Dement Geriatr Cogn Disord 2015;40:22–32 DOI: 10.1159/000377673 Accepted: February 3, 2015 Published online: April 14, 2015
© 2015 S. Karger AG, Basel 1420–8008/15/0402–0022$39.50/0 www.karger.com/dem
Original Research Article
Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians Kylie Radford a, b Holly A. Mack a Brian Draper a–c Simon Chalkley a–c Kim Delbaere a, b Gail Daylight c Robert G. Cumming a, d Hayley Bennett a Gerald A. Broe a, b a
Neuroscience Research Australia, b Faculty of Medicine, University of New South Wales, South Eastern Sydney Local Health District, NSW Health, and d School of Public Health, University of Sydney, Sydney, N.S.W., Australia c
Key Words Dementia · Indigenous population · Cognitive impairment · Community sample · Neuropsychological assessment · Mini-Mental State Examination · Kimberley Indigenous Cognitive Assessment · Rowland Universal Dementia Assessment Scale Abstract Background: Validated cognitive screening tools for use in urban and regional Aboriginal populations in Australia are lacking. Methods: In a cross-sectional community-based study, 235 participants were assessed on the Mini-Mental State Examination (MMSE), the Rowland Universal Dementia Assessment Scale (RUDAS) and an urban modification of the Kimberley Indigenous Cognitive Assessment (mKICA). Performance on these cognitive screening tools was compared to dementia diagnosis by clinical consensus. Results: All tests were culturally acceptable with good psychometric properties. Receiver operating characteristic curve analyses revealed that the MMSE and mKICA were the most accurate. Conclusion: The MMSE is an effective cognitive screening tool in urban Aboriginal populations. The mKICA is a good alternative when illiteracy, language or cultural considerations deem it appropriate. The © 2015 S. Karger AG, Basel RUDAS also has adequate validity in this population.
Introduction
Dr. Kylie Radford Neuroscience Research Australia PO Box 1165 Randwick, NSW 2031 (Australia) E-Mail k.radford @ neura.edu.au
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
There is a dearth of validated cognitive screening tools for use in Indigenous Australian populations. In older Aboriginal and Torres Strait Islander people, dementia is of growing concern. Recent research indicates that the prevalence of dementia in Aboriginal commu-
23
Dement Geriatr Cogn Disord 2015;40:22–32 DOI: 10.1159/000377673
© 2015 S. Karger AG, Basel www.karger.com/dem
nities is 3–5 times higher than experienced in the general Australian population, and higher than in many other populations worldwide [1, 2]. Hence, culturally appropriate, objective and reliable methods of assessing cognitive function and identifying cognitive impairment are urgently needed to improve timely and accurate dementia diagnosis in this population. The issue of dementia screening has been partially addressed with the development of the Kimberley Indigenous Cognitive Assessment (KICA) tool [3]. The KICA incorporates a formal cognitive screening test (KICA-Cog) along with medical history and collateral (family) report sections. It was developed to meet the needs of remote Aboriginal communities, where many different Indigenous languages are spoken (often instead of English), and levels of formal education and literacy tend to be low – factors which influence performance on standard cognitive tests, along with cultural differences. The KICA-Cog has been validated with Aboriginal Australians in the remote Kimberley region of Western Australia and in the Northern Territory of Australia [4]. A shortened version of the KICA-Cog has also been validated in Far North Queensland with a predominantly English-speaking Aboriginal (60%) and Torres Strait Islander (40%) sample [5]. The KICA-Cog has been shown to have strong diagnostic properties for dementia screening [4]. Alternatives include the Mini-Mental State Examination (MMSE) [6] and the Rowland Universal Dementia Assessment Scale (RUDAS) [7]. The MMSE is arguably the most common and extensively validated cognitive screening tool used in clinical practice and research around the world [8, 9]. Most research on the MMSE in indigenous populations has been conducted in North America. For Native Americans, poorer performance on the MMSE was found to be associated with socioeconomic disadvantage and low education [10] – a finding that has been consistently reported across many populations [8, 11, 12]. Whyte et al. [13] compared Native American and Caucasian patients with Alzheimer’s disease (matched on age, sex and education) and found no significant difference in MMSE scores. The authors concluded that cultural differences, therefore, did not have an impact on test performance. The RUDAS [7] is a cognitive screening tool that was specifically designed to minimize the impact of cultural differences on test performance. It has been validated in culturally diverse groups, both in Australia [14–16], where it was initially developed, and internationally [17– 19]. RUDAS performance is not associated with cultural or language background in multicultural settings [14, 15, 19]. The RUDAS is also said to be less influenced by sociodemographic factors such as education [14, 16, 19, 20], but this might not be the case in all populations [17, 21]. In direct comparison with the MMSE, the RUDAS has generally been found to have similar accuracy for dementia screening [15–20]. Whilst the KICA-Cog represented an important advance in cognitive assessment for remote Aboriginal Australians, the majority of Aboriginal Australians reside in urban and regional (i.e., non-remote) communities and there are currently no validated cognitive screening tools for this population. Urban/regional populations differ from remote populations in that they are more likely to have been exposed to formal education systems, have higher levels of education and literacy, and speak English as a first or primary language [22]. Therefore, standard cognitive tools such as the MMSE and RUDAS should be appropriate for use in urban Aboriginal populations, along with the culturally specific KICA-Cog (adapted for urban settings). A primary aim of the epidemiological Koori Growing Old Well Study (KGOWS) [22] was to assess the relative performance of these standard (MMSE and RUDAS) and adapted (KICA-Cog) cognitive screening instruments for the diagnosis of dementia and cognitive impairment in the urban/regional Aboriginal population of Australia.
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
Radford et al.: Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians
24
Dement Geriatr Cogn Disord 2015;40:22–32 DOI: 10.1159/000377673
© 2015 S. Karger AG, Basel www.karger.com/dem
Radford et al.: Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians
Materials and Methods A cross-sectional community-based study was conducted in partnership with Aboriginal communitycontrolled health organizations, and under the guidance of local community Elders and a study-specific Aboriginal reference group. The study was approved by the Aboriginal Health and Medical Research Council Ethics Committee (AHMRC; 615/07), University of New South Wales Human Research Ethics Committee (HREC 08003) and New South Wales (NSW) Population & Health Services Research Ethics Committee (AU RED Ref: HREC/09/CIPHS/65; Cancer Institute NSW Ref: 2009/10/187). All participants gave written informed consent or, in the case of reduced capacity to consent, gave their assent and proxy informed consent was obtained. Participants A representative sample of 336 older Aboriginal men and women was recruited from 5 urban and regional catchment areas in NSW, Australia, as part of the epidemiological KGOWS [22]. Inclusion criteria were Aboriginal or Torres Strait Islander identity, aged 60 years and older, and having lived in 1 of the 5 catchment areas for 6 months or longer. Exclusion criteria were current incarceration and stroke within the last 3 months. Study recruitment was mainly through local Aboriginal research assistants, with the support of community organizations and local guidance groups. Further details of this sample have been previously reported [1].
Procedure Participants were screened using all 3 cognitive measures, presented in the order above and intermixed with other survey questions or psychological scales, as part of a comprehensive structured interview of life span health and well-being [22]. Where items overlapped across screening tools (i.e., orientation items and animal fluency task), these were only administered once (on the first occasion). Screening was conducted by trained research assistants, with at least graduate qualifications in psychology or a related field, working alongside Aboriginal research assistants from the local community. Across the course of the multisite study, 14 research assistants were involved in collecting screening data. Data were collected at the participant’s place of residence or at local community health or research centers. Based on previous piloting [25], all participants who scored below conservative study cutoffs on any one of these measures (i.e., MMSE ≤26, mKICA ≤35 and RUDAS ≤25) proceeded to a detailed medical assessment, including medical and social history, neurological examination, further cognitive testing and contact person interview with a close relative or friend (for full details, see Radford et al. [22]). The remaining participants were classified as cognitively intact, given the reasonably stringent screening criteria (i.e., a high score on all 3 tests). A 20% random sample of the cognitively intact population was selected to proceed to medical assessment. Medical assessments were conducted by physicians with expertise and training in geriatric medicine (7 physicians across 5 sites). The results of this medical assessment were reviewed by a panel of ≥3 clinicians (including geriatricians and clinical neuropsychologists), who determined by consensus a
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
Measures A structured interview was used to collect demographic information including age, sex, language background and education level. In addition, 3 cognitive screening tests were administered. (1) The MMSE [6] is scored out of 30, with 11 items assessing different cognitive domains of orientation, attention, memory, language and visuospatial skills. Detailed scoring guidelines developed by Mioshi et al. [23] were used in the current study. Notably, both attention items were administered (i.e., serial 7s and ‘world’ backwards), with the higher score included in the total. The accepted MMSE dementia screening cutoff is a score
© 2015 S. Karger AG, Basel 1420–8008/15/0402–0022$39.50/0 www.karger.com/dem
Original Research Article
Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians Kylie Radford a, b Holly A. Mack a Brian Draper a–c Simon Chalkley a–c Kim Delbaere a, b Gail Daylight c Robert G. Cumming a, d Hayley Bennett a Gerald A. Broe a, b a
Neuroscience Research Australia, b Faculty of Medicine, University of New South Wales, South Eastern Sydney Local Health District, NSW Health, and d School of Public Health, University of Sydney, Sydney, N.S.W., Australia c
Key Words Dementia · Indigenous population · Cognitive impairment · Community sample · Neuropsychological assessment · Mini-Mental State Examination · Kimberley Indigenous Cognitive Assessment · Rowland Universal Dementia Assessment Scale Abstract Background: Validated cognitive screening tools for use in urban and regional Aboriginal populations in Australia are lacking. Methods: In a cross-sectional community-based study, 235 participants were assessed on the Mini-Mental State Examination (MMSE), the Rowland Universal Dementia Assessment Scale (RUDAS) and an urban modification of the Kimberley Indigenous Cognitive Assessment (mKICA). Performance on these cognitive screening tools was compared to dementia diagnosis by clinical consensus. Results: All tests were culturally acceptable with good psychometric properties. Receiver operating characteristic curve analyses revealed that the MMSE and mKICA were the most accurate. Conclusion: The MMSE is an effective cognitive screening tool in urban Aboriginal populations. The mKICA is a good alternative when illiteracy, language or cultural considerations deem it appropriate. The © 2015 S. Karger AG, Basel RUDAS also has adequate validity in this population.
Introduction
Dr. Kylie Radford Neuroscience Research Australia PO Box 1165 Randwick, NSW 2031 (Australia) E-Mail k.radford @ neura.edu.au
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
There is a dearth of validated cognitive screening tools for use in Indigenous Australian populations. In older Aboriginal and Torres Strait Islander people, dementia is of growing concern. Recent research indicates that the prevalence of dementia in Aboriginal commu-
23
Dement Geriatr Cogn Disord 2015;40:22–32 DOI: 10.1159/000377673
© 2015 S. Karger AG, Basel www.karger.com/dem
nities is 3–5 times higher than experienced in the general Australian population, and higher than in many other populations worldwide [1, 2]. Hence, culturally appropriate, objective and reliable methods of assessing cognitive function and identifying cognitive impairment are urgently needed to improve timely and accurate dementia diagnosis in this population. The issue of dementia screening has been partially addressed with the development of the Kimberley Indigenous Cognitive Assessment (KICA) tool [3]. The KICA incorporates a formal cognitive screening test (KICA-Cog) along with medical history and collateral (family) report sections. It was developed to meet the needs of remote Aboriginal communities, where many different Indigenous languages are spoken (often instead of English), and levels of formal education and literacy tend to be low – factors which influence performance on standard cognitive tests, along with cultural differences. The KICA-Cog has been validated with Aboriginal Australians in the remote Kimberley region of Western Australia and in the Northern Territory of Australia [4]. A shortened version of the KICA-Cog has also been validated in Far North Queensland with a predominantly English-speaking Aboriginal (60%) and Torres Strait Islander (40%) sample [5]. The KICA-Cog has been shown to have strong diagnostic properties for dementia screening [4]. Alternatives include the Mini-Mental State Examination (MMSE) [6] and the Rowland Universal Dementia Assessment Scale (RUDAS) [7]. The MMSE is arguably the most common and extensively validated cognitive screening tool used in clinical practice and research around the world [8, 9]. Most research on the MMSE in indigenous populations has been conducted in North America. For Native Americans, poorer performance on the MMSE was found to be associated with socioeconomic disadvantage and low education [10] – a finding that has been consistently reported across many populations [8, 11, 12]. Whyte et al. [13] compared Native American and Caucasian patients with Alzheimer’s disease (matched on age, sex and education) and found no significant difference in MMSE scores. The authors concluded that cultural differences, therefore, did not have an impact on test performance. The RUDAS [7] is a cognitive screening tool that was specifically designed to minimize the impact of cultural differences on test performance. It has been validated in culturally diverse groups, both in Australia [14–16], where it was initially developed, and internationally [17– 19]. RUDAS performance is not associated with cultural or language background in multicultural settings [14, 15, 19]. The RUDAS is also said to be less influenced by sociodemographic factors such as education [14, 16, 19, 20], but this might not be the case in all populations [17, 21]. In direct comparison with the MMSE, the RUDAS has generally been found to have similar accuracy for dementia screening [15–20]. Whilst the KICA-Cog represented an important advance in cognitive assessment for remote Aboriginal Australians, the majority of Aboriginal Australians reside in urban and regional (i.e., non-remote) communities and there are currently no validated cognitive screening tools for this population. Urban/regional populations differ from remote populations in that they are more likely to have been exposed to formal education systems, have higher levels of education and literacy, and speak English as a first or primary language [22]. Therefore, standard cognitive tools such as the MMSE and RUDAS should be appropriate for use in urban Aboriginal populations, along with the culturally specific KICA-Cog (adapted for urban settings). A primary aim of the epidemiological Koori Growing Old Well Study (KGOWS) [22] was to assess the relative performance of these standard (MMSE and RUDAS) and adapted (KICA-Cog) cognitive screening instruments for the diagnosis of dementia and cognitive impairment in the urban/regional Aboriginal population of Australia.
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
Radford et al.: Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians
24
Dement Geriatr Cogn Disord 2015;40:22–32 DOI: 10.1159/000377673
© 2015 S. Karger AG, Basel www.karger.com/dem
Radford et al.: Comparison of Three Cognitive Screening Tools in Older Urban and Regional Aboriginal Australians
Materials and Methods A cross-sectional community-based study was conducted in partnership with Aboriginal communitycontrolled health organizations, and under the guidance of local community Elders and a study-specific Aboriginal reference group. The study was approved by the Aboriginal Health and Medical Research Council Ethics Committee (AHMRC; 615/07), University of New South Wales Human Research Ethics Committee (HREC 08003) and New South Wales (NSW) Population & Health Services Research Ethics Committee (AU RED Ref: HREC/09/CIPHS/65; Cancer Institute NSW Ref: 2009/10/187). All participants gave written informed consent or, in the case of reduced capacity to consent, gave their assent and proxy informed consent was obtained. Participants A representative sample of 336 older Aboriginal men and women was recruited from 5 urban and regional catchment areas in NSW, Australia, as part of the epidemiological KGOWS [22]. Inclusion criteria were Aboriginal or Torres Strait Islander identity, aged 60 years and older, and having lived in 1 of the 5 catchment areas for 6 months or longer. Exclusion criteria were current incarceration and stroke within the last 3 months. Study recruitment was mainly through local Aboriginal research assistants, with the support of community organizations and local guidance groups. Further details of this sample have been previously reported [1].
Procedure Participants were screened using all 3 cognitive measures, presented in the order above and intermixed with other survey questions or psychological scales, as part of a comprehensive structured interview of life span health and well-being [22]. Where items overlapped across screening tools (i.e., orientation items and animal fluency task), these were only administered once (on the first occasion). Screening was conducted by trained research assistants, with at least graduate qualifications in psychology or a related field, working alongside Aboriginal research assistants from the local community. Across the course of the multisite study, 14 research assistants were involved in collecting screening data. Data were collected at the participant’s place of residence or at local community health or research centers. Based on previous piloting [25], all participants who scored below conservative study cutoffs on any one of these measures (i.e., MMSE ≤26, mKICA ≤35 and RUDAS ≤25) proceeded to a detailed medical assessment, including medical and social history, neurological examination, further cognitive testing and contact person interview with a close relative or friend (for full details, see Radford et al. [22]). The remaining participants were classified as cognitively intact, given the reasonably stringent screening criteria (i.e., a high score on all 3 tests). A 20% random sample of the cognitively intact population was selected to proceed to medical assessment. Medical assessments were conducted by physicians with expertise and training in geriatric medicine (7 physicians across 5 sites). The results of this medical assessment were reviewed by a panel of ≥3 clinicians (including geriatricians and clinical neuropsychologists), who determined by consensus a
Downloaded by: NYU Medical Center Library 128.122.253.228 - 4/21/2015 12:35:35 PM
Measures A structured interview was used to collect demographic information including age, sex, language background and education level. In addition, 3 cognitive screening tests were administered. (1) The MMSE [6] is scored out of 30, with 11 items assessing different cognitive domains of orientation, attention, memory, language and visuospatial skills. Detailed scoring guidelines developed by Mioshi et al. [23] were used in the current study. Notably, both attention items were administered (i.e., serial 7s and ‘world’ backwards), with the higher score included in the total. The accepted MMSE dementia screening cutoff is a score
