Cognitive Impairment in the Amish: A Four County Survey
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
International Journal of Epidemiology Vol. 26, No. 2
© International Epidemiological Association 1997 Printed in Great Britain
Cognitive Impairment in the Amish:
A Four County Survey
C C JOHNSON,* B A RYBICKI,* G BROWN,** E D’HONDT,* B HERPOLSHEIMER,* D ROTH* AND
C E JACKSON†
Johnson C C (Division of Biostatistics and Research Epidemiology, Henry Ford Health System, One Ford Place, 3E,
Detroit, MI 48202–3450, USA), Rybicki B A, Brown G, D’Hondt E, Herpolsheimer B, Roth D and Jackson C E. Cognitive
impairment in the Amish: A four county survey. International Journal of Epidemiology 1997; 26: 387–394.
Background. The prevalence of probable dementia was determined in a rural, homogeneous community of Amish indi-
viduals in the Midwestern USA. The Amish are a genetically isolated group with a low level of formal education (ø8 years)
and few exposures to modern life, who live in intergenerational settings and have strong social support networks.
Methods. Using community directories, trained interviewers administered the Mini Mental State Examination (MMSE)
and a medical history survey to all Amish over 64 years old in a four county area. Individuals with scores ,27 (out of a
maximum of 30 points) were given additional neuropsychological tests. Results were reviewed by a neuropsychologist
and subjects were classified with regard to probable cognitive impairment.
Results. The MMSE scores were inversely related with age and directly with education. The Amish have higher MMSE
scores than reported for the general US population. The overall prevalence of probable cognitive impairment for those
over 64 years was 6.4%. The prevalence increased with age and lower education and was lowest among married
individuals.
Conclusions. The MMSE scores among the Amish were higher than the general population despite their low level of
formal education. The lower level of cognitive impairment among the Amish could reflect a lack of inherited susceptibility
to dementing diseases, or environmental factors characteristic of their traditional lifestyle. Even among this population
with ø8 years of formal education, education may protect against cognitive impairment.
Keywords: cognitive impairment, prevalence, dementia, community survey, MMSE, Amish
A number of studies have measured the prevalence of Another important difference between the Amish and
cognitive impairment in defined, characterized popula- most other populations previously studied is that the
tions.1–10 Most surveys have been in urban areas or Amish tend to care for infirm family members at home
retirement communities2,5,7–9 and often exclude resid- rather than institutionalizing them. Therefore, the popu-
ents of long term care institutions. The purpose of this lation enumerated for our study should have provided a
study was to evaluate the prevalence of dementia in more valid estimate of the prevalence of dementia com-
rural communities of Amish centered in Indiana and pared with previous efforts limited to older individuals
Michigan, where together with Ohio and Pennsylvania, capable of autonomous living.
about 80% of the Amish reside in the US. The Amish, We compared the age-specific distribution of Mini
of Swiss descent and genetically isolated, live a Mental State Examination (MMSE) scores in the Amish
traditional lifestyle with few exposures to modern life, population with the general US population, and report
but have strong social support networks and high stand- the overall and age- and sex-specific prevalence of de-
ards of living and medical care.11 A number of studies mentia. We were also able to test whether formal edu-
have suggested that a low level of formal education is cation is associated with dementia after adjustment for
positively associated with dementia, and specifically age and sex in a setting in which education does not go
Alzheimer’s disease.12,13 Amish individuals generally beyond the eighth grade.
attend their own schools and, for religious and cultural
reasons, leave formal schooling after the 8th grade.
METHODS
Amish communities in the Midwest periodically pub-
* Division of Biostatistics and Research Epidemiology, † Department lish directories that list all Amish residents in a county,
of Medical Genetics, Henry Ford Health System, One Ford Place, 3E,
Detroit MI 48202–3450, USA.
by household, location of residence, names of family
** New York Hospital, Cornell Medical Center, White Plains, New members, gender, and dates of birth (and death for de-
York, USA. ceased spouses). Using the most recent directories
387388 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY which included a 1988 directory for two contiguous gender and vocabulary proficiency.18 Scores on this counties (Elkhart and LaGrange) in north central Indi- instrument are affected by age,19,20 however age effects ana, a 1987 directory for a Michigan county (St Joseph) were controlled to some extent in this study because bordering these counties to the North, and a 1992 dir- only those individuals .65 years were evaluated and ectory for a non-contiguous north eastern Indiana age-specific scores were calculated. While the MMSE county (Adams), a sampling frame was enumerated of has been shown to vary by ethnicity and years of educa- all Amish residents ù65 years as of 1 January 1991. In tion,20,21 the Amish represent a homogeneous Caucasian contrast to many other populations, extended families community with a culturally mandated low level of tend to live on the same property, and it is uncommon formal education, but high literacy. for older Amish citizens to move to senior residential The MMSE score was computed by summing the communities or apartments (estimated at ,2%), but correct responses to obtain a total score that falls in a these people are included in the directories. It is also range 0–30. Non-parametric tests (Spearman’s rank not customary for the Amish to place invalids in ex- correlation coefficient and Wilcoxon 2-sample test) tended care facilities such as nursing homes; in the were used to evaluate the relationship between demo- course of this survey we did not identify any Amish graphic variables and MMSE scores in the Amish popu- person residing in a long term care institution. There- lation. The distribution of MMSE scores, lower quartile fore virtually every Amish person ù65 years in the or 25% percentile, median or 50% percentile, and upper study counties was included in the sample, except for quartile or 75% percentile, were calculated, assigning any who may have actively declined to be included in a averaged ranked scores for tied data values. The dis- directory, which would number less than ten families tribution of scores in the Amish was compared to a according to community leaders. national household survey that reported age and edu- One of the investigators (CEJ) had an established cation level specific population-based normative values medical research relationship with these communities. for the MMSE.20 The study was first introduced to community leaders to Individuals whose score on the MMSE was ù27 were obtain their support. Trained interviewers (ED, BH, DR), classified as cognitively normal. Those with lower scores going door-to-door, visited the enumerated households received a more extensive neuropsychological evalu- in the four counties during three summers, 1991–1993. ation comprised of the Dementia Rating Scale (DRS),22 Those who had died since the publication of the dir- the Boston Naming Test (BNT), 23 and the Level 2 Read- ectories or who had moved out of the four counties ing subtest from the Wide Range Achievement Test- were excluded from the study population. People who Revised (WRAT-R2).24 The DRS is lengthier than the could not validly complete the MMSE due to physical MMSE, assesses a broader range of behaviours,22 is disabilities such as blindness or deafness were likewise highly reliable over time,25 and is a more sensitive excluded from both the denominator and the numerator measure of decline than the MMSE.26,27 The BNT as- for analyses. sesses confrontational naming, a behaviour commonly Participating subjects were given the MMSE14 to impaired with degenerative dementia. 28 We used screen for evidence of cognitive impairment. This Version 1 of Mack et al.29 15-item short-form of the instrument has been used in this manner in numerous BNT. The WRAT-Reading subtest is a measure of single epidemiological surveys and clinical studies.15,16 The word reading, a skill that is more resistant to dementia training of the research assistants in the administration than many other cognitive skills.30 Low WRAT-R2 of the MMSE was overseen by a neuropsychologist reading scores among individuals with mild cognitive (GB). The MMSE was administered at the beginning of impairment would raise the possibility of a lifelong the interview; demographic and medical information cognitive dysfunction rather than adult-onset decline. was collected later in the interview via standardized Individuals whose MMSE score was ,27 were forms. As in the original version of the MMSE, par- classified as cognitively impaired if their DRS score ticipants were given the option of spelling ‘world’ fell below an age-adjusted cutting score, provided their backward if they had difficulties with the serial sevens test results were valid. Participants with a clinical diag- (subtraction by sevens from 100) and the question that nosis of dementia who could not attempt the MMSE had the highest score was used. were also classified as cognitively impaired and given The MMSE was chosen as the primary screening an MMSE score of 0. Total, age, gender and education instrument because it has acceptable levels of sensitivity level-specific prevalence ratios of cognitive impairment and specificity in detecting dementia and delirium; 17 has were calculated. To explore the interrelationships be- a high degree of inter-rater and test-retest reliability;14 tween these variables, adjusted odds ratios were cal- and is not influenced by such confounding variables as culated using logistic regression.
COGNITIVE IMPAIRMENT IN THE AMISH 389
TABLE 1 Disposition of population under study physician-diagnosed dementia were unable to be tested
due to lack of comprehension, leaving 516 people who
Sampling frame 771 undertook the MMSE (Table 1). The latter were given
Discovered moved 15 scores of 0 and thus placed in the lowest quartile of
Discovered deceased 106 MMSE scores so as not to upwardly bias reported
Physically unable to complete valid MMSE 18
Available sample 632
results.
Unable to contact 28 The MMSE score decreased with age (Spearman
Refused interview 78 correlation coefficient of –0.39; P , 0.001) (Table 3).
Participants 526 Of the 509 with MMSE scores and education level,
Completed MMSE 516
114 (22%) had ,8 years education, and the mean age of
Incomplete MMSE 4
History of dementia 6 these individuals was significantly higher than those
who completed 8 years of school (76.8 versus 73.8;
P , 0.001). MMSE score was directly related with edu-
cation (mean score for ,8 years of 27.3 versus mean
RESULTS score for .8 years of 28.2; P , 0.0001). Compared to
A total of 771 individuals were enumerated in the four a national sample with 5–8 years of formal education,
counties (Table 1). Fifteen were determined to have the Amish have a higher median MMSE score, by at
moved, and 106 were deceased. Eighteen people were least two points, for every age group and for both levels
unable to attempt or complete the MMSE due to phys- of education. Except for the 85+ age group, Amish
ical disability or illness, leaving 632 eligible people. individuals at the 25 percentile and the 75 percentile
Of this number, 28 (4.4%) could not be contacted, scored higher than the general population sample. This
78 (12.3%) refused to participate, and 526 (83.2%) con- pattern was sustained in the two educational strata.
sented to be included. An MMSE score of ù27 was attained by 481 indi-
The 526 participants included 236 men and 290 viduals, who were classified as cognitively normal.
women (Table 2). The average age was 74.7 years (SD Twelve people had scores from 24 through 26 on the
6.3). Of the participants, 507 (96.4%) lived in Indiana MMSE. All but one subject within this range of scores
and 19 (3.6%) in Michigan. The 106 who refused or were completed the DRS. This 76 year old participant, who
not contacted did not differ statistically from parti- had an MMSE score of 24, could not complete the
cipants in terms of gender or age (Table 2), but Mich- assessment because she was too busy. She recalled two
igan residents had a lower proportion participating, of three words after a delay on the MMSE and was
although the number of Amish in Michigan was small, oriented to time, place, and person. Because she did not
so the absolute difference was only 4.9%. appear to have any impairment of memory or of daily
Of the 526 consenting to participate, four refused activities, she was classified as cognitively normal. To
to complete the MMSE, and six with a history of classify the remaining 11 individuals with intermediate
TABLE 2 Comparison of respondents (n = 526) versus non-respondents (n = 106)
Respondents (%) Non-respondents (%) P-valuea
Gender
Males 236 (44.9) 42 (39.6) 0.321
Females 290 (55.1) 64 (60.4)
Age
65–69 146 (27.8) 27 (25.5) 0.222
70–74 156 (29.7) 23 (21.2)
75–79 115 (21.9) 24 (22.6)
80–84 71 (13.5) 21 (19.8)
85+ 38 (7.2) 11 (10.4)
Location
IN residence 507 (96.4) 97 (91.5) 0.036
MI residence 19 (3.6) 9 (8.5)
a
χ2 test.390 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
TABLE 3 Mini Mental State Examination (MMSE) score by age group for Amisha (n = 522) and normative population sampleb
Age N Lower Quartile Median Upper Quartile
Amish
Total 65–69 146 28 29 30
70–74 156 27 28 29
75–79 115 27 28 29
80–84 69 27 27 28
85+ 36 20.5 27 28
Education levelc
,8 years education 65–69 18 27 29 30
70–74 31 27 28 29
75–79 32 27 28 29
80–84 21 25 27 28
85+ 12 19 27 27
8 years education 65–69 127 28 29 30
70–74 121 28 29 29
75–79 81 28 28 29
80–84 47 27 27 28
85+ 19 25 27 28
Normative sample
5–8 years education 65–69 633 24 27 29
70–74 533 24 26 28
75–79 437 22 26 28
80–84 241 22 25 27
85+ 134 21 24 27
a
People who attempted but were unable to undertake the MMSE and had a physician diagnosis of dementia received a 0 score and are included.
b
Data from household community surveys, 5 US sites, and weighted based on the 1980 US census by age, sex and race.20
c
Missing education level on 13 individuals; one 65–69 years, four 70–74 years, two 75–79 years, one 80–84 years and five 85+ years of age.
MMSE scores, we used the age-adjusted cutting scores lower education among the oldest age groups. Among
on the DRS of 125 for subjects ,80 years of age and the 70–79 year olds the prevalence was higher in males,
123 for subjects .79 years.31 One 80 year old subject but was comparable by gender in the oldest old. The
with a DRS score of 125 was excluded from the cal- highest prevalence of probable dementia was among
culation of prevalence ratios as he had an impaired those in Adams County in Northeast Indiana, particu-
single word reading standard score of 76, raising the larly in those 80+ years. Almost everyone was married
possibility of mild, lifelong cognitive dysfunction. Three or widowed; the overall prevalence was lowest for
additional subjects were classified as cognitively nor- married individuals.
mal, and seven as impaired. In a logistic regression model, age (65–74, 75–79,
There were 29 people with scores ,24 on the MMSE. 80–84, 85+ years) was significantly associated with
One person was judged to be cognitively normal based risk of probable dementia (Table 5). Education level
on the complete battery and complicating medical (ù8 years versus ,8 years) remained protective but
conditions. Two subjects, one with visual limitations, decreased in magnitude and was no longer statistically
did not complete the battery due to lack of co-operation, significant after adjusting for age and gender.
could not be classified, and were excluded from the
analysis of prevalence. Six, as mentioned above, had a
previous history of physician-diagnosed dementia and DISCUSSION
were assigned a score of 0; these and the remaining Due to the size of the Amish population, the numbers in
20 people were judged to have probable dementia. many of the stratified categories were small and our
A total of 33 people were categorized as having adjusted risk estimates for probable dementia were
probable dementia out of 519 people with a complete accompanied by wide confidence intervals. However,
and valid assessment, for an overall prevalence of 6.4% increasing age and fewer years of schooling appeared to
(Table 4). The prevalence increased with age, and with be associated with poor performance on the MMSE.COGNITIVE IMPAIRMENT IN THE AMISH 391
TABLE 4 Prevalence ratios (%) of probable dementia by age, gender, education level, residence and marital status (n = 519)
Totala Age
65–69 70–74 75–79 80–84 85+
Total 6.4 (33/519) 0.0 (0/146) 1.9 (3/156) 2.6 (3/114) 16.4 (11/67) 44.4 (16/36)
Gender
Males 6.5 (15/232) 0.0 (0/70) 2.9 (2/70) 4.2 (2/48) 16.7 (5/30) 42.9 (6/14)
Females 6.3 (18/287) 0.0 (0/76) 1.2 (1/86) 1.5 (1/66) 16.2 (6/37) 45.5 (10/22)
Educationb
,8 years 9.7 (11/113) 0.0 (0/18) 0.0 (0/31) 0.0 (0/31) 28.6 (6/21) 41.7 (5/12)
8 years 3.1 (12/393) 0.0 (0/127) 0.8 (1/121) 1.2 (1/81) 8.9 (4/45) 31.6 (6/19)
Area of Residence
N Central Indiana 6.4 (27/422) 0.0 (0/127) 2.5 (3/121) 3.4 (3/87) 14.0 (8/57) 43.3 (13/30)
N East Indiana 7.7 (6/78) 0.0 (0/17) 0.0 (0/26) 0.0 (0/23) 42.9 (3/7) 60.0 (3/5)
Michigan 0.0 (0/19) 0.0 (0/2) 0.0 (0/9) 0.0 (0/4) 0.0 (0/3) 0.0 (0/1)
Marital Status
Never Married 22.2 (4/18) 0.0 (0/6) 16.7 (1/6) 25.0 (1/4) 100.0 (1/1) 100.0 (1/1)
Married 4.0 (15/373) 0.0 (0/125) 1.6 (2/124) 2.7 (2/73) 18.4 (7/38) 30.8 (4/13)
Widowed 10.9 (14/128) 0.0 (0/15) 0.0 (0/26) 0.0 (0/37) 10.7 (3/28) 50.0 (11/22)
a
Excluded from analysis are two subjects with incomplete neuropsychological testing and one subject with possible chronic cognitive impairment.
b
Missing education level on 13 individuals; one 65–69 years, four 70–74 years, two 75–79 years, one 80–84 years and five 85+ years of age.
TABLE 5 Crude and adjusted odds ratios for risk of probable dementia, by age, gender, and education level (n = 519)
Crude Odds Ratio 95% Confidence Adjusted Odds Ratioa 95% Confidence
Interval Interval
Age
65–74 1.0 – 1.0 –
75–79 2.7 (0.6–12.8) 2.4 (0.2–39.5)
80–84 19.6 (7.4–52.0) 47.6 (5.9–381.9)
85+ 79.7 (35.7–178.0) 145.3 (17.7–999.0)
Gender
Male 1.0 – 1.0 –
Female 0.97 (0.5–2.0) 0.70 (0.3–1.8)
Educationb
,8 years 1.0 – 1.0 –
.8 years 0.29 (0.1–0.7) 0.44 (0.2–1.2)
a
Odds ratios adjusted for all other factors included in Table.
b
Missing education level on 13 individuals; one 65–69 years, four 70–74 years, two 75–79 years, one 80–84 years and five 85+ years of age.
Although the MMSE scores decreased with age in this if Amish individuals who would have been institu-
Amish population, the scores were higher than those tionalized in other settings were to have been excluded
found in the general population. The prevalence of cog- from our study. Although we were unable to contact
nitive impairment was less in younger age groups but 28 people (4.3%), some of whom could have theoret-
comparable to other study populations for older age ically been located in nursing homes, none of the study
groups.3,4,6–10 We believe that the prevalence would participants was residing in a long-term care facility.
have also been comparably less in the older age groups However, in other populations individuals with dementia392 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY by history, and many with low MMSE scores, would dementia, particularly Alzheimer’s disease.12,13,38–43 have been placed in extended care facilities and not The association is difficult to evaluate as test scores included in prevalence figures. independently have a direct positive relationship with Although we did not measure inter-rater reliability in education, 40 although a recent study in a literate this study, we utilized a standardized version of the Swedish population suggested that the positive pre- MMSE developed for a multicentre trial. This version dictive value of the MMSE as a screening test for of the MMSE more explicitly describes the test’s ad- dementia differs little by level of education.44 The ministration and scoring than earlier forms. Intraclass Amish are an interesting group to study in this regard as correlations for standardized versions of the MMSE they maintain their own schools because of cultural range around 0.90,32 indicating excellent inter-rater reasons, which do not go beyond the eighth grade, and agreement. All examiners were trained under the super- therefore the entire population can be considered as vision of the same neuropsychologist to develop a high having low levels of formal education relative to the uniformity of test administration. general US population. However, unlike other people Although the MMSE has been used as a marker of with this level of education in the general population, dementia in large sample epidemiological studies of the Amish population should represent a full continuum dementia,33,34 it has been criticized for being insensitive of intellectual ability. It is intriguing that among this to some forms of cognitive dysfunction, especially when population an association of lower education with cog- a cutting score of 24 is used.35,36 We compensated for nitive impairment seems to be present, although not the potential insensitivity of the MMSE by using a statistically significant, after controlling for age and cutting score of 27 to more leniently identify demented gender. A limitation of this analysis is that level of edu- individuals. With a cutting score of 27, the MMSE has cation was not included in the Amish directories and a sensitivity near 100% when identifying cognitive had to be collected from the subjects, and was a missing impairment in neurological samples.35 Even though we value for a considerably higher proportion of those who did not use the more comprehensive neuropsycholo- were cognitively impaired, 10 of 33 versus 3 of 486 gical battery to study individuals with MMSE scores classified as cognitively normal. above 27, the high sensitivity of a cutting score of 27 on A consideration is that there is virtually no in- the MMSE to cognitive dysfunction suggests that few, migration of non-Amish, but some young adults leave if any, individuals with cognitive impairment went un- the community. It is possible that there is an out- detected in our study. migration of individuals with higher intellect (perhaps The apparent lower level of cognitive impairment reflected by the desire for more educational opportun- among the Amish may have several explanations. It may ities) and that the surveyed population may be skewed be that the Amish have a lower prevalence of dement- somewhat towards a lower distribution of intellectual ing conditions, such as Alzheimer’s disease. Factors ability, but still have higher median MMSE scores than associated with Amish life may independently, or inter- found in the general population. Another potential actively with inherited characteristics, play a role in this limitation is that those who refused to participate or apparent protection from dementia. Since the Amish were not contacted may have had lower levels of represent a relatively genetically homogeneous group, cognitive function, as has been shown in a recent this may reflect a lower prevalence of inherited suscept- study.45 ibility to one or more of the dementing diseases. The Many other factors associated with higher education very low prevalence among those aged 65–79 supports in the general population are characteristic of the this hypothesis as early onset dementia is considered Amish despite their low number of years in school, to have a stronger genetic component.37 Equally, the including economic well-being, social stability, excep- Amish have a unique as well as homogeneous lifestyle tionally high levels of literacy, good nutrition and ac- and physical and social environment, characterized, for cess to medical care. While the Amish deem a high example, by high levels of physical activity, a limited school education and beyond as too ‘worldly’, they number of occupations, rural settings, and extended spend considerable time in literate pursuits as their cul- families and community support. Of interest is that ture emphasizes lifelong reading and study of religious Jorm, in his 1987 summary of dementia prevalence scriptures, and all Amish are bilingual. Although the studies, noted that prevalence ratios were notably lower analysis of education was somewhat ambiguous, our in rural settings, although only three of 47 studies were study adds some support to Katzman’s theory of ‘neur- of rural populations.1 onal reserve’, or greater brain capacity, as a protection Controversy exists concerning the possible inverse against the deterioration associated with the dementing association of increasing level of formal education with disorders.13
COGNITIVE IMPAIRMENT IN THE AMISH 393
This remarkably homogeneous group of people with 11
Cross H E, McKusick V A. Amish demography. Soc Biol 1978;
an active, rural lifestyle, low formal education, and strong 17: 83–101.
12
social networks appears to have a lower than expected Mortimer J A, Graves A B. Education and other socioeconomic
determinants of dementia and Alzheimer’s disease.
level of cognitive impairment among those .65 years Neurology 1993; 43 (Suppl. 4): S39–44.
of age. Further genetic studies, including neurological 13
Katzman R. Education and the prevalence of dementia and
examination of the cases and testing for the prevalence Alzheimer’s disease. Neurology 1993; 43: 13–20.
of putative markers such as apolipoprotein E446 asso- 14
Folstein M, Folstein S, McHugh P R. Mini-mental state: A
ciated with Alzheimer’s disease, is ongoing.47 practical method for grading the cognitive state of patients
for the clinician. J Psych Res 1975; 12: 189–98.
15
Gagnon N, Letenheur L, Dartigues J-F et al. Validity of the Mini
Mental State examination as a screening instrument for
ACKNOWLEDGEMENTS cognitive impairment and dementia in French elderly com-
The authors acknowledge the work of others who made munity residents. Neuroepidemiology 1990; 9: 143–50.
16
this study possible: Carol Vandenberg, data manager; Tombaugh T N, McIntyre N J. The Mini-Mental State Exam-
Cecelia Day, scientific programmer; and Susan Mc- ination: a comprehensive review. J Am Geriatr Soc 1992;
Guinness, secretary; members of the Amish commun- 40: 922–35.
17
Anthony J C, LeResche L R, Vonkorff M R et al. Limits of the
ities, and Freda E Yoder, Rockville, MD, who provided ‘Mini-Mental State’ as a screening test for dementia and
us with her observation that dementia seemed less delirium among hospital patients. Psychol Med 1982; 12:
common among the Pennsylvania Amish. Supported by 397.
18
Dykstra Foundation, Detroit, MI; the Dorothy L. Stub- Folstein M, Anthony J C, Parhad I, Duffy B, Gruenberg E M.
nitz Foundation, Adrian, MI; and the Henry Ford Health The meaning of cognitive impairment in the elderly. J Am
Geriatr Soc 1985; 33: 228–35.
System Fund for Henry Ford Hospital and summer 19
Bleecker M L, Bolla-Wilson K, Kawas C, Agnew J. Age-
research trainee program. specific norms for the Mini-Mental State Exam. Neurology
1988; 38: 1565–68.
20
Crum R M, Anthony J C, Bassett S S, Folstein M F. Population-
REFERENCES based norms for the Mini-Mental State Examination by age
1
Jorm A J, Korten A E, Henderson A S. The prevalence of and educational level. JAMA 1993; 269: 2386–91.
21
dementia: A quantitative integration of the literature. Acta Bohnstedt M, Fox P J, Kohatsu N D. Correlates of Mini-Mental
Psychiatr Scand 1987; 76: 465–79. Status Examination scores among elderly demented pa-
2
Evans D A, Funkenstein H H, Albert M S et al. Prevalence of tients: The influence of race-ethnicity. J Clin Epidemiol
Alzheimer’s disease in a community population of older 1994; 47: 1381–87.
22
persons higher than previously reported. JAMA 1989; 262: Mattis S. Mental status examination for organic mental
2551–56. syndrome in the elderly patient. In: Bellak L, Karasu T B
3
Heyman A, Fillenbaum G, Prosnitz B et al. Estimated pre- (eds). Geriatric Psychiatry. New York, NY: Grune &
valence of dementia among elderly black and white com- Stratton, 1976.
23
munity residents. Arch Neurol 1991; 48: 594–98. Kaplan E, Goodglass H, Weintraub S. The Boston Naming Test.
4
Hofman A, Rocca W A, Brayne C et al. The prevalence of de- Philadelphia, PA: Lea & Febiger, 1976.
24
mentia in Europe: a collaborative study of 1980–1990 find- Jastak S, Wilkinson G S. Wide Range Achievement Test–Revised
ings. Int J Epidemiol 1991; 20: 736–48. 1984. Wilmington, Delaware: Jastak Associates.
5 25
Pfeffer R I, Afifi A A, Chance J M. Prevalence of Alzheimer’s Coblenz J M, Mattis S, Zingesser L H, Kasoff S S, Wisniewski
disease in a retirement community. Am J Epidemiol 1987; H M, Katzman R. Presenile Dementia. Arch Neurol 1973;
125: 420–36. 29: 299–308.
6 26
Rocca W A, Bonaiuto S, Lippi A et al. Prevalence of clinically Haxby J V, Raffade K, Gillette J, Schapiro M B, Rappoport S I.
diagnosed Alzheimer’s disease and other dementing Individual trajectories of cognitive decline in patients with
disorders: a door-to-door survey in Appignano, Macerato dementia of the Alzheimer type. J Clinical Exp Neuro-
Province, Italy. Neurology 1990; 40: 626–31. psychol 1992; 14: 575–92.
7 27
Skoog I, Nilsson L, Palmertz B et al. A population-based study of Salmon D P, Thal L J, Butters N, Heindel W C. Longitudinal
dementia in 85-year olds. N Engl J Med 1993; 328: 153–58. evaluation of dementia of the Alzheimer type: A com-
8
Saunders P A, Copeland J R M, Dewey M E et al. The pre- parison of 3 standardized mental status examinations.
valence of dementia, depression and neurosis in later life: Neurology 1990; 40: 1225–30.
28
The Liverpool MRC-ALPHA Study. Int J Epidemiol 1993; Bayles K A, Tomoeda C K. Confrontation naming impairment in
22: 838–47. dementia. Brain Lang 1983; 19: 98–114.
9 29
Copeland J R M, Gurland B J, Dewey M E, Kelleher M J, Smith Mack W J, Freed D M, Williams B W, Henderson V W. Boston
M R, Davidson I A. Is there more dementia, depression and Naming Test: Shortened versions for use in Alzheimer’s
neurosis in New York? A comparative study of the elderly Disease. Neuropsychologia 1992; 47: 154–58.
30
in New York and London using the computer diagnosis Nelson H E. National Adult Reading Test (NART): Test Manual
AGECAT. Br J Psychiatry 1987; 151: 466–73. 1982. Windsor: NFER Nelson.
10 31
Livingston G, Sax K, Willison J et al. The Gospel Oak Study Vangel S J Jr, Lichtenberg P A. Mattis dementia rating scale:
Stage II: The diagnosis of dementia in the community. Clinical utility and relationship with demographic vari-
Psychol Med 1990; 20: 881–91. ables. Clin Neuropsychologist 1995; 9: 209–13.394 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
32 41
Molloy D W, Alemayehy E, Roberts R. Reliability of a stand- Bonaiuto S, Rocca W A, Lippi A et al. Education and occu-
ardized Mini-Mental State Examination compared with the pation as risk factors for dementia: a population-based
traditional Mini-Mental State Examination. Am J Psy- case-control study. Neuroepidemiol 1995; 14: 101–09.
42
chiatry 1991; 148: 102–05. Beard I M, Kokmen E, Offord K P, Kurland L T. Lack of
33
Folstein M, Anthony J C, Parhad I, Duffy B, Gruenberg E M. association between Alzheimer’s disease and education,
The meaning of cognitive impairment in the elderly. J Am occupation, marital status, or living arrangement.
Geriatr Soc 1985; 33: 228–35. Neurology 1992; 42: 2063–68.
34 43
Zhang M, Katzman R, Salmon D et al. The prevalence of Evans D A, Beckett L A, Albert M S et al. Level of educa-
dementia and Alzheimer’s disease in Shanghai, China: The tion and change in cognitive function in a community
impact of age, gender, and education. Ann Neurol 1990; population of older persons. Ann Epidemiol 1993; 3:
27: 428–37. 71–77.
35 44
Dick J P R, Guiloff R J, Stewart A et al. Mini-mental state Fratiglioni L, Jorm A F, Grut M et al. Predicting dementia from
examination in neurological patients. J Neurol Neurosurg the Mini-Mental State Examination in an elderly
Psychiatr 1984; 47: 496–99. population: the role of education. J Clin Epidemiol 1993;
36
Babikian V L, Wolfe N, Linn R, Knoefel J E, Albert M L. Cog- 46: 281–87.
45
nitive changes in patients with multiple cerebral infarcts. Launer L J, Wind A W, Deeg D J M. Non-response pattern and
Stroke 1990; 21: 1013–18. bias in a community-based cross-sectional study of
37
Goedert M, Strittmatter W J, Roses A D. Risky apolipoprotein in cognitive functioning among the elderly. Am J Epidemiol
brain. Nature 1994; 372: 45–46. 1994; 139: 803–12.
38 46
Friedland R P. Epidemiology, education, and the ecology of Corder E H, Saunders A M, Strittmatter W J et al. Apolipo-
Alzheimer’s disease. Neurology 1993; 43: 246–49. protein E4 gene dose and the risk of Alzheimer’s disease in
39
Brayne C, Calloway P. The association of education and socio- late onset families. Science 1993; 261: 921–23.
47
economic status with the Mini-Mental State Examination Pericak-Vance M A, Johnson C C, Rimmler J B et al.
and the clinical diagnosis of dementia in elderly people. Alzheimer’s disease and apolipoprotein E-4 allele in an
Age Ageing 1990; 19: 91–96. Amish population. Ann Neurol 1996; 39: 700–04.
40
O’Connor D W, Pollitt P A, Treasure F P. The influence of edu-
cation and social class on the diagnosis of dementia in a
community population. Psychol Med 1991; 21: 219–24. (Revised version received June 1996)You can also read
