responsibilities. Owing to their roles in the society, all were presumed to function within the normal cognitive range and to be free of cognitive impairments or dementia that would interfere with their work or societal functions.
Materials and administration
A trained examiner, native of Sierra Leone, administered the tests individually in Krio, the lingua franca for trading and social interactions, which blends English, French and Portuguese. Participants completed testing in one session in their community settings, and short trials established adequacy of naming test stimuli.
AQT CN and CA each consists of three tests. Test 1 requires naming of 40 squares (e.g. black, blue, red or yellow) and Test 2 naming either 40 numbers (e.g. 2, 4, 5 and 7) or animals (e.g. bird, cat, fish, rat, snake and spider) rendered in black. Tests 1 and 2 measure perceptual speed (i.e. perception + response time). Test 3 requires naming of 40 CN (e.g. red circle and blue square) or CA (e.g. yellow fish and black spider) combinations and measures cognitive speed (i.e. perceptual speed + cognitive overhead). Testretest reliability (r) is high for CN (0.91) and CA naming (0.96).
Naming times for each test were measured digitally, beginning at voice onset, and recorded in seconds and in fractions of seconds. Data from 144 American speakers with a mean age of 39.6 years (SD 15.9) were used for comparison. One-way ANOVA indicated no significant differences in age between Krio and American speakers (F = 3.20; P>0.05).
Statistical analyses
Naming times (seconds) were used in the analyses for the total group (n = 164), men (n = 105) and women (n = 59), preliterate (n = 41) and literate (n = 123) speakers and speakers with (n = 45) and without (n = 119) malaria. The analyses (SPSS Version 11.0.2 for Macintosh) produced descriptive statistics, one-way and univariate ANOVA comparisons, correlations (r) and linear regression coefficients (ß). Measures of variability (SD) and cumulative frequencies were used to establish criterion cut-off times (seconds).
| Results |
|---|
|
|
|---|
Background variables
AQT naming times (seconds) for Krio speakers are summarised in Table 1. Single-dimension naming times are about one-third shorter than the dual-dimension times. Means for men and women did not differ significantly for colour, number, animal or CA naming (F = 0.962.37; P>0.05). CN naming was significantly faster among women than men (F = 6.55; P<0.01).
|
Mean time differences between (i) preliterate (n = 41) and literate adults (n = 123) and (ii) adults with (n = 45) and without malaria (n = 119) were tested. One-way ANOVA with literacy as a factor (i.e. education above age 15/Grade 8) [25] indicates significantly shorter CN (M = 55.30, SD 20.37) and CA (M = 76.14, SD 23.52) times for literate than for preliterate speakers (M = 73.10, SD 17.67; 82.98, SD 14.36) (F = 8.19 and 12.01; P<0.01). Naming times by literate Krio (n = 123) and American speakers (n = 144) were compared. Means (seconds) for Americans were as follows: colour20.91 (SD 3.31), number14.26 (SD 2.72), CN41.04 (SD 6.75), animal25.74 (SD 3.75) and CA48.29 (SD 7.18). All differences were significant (P<0.01) (F = 43.09, 21.65, 11.08, 44.30 and 36.85). Univariate ANOVA with age and years of education as independent variables tested interaction effects. For CN, age was not significant (F = 1.32), whereas education (F = 2.24; P<0.05) and the ageeducation interaction were significant (F = 0.81; P>0.05). For CA, the effects for age (F = 1.52), education (F = 1.23) and interaction (F = 1.11) were all non-significant (P>0.05).
Mean differences between speakers with (n = 45) and without (n = 119) malaria were significant (P<0.05) for colour, CN and CA naming (F = 5.66, 20.63 and 9.50). Adults with malaria completed 12.0 (SD 2.5) and adults without malaria completed 9.8 (SD 3.7) years of education, which is a significant difference (F = 11.67; P<0.01).
Age and education effects
In the total group (n = 164), correlations (r) between age and colour and animal naming were significant (0.23 and 0.17; P<0.05), whereas correlations with number, CN and CA were not (0.09, 0.15 and 0.03; P>0.05). Significant relationships between age and naming times were analysed using linear regression (Table 2). Coefficients for age effects for colour and animal naming indicated increases (slowing of speed) of 0.05 and 0.08 s per year. Age accounted for 5% of the variance in colour and 3% in animal naming time.
|
Correlations (r) were significant between education and colour, CN and CA naming (0.19, 0.38 and 0.28; P<0.05), whereas correlations with number and animal naming were not (0.03 and 0.13; P>0.05). Linear regression (Table 2) indicated decreases in time (increased speed) of 0.14 s for colour, 0.43 s for CN and 0.46 s for CA naming per added year of education. The relationships were described by the following equations: CN = 82.96 + (2.25) (education) and CA = 94.97 + (1.67) (education). Education accounted for 3% of the variance in colour, 14% in CN and 8% in CA naming time. In the previous research, education after the age of 15 years (Grade 8) did not affect AQT dual-dimension naming times [1, 19]. Therefore, correlations (r) were calculated between years of education and CN and CA naming for literate Krio speakers (n =123). There was no significant correlation for CN (r = 0.14; P>0.05), but there was for CA naming (r = 0.29; P<0.01).
Criterion cut-offs
Normative ranges for typical, slower-than-typical and atypical naming times for literate (n = 123) and preliterate (n = 41) Krio speakers were developed by previously employed methods [9, 17] (Table 3). Times (seconds) shorter than +1 SD above the mean indicated typical, longer than +2 SD above the mean indicated atypical (potentially pathological) and between +1 and +2 SD indicated slower-than-typical performance. Cut-off times were adjusted to lend parsimony and ease of reference for screening.
|
Cut-off criteria for literate Krio speakers were compared with criteria for literate American and Swedish speakers for CN (typical <50 s; slower 5159 s; atypical >60 s) and CA naming (typical <60 s; slower 6169 s; atypical <70 s) [9, 17]. This indicated a difference in cut-off criteria of 2530 s for CN and 4050 s for CA naming.
| Discussion |
|---|
|
|
|---|
Normative naming times were obtained from Krio speakers for AQT CN and CA naming, and criterion cut-offs were developed for potential cognitive screening. CF naming, explored in clinical research [9, 13, 1720], was not administered because of rare use of names for geometric forms in West African daily life. The study assumed a macro-perspective where dual-dimension naming times were considered symptomatic of broader attention/executive functions (e.g. attention, working memory and set shifting) [20, 2224].
Women named the CN stimuli on average 10 s faster than men, a counterintuitive finding, as men tend to be the traders. More importantly, men and women used similar times for dual-dimension naming, concurring with the findings for American and Swedish speakers [1, 5, 17]. Age did not significantly affect dual-dimension naming either.
Education (literacy) influenced CN, but not CA naming speed. This seems appropriate, as colour naming and number naming are part of the curriculum in West African schools, and longer exposure should result in greater cognitive speed. Interactions between age and education were not significant for dual-dimension naming. As a group, Krio speakers were significantly slower at naming both single- and dual-dimension stimuli than American speakers [5]. Even literate Krio speakers used considerably longer time than literate Americans to name stimulus combinations.
Age and education
Associations between age and naming times were low and only reached significance for perceptual speed measures. Age accounted for less variance in colour naming for Krio (5%) than for American speakers (8%) [5]. Associations between years of education and dual-dimension naming were low, but significant, and speed increased by about 2 s for each added year of education. Education accounted for relatively large proportions of the variance in CN and CA naming (15 and 8%, respectively).
The association between education and naming time was evident in adults with and without malaria. Adults with malaria, and more advanced education, used on average 10 s less for CN and 20 s less for CA naming than malaria-free adults. This supports that achievement of literacy results in greater cognitive flexibility. In the literate group, there was no significant association between education and CN naming. This concurs with the findings that education past established literacy has little effect on dual-dimension naming among American and Swedish adults [9, 17].
Implications for screening
Criterion cut-offs for AQT naming are identical for American and Swedish speakers [9, 17, 18, 20]. Because Krio shares features of English in content and structure, similarities in naming times would be expected. However, criterion cut-offs for the lower limits of atypical performance for Krio speakers are more than 30 s longer than cut-offs for American speakers [9]. The difference may reflect cultural traditions for rate of speech in West Africa or other uncontrolled factors.
Speed, ease of administration and equity of access are desirable screening-test qualities in societies with limited education. The advantages of using CN over CA naming with West African Krio speakers were therefore considered based on variability, biases and expressed preferences. CN and CA naming showed similar variability, but the majority preferred CN to CA naming. Although CN or CA naming may be used alone for screening, best clinical practice suggests administering both the tasks (57 min total) until clinical utility data become available. We also suggest repeating administration of dual-dimension tests with atypical naming times (>+2 SD) for consistency and validation.
Results suggest that AQT CN and CA can be used for first-line screening of Krio speakers in West Africa, Europe or North America to identify adults, who may require further evaluation for cognitive impairment and/or dementia. The tests are not intended for diagnostic purposes, because they are sensitive to cognitive impairments resulting from a range of neurological and psychiatric disorders. If naming time consistently falls in the atypical range, referral for medical workup is recommended. If naming time falls within the slower-than-typical range, periodic screening is recommended to identify if significant reductions in cognitive speed (+15 s or more), which may be indicative of a progressive disorder, have occurred.
Current health priorities in West Africa concern individuals who are HIV positive. Growth in the elderly population will, however, demand attention to detecting cognitive impairments, resulting from dementia or other causes. While prescription medications for AD are unavailable to most West Africans, the cultures have long histories of treating illness, including memory loss, with herbal constituents and compounds, justifying early identification.
| Key points |
|---|
|
|
|---|
- Normative data for AQT CN and CA naming were collected from 164 functionally intact speakers of Krio in Sierra Leone.
- Naming speed was significantly faster for literate than for preliterate speakers.
- Age had no significant effect on dual-dimension naming times.
- Years of education affected dual-dimension naming times in the total group, but not in the group of literate speakers.
- Criterion cut-offs were developed separately for screening literate and preliterate West African speakers of Krio.
| Conflicts of interest |
|---|
|
|
|---|
None.
| Acknowledgements |
|---|
|
|
|---|
We express our gratitude to the Evangelical Lutheran Church of Sierra Leone, its staff and members. We are especially indebted to Bishop Tom Barnett for supporting the study and to the Reverend Marie Barnett for collecting the data. This research received no third party funding.
| References |
|---|
|
|
|---|
- Institute for the Study of Aging & International Longevity Center-USA. Achieving and maintaining cognitive vitality with aging: workshop report. New York: ISOA & ILC-USA, 2001.
- Salthouse TA. Theoretical Perspectives on Cognitive Aging. Hillsdale, NJ: Erlbaum, 1991.
- Christensen H, Kumar R. Cognitive changes and the ageing brain. In: Sachdev PS, ed. The Ageing Brain: The Neurobiology and Neuropsychiatry of Ageing. Lisse: Swets & Zeitlinger; 2003, 7595.
- Fleischman DA, Wilson RS, Gabrieli JDE et al. A longitudinal study of implicit and explicit memory in old persons. Psychol Aging 2004; 4: 61725.
- Jacobson JM, Nielsen NP, Minthon L et al. Multiple rapid automatic naming measures of cognition: normal performance and effects of aging. Percept Mot Skills 2004; 98: 73953.[CrossRef][Web of Science][Medline]
- Price L, Said K, Haaland KY. Age-associated memory impairment of logical memory and visual reproduction. J Clin Exp Neuropsychol 2004; 26: 53138.[Web of Science][Medline]
- Small SA, Stern Y, Tang M et al. Selective decline in memory function among healthy elderly. Neurology 1999; 52: 13926.
[Abstract/Free Full Text] - Cumings JL.The Neuropsychiatry of Alzheimers Disease and Related Dementias. London: Taylor and Francis, 2003.
- Wiig EH, Nielsen NP, Minthon L, Warkentin S. Alzheimers Quick Test: Assessment of Parietal Function (AQT). San Antonio (TX): Harcourt/PsychCorp, 2002.
- Folstein MF, Folstein SE, McHugh PR. "Mini-mental state": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 18998.[CrossRef][Web of Science][Medline]
- Molloy D, Alemayehu E, Roberts R. Reliability of a standardized mini-mental state examination compared with the traditional mini-mental state examination. Am J Psychiatry 1991; 148: 1025.
[Abstract/Free Full Text] - Burns A, Lawlor B, Craig S. Rating scales in old age psychiatry. Br J Psychiatry 2002; 180: 1617.
[Abstract/Free Full Text] - Nielsen NP, Wiig EH, Minthon L, Warkentin S. Clinical utility of colour-form naming in Alzheimers disease: Preliminary evidence. Percept Mot Skills 2004; 99: 12014.[CrossRef][Web of Science][Medline]
- Goetz DW, Jacobson JM, Murnane JE et al. (1989). Prolongation of simple and choice reaction times in a double-blind comparison of twice-daily hydroxyzine versus terfenadine. J Allergy Clin Immunol 1989; 84: 31622.[CrossRef][Web of Science][Medline]
- Strauss E, Loring D, Chelune G et al. Predicting cognitive impairment in epilepsy: Findings from the Bozeman epilepsy consortium. J Clin Exp Neuropsychol 1995; 17: 90917.[Web of Science][Medline]
- Vendrell P, Junque C, Pujol J et al. The role of prefrontal regions in the Stroop task. Neuropsychologia 1995; 33: 34152.[CrossRef][Web of Science][Medline]
- Wiig EH, Nielsen NP, Minthon L, Warkentin S. Alzheimers Quick Test: Assessment of Parietal Function. Svensk Version & Norsk Versjon. Stockholm, Sweden: Psykologiførlaget, 2003.
- Warkentin S, Tsantali E, Kiosseoglou G et al. The AQT as a useful short screening test for dementia: Evidence from two European cultures. Alzheimers Assoc Int Conf on Prevention of Dementia. Washington, DC, 2005.
- Warkentin S. AQS assessment of temporoparietal lobe function. Scientific Presentation. Euro Union Consortium on Alzheimers Disease. Liege, Belgium, 2003.
- Wiig EH, Nielsen NP, Minthon L et al. Parietal lobe activation in rapid, automatized naming by adults. Percept Mot Skills 2002; 94: 123044.[CrossRef][Web of Science][Medline]
- Londos E, Warkentin S, Minthon L. "AQT" a measure of cognitive speed is a useful tool in the diagnosis of dementia with Lewy bodies. 9th Internat Psychogeriat Assoc Conf. Stockholm, Sweden, 2005.
- Baddeley A, Bressi S, Della Sala S et al. The decline of working memory in Alzheimers disease: a longitudinal study. Brain 1991; 114: 252142.
[Abstract/Free Full Text] - Fry AF, Hale S. Processing speed, working memory, and fluid intelligence: evidence for a developmental cascade. Psychol Sci 1996; 7: 23741.
- Fockert JW, Rees G, Rith CD et al. The role of working memory in selective visual attention. Science 2001; 291: 18036.
[Abstract/Free Full Text] - The World Factbook 2003. Superintendent of Documents, Pittsburgh, PA.
![]()
CiteULike
Connotea
Del.icio.us What's this?
This article has been cited by other articles:
![]() |
M. Andersson, E. H. Wiig, L. Minthon, and E. Londos A Quick Test for Cognitive Speed: A Measure of Cognitive Speed in Dementia With Lewy Bodies American Journal of Alzheimer's Disease and Other Dementias, September 1, 2007; 22(4): 313 - 318. [Abstract] [PDF] |
||||
![]() |
N. P. Nielsen, R. Ringstrom, E. H. Wiig, and L. Minthon Associations between AQT Processing Speed and Neuropsychological Tests in Neuropsychiatric Patients American Journal of Alzheimer's Disease and Other Dementias, July 1, 2007; 22(3): 202 - 210. [Abstract] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
