International Journal of Neuroscience

ISSN: 0020-7454 (Print) 1543-5245 (Online) Journal homepage: http://www.tandfonline.com/loi/ines20

Primary or Depressive Dementia: Mental Status Screening Gabriel desRosiers To cite this article: Gabriel desRosiers (1992) Primary or Depressive Dementia: Mental Status Screening, International Journal of Neuroscience, 64:1-4, 33-67, DOI: 10.3109/00207459209000533 To link to this article: http://dx.doi.org/10.3109/00207459209000533

Published online: 07 Jul 2009.

Submit your article to this journal

Article views: 13

View related articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=ines20 Download by: [New York University]

Date: 19 March 2016, At: 01:49

0 1992 Gordon and Breach Science Publishers S.A. Printed in the United States of America

Intern. J . Neuroscience, 1992, Vol. 64, pp. 33-61

Reprints available directly from the publisher Photocopying permitted by license only

PRIMARY OR DEPRESSIVE DEMENTIA: MENTAL STATUS SCREENING GABRIEL DEsROSIERS Neurobehavioural Clinic,Department of Neurology, and Department of Psychiatry, Addenbrooke’s Hospital, Cambridge, CB2 2QQ, England, UK

Downloaded by [New York University] at 01:49 19 March 2016

(Received July 8, 1991) In the course of interviewing a patient, several aspects of everyday functioning must be covered to provide a range of observations necessary to suggest a provisional diagnosis. First organized by Adolf Meyer, the mental status examination consists of several techniques which, in recent times, have been shortened, structured and standardized to cover maximal ground in minimal time. In this article, the most popular scales are reviewed psychometrically for their capacity to detect, as first-stage instruments, cognitive impairment suggestive of primary dementia in the context of varying prevalence rates and confounding factors like sensory impairments, sociodemographics and depressive states. Several of the measures are found adequate in some respects though not in others, but all of the better ones, when used as front line implements during clinical intake, regularly improve detection over base rates. An analytical method modelled on ROC procedures is then described contrasting two of them before newer instruments are considered which aim to improve sensitivity at relatively little cost to specificity. Keywords: Alzheimer’s Disease, depression, dementia, mental status screening.

To assist diagnostic intake in patients with neurobehavioral disorders, clinicians have developed relatively short ‘bedside’ screening measures to evaluate mental status. In a way, a mental status examination is akin to the physician’s quick check of a patient’s vital organs and life systems. It consists of a limited set of observations, routines and brief tests designed to reveal, with a minimum of cooperation required from the patient, mental dysfunction in the absence of any prior indication of pathology on a standard neurological workup (Strub & Black, 1985). According to Strub and Black, one of the most frequently confusing situations in behavioral neurology is when a patient presents with subtle signs of cognitive decline in the context of depressive affect and it is uncertain whether the diagnosis should be one of depressive dementia’ or one of primary dementia arising from some degenerative process like Alzheimer’s disease (AD). In this situation, the authors believe, a quick mental status examination is most suited as a first-stage assessment tool to determine all but the most ambiguous cases. Although Adolf Meyer is recognized as first organizing a procedural algorithm that systematically covered, in a hierarchical progression, several aspects of a patient’s ‘everyday’ psychological functioning, the practice itself is a much older one. Requests for reprints should be addressed to Gabriel desRosiers, Neurobehavioral Clinic, Department of Neurology, Addenbrooke’s Hospital, Cambridge, CB2 2QQ,UK. 1. The term ’pseudodementia” was introduced last century by Carl Wernicke to refer to a chronic psychiatric state mimicking mental deterioration. Although this term has often been used in the past to describe the cognitive impairment observed in some depressives, objections have also been raised about its connotations in this context. Accordingly the terms ‘depressive dementia’ and ‘primary dementia’ are used instead to refer to cognitive decline observed in depression and Alzheimer’s disease, respectively. 33

Downloaded by [New York University] at 01:49 19 March 2016

34

G. DEsROSIERS

A search of British archives revealed to Neugebauer (1979) the existence of similar mechanisms as early as in the 13th century. At that time, royal mandates concerning ‘idiots or natural fools’ and those ‘non compos mentis’-whose wit and memory happened to fail, some of whom with periods of temporary remission, lucid intervals, and even complete recovery-empowered the crown to protect, govern, and also derive profits from their estates while providing the ailing individual with custodial care. This legal incompetency jurisdiction (Prerogativa Regis) covered people of all ranks and occupations in the form of inquisitions mindful to distinguish those with normal senescence, depression, psychosis and dementia. Mental status testing served as a certification process carried out by a court-appointed official to determine a person’s orientation, memory, and basic intellectual competence like counting backwards, reading skills and abstract numerical operations, and inquiries were made into her/his personal and social daily life and habits. In his Outlines of Exurninations first published in 1918, Meyer (195 1) did much to provide guidelines on the conduct of clinical interviewing covering personal and family history, physical and neurological checkup, and mental status evaluation. A structured set of questions and tests were intended to bear on a patient’s orientation to time, place and persons, elicit data on personal identification and dates, measure recall and retention (pictorial, narratives, word paired associates, tests of overlearned material and digit span), evaluate calculation and reading, as well as ascertain the patient’s volition, capacity of activity, and attention. Inquiries also focused on affect and emotional preoccupations, psychiatric ideation and personality make-up, and clinicians were encouraged to record comprehensive narratives on each patient’s reactions in order to ”illustrate the stream of mental reality” (p. 243). It is not hard to see how, as a standard exercise, this extensive scheme could easily overtax the busy clinician for time, skills and motivation and, to a certain degree, subjectivity could still nuance the end product. In times, other objections were raised about its tractibility when faced with emergency situations or when group studies were carried out (Slaby, 1984). With later improvements, the consensus was that, when standardized, structured procedures proved more reliable and efficient than non-structured methods (Withers, & Hinton, 1971). While saving time, effort and money, they still tapped a common set of characteristics that most clinicians agreed would be essential in such an examination. Its primary features were summarized by Bell and Hall (1 977) as follows: 1. It should be designed to distinguish functional from organic patients. 2. It should be in a standardized format so examination can be repeated by different interviewers, or at different times, to evaluate change in status. 3. It should be capable to act as a legal document attesting to the specific examination of mental functioning. 4. It should also be amenable to quantification. Other requirements could be added. As a screening device, acceptability to patients should be emphasized. In addition, since patients with early AD often present initially with specific deficits in different functions (Celsis, Agniel, Fuel, Le Tinnier, Viallard, & Demonet, 1990), a mental status examination should sample various domains of cognitive functioning while remaining as brief as possible. Truely brief versions of Meyer’s technique began to appear just before the second world war (e.g., Mayer-Gross & Guttman, 1937) leading to a streamlined set of procedures-the Clinical Tests of the Sensorium-which, in the 1 9 5 0 ’ ~attracted ~ considerable scrutiny on its ability to distinguish organic from functional syndromes. Roth and Hopkins (1953) for example found that certain tasks (e.g., digit span) were not successful in telling apart elderly depressives from dements but others (e.g., orientation, information) fared much better. Shapiro and associates (1 956) cast a

Downloaded by [New York University] at 01:49 19 March 2016

PRIMARY OR DEPRESSIVE DEMENTIA

35

similar verdict though, despite weeding out the unreliable subtests, Birkett and Boltuch (1977) still managed to misclassify 28% of organic patients with right hemisphere injury and some 14% of depressives. Wechsler (1945) was the first to provide normative data allowing single case assessment on some of the subscales and Walton’s (1958) validation study demonstrated that depressives and dementing patients were best contrasted when elaborate tests of new learning (Logical Memory and Associate Learning) were also incorporated in the assessment. Walton followed up his patients for two years to identify those who had originally been wrongly classified as organically dementing (8%) and others first thought depressed but confirmed organically dementing at retest (6%). As reviewed in desRosiers (in press), complex measures of learning and retention are particularly useful as second-stage assessment tools. Shrout, Skodol and Dohrenwend (1986) argued that when the prevalence of a serious condition is in the low range, a two-stage procedure is most useful particularly if the cost of administering a screen is far less than performing a criterion test for accurate diagnosis. Highly sensitive instruments that are simple, inexpensive and quick to administer provide the most viable strategy to detect primary dementia at its earliest stage and, in the long run, enable quicker management while also increasing the preparatory period during which caregivers can be helped to develop coping skills (zarit, & Zarit, 1982). At the same time specific (see below) tests also have for effect to reduce the need for costly specialized testing (e. g ., brain imaging; neuropsychological workup) in all but the most difficult cases which can then be referred to second-stage facilities (e.g., Bayer, Pathy, & Twining, 1987; Reding, Haycox, & Blass, 1985). A survey by Somerfield and colleagues (1991) revealed that, while physicians were inclined to refer ambiguous patients for neuropsychological evaluation, they were also willing to apply a quick mental status scale to avoid needless referrals. Since the sixties, numerous measures have been developed to fulfill this role and reviews were published by Cohen, Eisdorfer, and Holm (1984), Gurland (1980) and Nelson, Fogel, and Faust (1986). These however concentrated on the tests’ ability to detect organicity, their utility in charting the course of illness, and in management outcome. There has been little effort to examine their validity from a psychometric point of view. For example, none has looked at their potential in relation to different base rates (antecedent probabilities) prevailing in the various clinical settings where the scales are most often used (Meehl & Rosen, 1955). Often measures are assessed with samples mimicking a 50% base rate, a situation not reflective of most clinical settings. Additionally, little has been summarized about the tests’ susceptibility to confounding factors like sociodemographics, sensory impairments and others. This is of special importance because of the problems in defining primary dementia and their effects on gauging its prevalence. GOLD STANDARD The evaluation of a test’s accuracy depends on its relationship to some mean of establishing whether the suspected condition is truely present or absent. But in the case of Alzheimer’s disease, the most common disorder leading to primary dementia, no antemortem marker exists and a definite diagnosis can only be given on micropathologic evidence (Jones & Richardson, 1990). This means that in living patients diagnosis usually rests on clinical and behavioral criteria. Since the introduction of standardized algorithms (e.g., Mckahnn, Drachman, Folstein, Katzman, Price, & Stadlan, 1984), diagnostic accuracy has improved considerably though several stud-

Downloaded by [New York University] at 01:49 19 March 2016

36

G. DEsROSIERS

ies still report a substantial number of patients being misdiagnosed in Britain (Burns, Luthert, Levy, Jacoby, & Lantos, 1990) and USA (Risse, Raskind, Nochlin, Sumi, Lampe, & Bird, 1990). One reason for this is that not all neuropathologists use the same criteria, some requiring that neurofibrillary tangles and senile plaques be present in either hippocampus, neocortex, or both (Jones & Richardson, 1990; Terry & Hansen, 1990). Conversely, other histological reports note the same presence of micropathologic findings in the brains of nondemented elderly (Mann, Brown, Jones, & Davies, 1990) and affectively ill patients (Buhl & Bojsen-Moller, 1988). It would be very useful if a screen test could be compared to a 100% trustworthy reference criterion but, in the interim, clinicians must contrast results against clinical diagnosis which does not always correlate with postmortem diagnosis. This leaves the possibility that a potentially more sensitive test might identify true cases which would be regarded as false positives by an old standard. When the disease is not self-limited and becomes overt after some years, such as in AD, follow-ups remain the best strategy to evaluate a test’s potential. DEFINITION

For the present, the diagnosis of AD in living patients revolves around the notion of threshold of symptoms, signs and other features, the most prominent of which is a progressive memory impairment. Some ambiguities arise however when other descriptives of ‘normal’ cognitive attenuation like benign senescent forgetfulness and age-related cognitive impairment are also considered (Ritchie, 1988). This dilemma becomes more pronounced when instruments developed to assess cognitive impairment in the clinic are also used as part of the algorithm to define primary dementia in the first place. In recognition of this, staging measures were developed using other behavioral features to longitudinally characterize patients into stages from early to terminal phases. This strategy acknowledges that the earliest symptoms of primary dementia may not only manifest as poor memory performance but also in the domains of emotions, personality, comportment and others. New staging tools in turn enable clinicians to examine a mental status scale’s external validity (e.g., Cooper, Mungas, & Weiler, 1990; Davis, Morris, & Grant, 1990) for example by determining which scales, subtests, or items are most sensitive at different stages of AD (e.g., Ashford, Kolm, Colliver, Bekian, & Hsu, 1989). ACCURACY The sensitivity of a test refers to its ability to detect all positive cases (early AD patients) for whom screening is targeted. It represents the likelihood of a positive test result in a person with the target condition (true positive rate). A highly sensitive test is most useful when a person’s result is negative as it increases a clinician’s confidence that his/her clinical make-up is due to other than AD. Specificity concerns the ability of a test to exclude all negative cases (e.g., depressives) for whom testing is not relevant. It represents the likelihood of a negative test result in a patient who does not have the condition (true negative rate). A highly specific test is most useful when returning a positive score as it increases a clinician’s suspicion that the patient’s results are not due to other confounding conditions like depressive illness. Because mental status scales are never completely accurate, a trade-off between these parameters must take place (Longsreth, Koepsell, & van Belle, 1987). As graphically

PRIMARY OR DEPRESSIVE DEMENTIA

Downloaded by [New York University] at 01:49 19 March 2016

0

2

x

4

Global Deterlorallon

FIGURE I

Y 6

37

8

Scale

Mean transformed scores for three diagnostic tests according to dementia stages (GDS)

depicted in desRosiers (in press), the overlap between distribution scores of AD and depressives on a particular instrument means that when sensitivity increases specificity diminishes, and vice versa, and one empirical exercise is to determine the best cutpoint at which separation is maximized and whether the results can be replicated in successive applications of the cutpoints (Ley, 1972). As we will see shortly, this is where the effects of extraneous factors (e.g., education) are most important. The staging approach recognizes that medical conditions are seldom if ever static and so it would be misguided to think of sensitivity and specificity as invariants. Rather, sensitivity and specificity for the same instrument will fluctuate according to temporal stages in the condition tested (Begg, 1987; Chang, 1989). This is illustrated in Figure 1 fashioned after Ferris and colleagues (1986). As dementia progresses to deeper stages on one such staging measure (GDS1 to GDS7), so does impairment on Test A (which might include more perceptual items) and Test B (which might include more verbal items), but the slope is different for each curve indicating that sensitivity will favour one measure over the other depending on whether testing is instituted at point X or point Y. As in other areas of clinical assessment (e.g., tumours and computed tomography), investigative studies not taking account of the problem of case-mix will tend to stack the odds for or against an instrument being evaluated (Kent, & Larson, 1988; Longstreth, Koepsell, & van Belle, 1987). As we will see later, a more practical index for the clinician in this context is the predictive power of a test which takes account of the antecedent probabilities of a condition in one’s clinical setting. PREVALENCE Problems in the definition of primary dementia in turn affect estimates of its prevalence. As in Jorm and associates’ (1987) report however, approximate values can be derived when compiling several studies and Table 1 identifies five distinct settings where clinicians regularly confront dementia. Emphasis is put on studies reporting prevalence for both organic dementia and clinical depression in the same survey. Of note is that these surveys often did not distinguish between mild and severe cases

G. DEsROSIERS

38

TABLE I Recent Prevalence Studies of Cognitive Impairment and Clinical Depression in Five Different Settings Setting Community Kay (1985) Weissman (1985) Regier (1988) Bland (1988) O’Connor (1989)

Downloaded by [New York University] at 01:49 19 March 2016

Lindesay ( 1989)

N

Age

Site

How

158 116 2588 4087 5702 363 363 358 2179 132 542 348

70-79 80+ 65 + 45-64 65 + 45-54 55-64 65 + 75-89 90+ 65-74 75 +

AUS AUS US A USA US A CAN CAN CAN GB GB GB GB

GMS GMS DIS DIS DIS DIS DIS DIS CAMDEX CAMDEX CARE CARE

266 70 217 131

60+ 65+ 75* 65+

USA

CAN GB GB

Dementia (Memory) Clinic Maletta (1982) Bayer (1987) Philpot (1987) Van Der Cammen (1987)

100

100 100 50

60+ 50-89 67* 61-90

USA GB GB GB

3.4 1.2 4.9 1.4 1 .o

3.2 4.5 18.9 0.3 2.2

4.6 13.6 1.3 7.1

I .9 0.9 1.7 2.0 0.7 3.1 4.5 1.1

4.1 3.6

DSM2 DSM3 DSM3 COM

46 51 25 17

23 17 23 24

COM DSM3 DSM3 COM

53 58 50 72

24 21 12 10

36

44

30 30

38

22

Mild to Severe 302 537

65+ 84*

ITALY USA

Psychiatric Hospital Blessed (1982)

1.3 6.1

Mild to Severe

Nursing Homes Spagnoli (1986) Parmalee ( 1989)

10.8 16.4 12.7

Clincial Depression

Mild to Severe

Liaison Psychiatry Popkin (1984) Mainprize (1987) Scott (1 988) Poynton (1988)

Cognitive Mild Sev

DSM3 DSM3R Mild to Severe

320

65+

GB

COM

*Mean Age; DIS: Diagnostic Interview Schedule; COM: Combination of Criteria; SEV: Severe; HOW: Diagnostic Criterion; GB: Great Britain; CAN: Canada; AUS: Australia.

of dementia. The evidence indicates that base rates across the severity spectrum can vary for both conditions depending on which setting is considered and clinicians must adjust estimates accordingly. Other difficulties are that some surveys tend to treat diagnostic categories in a hierarchical fashion. Because some overlap between the categories of major depressive illness and primary dementia often leads to misdiagnosis (e.g., Alexopoulos, 1989; Kral & Emery, 1989), a more suitable strategy might be to pool figures derived from large samples reclassified at follow-up (e.g., Reding et al., 1985). Kassirer and Kopelman ( 1989) identified several stages where decisions can lead to diagnostic errors. Among others, failure to consider the prevalence of conditions as they occur in the work place can increase confusion particularly when differential diagnosis is already difficult. When clinical features are not specific enough to warrant a definite diagnosis, further information is often collected from specialized tests drawing on normative tables. In certain situations, the use of highly reliable measures can still lead to more diagnostic errors than if allocating a decision on the basis of

Downloaded by [New York University] at 01:49 19 March 2016

PRIMARY OR DEPRESSIVE DEMENTIA

39

prevalence alone. Unless a clinician has a test with 100% sensitivity and specificity specialized assessment may not benefit decision taking when base rates are extremely high or low (Fletcher, Fletcher, & Wagner, 1988; Rosenquist, 1989). The probability of a condition being present after test results are known (post-test probability) depends on the sensitivity and specificity of a test interacting with antecedent (pre-test) probability. Often studies report on the sensitivity and specificity of screening scales derived with populations sampled in artificial ratios. A more practical estimate is when an instrument’s potential is assessed under realistic field conditions and one way to do this is to consider a test’s predictive power by taking base rates into account. Whether to order a test or not depends on its sensitivity and specificity but once results are known these are no longer of prime importance since they estimate the probabilities that a test will be positive (or negative) in people already known to have (or not) AD. For a clinician, the more urgent question is whether a patient with a positive test score is in fact showing signs of primary dementia (positive predictive value) and, more importantly for mental status screens, whether those with normal scores are indeed free from primary dementia (negative predictive value). The more sensitive a test is, the greater its negative predictive power and the more specific it is, the greater its positive predictive power, and equations to work out these values are described in desRosiers (in press). Before we examine selected mental status scales along these lines a few words should be said about common complications affecting their interpretation. Whether it is carried out in a hospital, outpatient clinic, the physician’s office, or at home as part of an epidemiological survey, mental testing can be complicated by a host of problems. Cognitive difficulties often increase with age and so does the rate at which these problems interfere with reliable assessment. Though it is likely that a skilled interviewer could manage to secure a reasonably cooperative response on the patient’s part, it is not always safe to assume so. For individuals who still retain some insight into their intellectual difficulties, a cognitive examination can prove anxiety provoking. Adjustment to new surroundings (e.g., hospital) may take longer in some elderly and there may be difficulties with motivation and cooperation depending on the purpose of the assessment. Some dissimulation for example can be expected if the evaluation is related to commitment procedures (Gurland, 1980). When depressive illness is present, manipulative or evasive behaviour can be an understandable reaction in someone who thinks the interview might be best focused on intrapsychic, not intellectual, discomforts. If not accounted for, agitation, sensory attenuation, ongoing medical indispositions, and others can all lead to erroneous conclusions due to breaks in communication. SENSORY ATTENUATION Visual, hearing, and speech i m m m e n t s often reduce performance on cognitive tasks despite normal intellectual functioning. This means that standardized testing conditions (e.g., display size, repeating instructions) are sometimes infringed and results may not be strictly comparable against normative data. Sometimes this is further complicated by an elderly’s concealing (or ignorance) of the problems. Questions not heard properly could elicit seemingly inappropriate or no answer at all and so complicate the task of discerning whether the behaviour is due to sensory or cognitive impairment. A vacuum cleaner left running during a home interview might be thought of as an indication of dementia rather than a consequence of hearing deficiency (Sal-

Downloaded by [New York University] at 01:49 19 March 2016

40

G . DEsROSIERS

omon, 1986). More complex is the observation that hearing dysfunctions and primary dementia seem to coexist to a certain degree (Colsher & Wallace, 1990). The interaction between the two ailments is not yet clear, but because hearing difficulties can often be remediated, it becomes imperative to consider to what extent an observed dementia is a consequence of hearing deterioration. in general, if the handicap contributes to cognitive difficulties one might look for higher rates in diagnosed organic dementia and there is some recent evidence supporting this suggestion (Uhlmann, Larson, Rees, Koepsell, & Duckert, 1989; Cutler & Grams, 1988). Thus the risk of dementia perhaps increases with progressively more impaired hearing even after controlling for other confounders like age, depression, medication, etc. Alternatively, the infirmity might facilitate the detection of dementia and amendments for its effects on standard cognitive tools may not always be justified unless correct performance can be evinced in some other way. Uhlmann, Ten, Rees, Mozlowski, and Larson ( I 989) for example found that patients with primary dementia and hearing disability did not obtain a higher mental status score when a standard instrument was given in a written format. More research is required to clarify the relationship between auditory acuity and cognitive functioning in AD patients and to what extent a mild unrecognized hearing deficit can confound cognitive assessment on mental status scales. SOCIODEMOGRAPHICS Educational achievement, ethnic and cultural make-up, primary language of communication, and age can all colour results on screen tests, even on relatively simple subscales like orientation to time (OTA, 1988). For instance, a bilingual patient might alternate between native and second language during testing leaving the clinician somewhat ambivalent as to whether the behaviour indicates resistance to testing, normal idiosyncratic patterns, regression associated with primary dementia or fluctuating motivation (effort) due to depression. Suggestions to counteract sociodemographic contaminations include searching for culture fair tests, add more nonverbal material, translate existing measures or use interpreters. Another way is to psychometrically adjust scoring procedures but, currently, there is a debate as to how judicious this strategy is with respect to educational attainments. Gurland (1980) for example entertained the notion that less educated persons are more at risk of developing primary dementia in later life through associations with lifestyle hazards, occupational exposures or others. Is low education of etiologic importance in primary dementia or is it simply the product of individual differences which should be counterbalanced when cognitive testing is carried out? In other words, d o people with poor education eventually develop dementia at a greater rate than those better educated and, if so, is this linked to increasing prevalence of brain pathology, psychiatric conditions, etc? In part this suggestion reflects our limited knowledge of the conditions causing primary dementia and this question for example should be easier to tackle looking at disorders with more predictable etiological mechanisms such as Huntington’s disease. This debate is important because a number of neuropsychological measures provide sociodemographic adjustments whereas few mental status questionnaires do, thus creating the right atmosphere for diagreement across assessment procedures. Berkman (1986) has argued that educational levels could be the end product of many factors, social, environmental, behavioural and biological. The case is made that with vascular dementia, risk factors like diet, control of blood pressure, or smoking

PRIMARY OR DEPRESSIVE DEMENTIA

41

appear more prominent in the lower educational echelons. Any adjustment for the latter would have obscured its association with vascular conditions like strokes. Reviewing the literature, Henderson (1988) concluded that the evidence for a gradient in the incidence of AD across educational levels was incoherent and suggested clinicopathological studies might be the best way to clarify this point. As there is evidence that the prevalence of reported everyday memory failures also covaries with education levels (Cutler & Grams, 1988), concurrent longitudinal surveys with different cohorts could also shed more light on this question. This is considered further when reviewing the Mini-Mental Status Exam (MMSE), a scale which, in the current literature, provides the most tested ground for this question.

Downloaded by [New York University] at 01:49 19 March 2016

PREMORBID ABILITY In a related spirit, La Rue, Dessonville and Jarvik (1985) reported a long-term study of elderly twins in which the likelihood of developing dementia seemed to covary with premorbid intellectual functioning. The question of premorbid ability is an old one, but in many instances assessment procedures do not take it into consideration when searching for AD. Yet, in most standard diagnostic systems, cognitive decline is a cardinal feature in the diagnosis of primary dementia (e.g., McKahnn et al., 1984). Part of the problem is that it is not always clear how to arrive at such estimate though the effort appears well worth it given that true decline may not occur in other conditions like depression. Currently the best technique appears to be the one developed by Nelson (1982) which relies on pronunciation of phonetically irregular words that can only be sounded out correctly by those with prior familiarity with them. Since richness of vocabulary is a good indication of intellectual levels, it is assumed that this familiarity ostensibly indicates former usage, hence premorbid intellectual functioning. Nelson’s National Adult Reading Test (NART) appears particularly well suited for AD patients in that they are invariably described to retain their pronunciation skills well into the later stages of the disease. Several studies with depressed and dementing patients confirm the utility of this approach (Crawford, Besson, Parker, Sutherland, & Keen, 1987; O’Carroll, Baikie, & Whittick, 1987) over other methods (cf. Silverstein, 1987). MMSE’s association with WAIS in AD (.81 to .83) and psychiatric (.78 to .83) patients is high (Farber, Schmitt, & Logue, 1988; Folstein, Folstein, & McHugh, 1975; Giordani, Boivin, Hall, Foster, Lehtinen, & Bluemlein, 1990) and, were its correlation with NART in a large normative sample known, adjustment rules could be derived (e.g., Bleeker, Bolla-Wilson, Kawas & Agnew, 1988). The procedure might be less than valid in the few instances when second language speakers are assessed or when premorbid abilities were below average to start with. Given the prospective nature of La Rue et al.’s (1985) study, their observation deserves further investigation as the 20 year follow-up seem too long to suggest that those with originally lower IQ were already beginning to show the effect of AD. This, as they recognized, remains an open question until more longitudinal studies come to press. Table 2 displays some information about the most often used mental status scales. Some general points can be made before each test is considered individually. First, it hasn’t always been possible to examine a measure’s pedigree with respect to degenerative dementia alone. In the past, undifferentiated groups of patients with dementia were common and, in very early studies, terms like organicity encompassed a whole gamut of conditions from hydrocephalus, strokes and closed head injury to diseases causing primary dementia. As far as possible, each study covered was as-

(70) (125)

(23) (74)

(38) (18)

(32) (196)

(49) (43)

(5)

(56) (41) (29) (336)

(28)

(15)

(47) (80)

Dementia Others

Dementia Others

Dementia Others

Dementia Others

Dementia Others

Dementia Depression

OMD Others OBS Others

Dementia Others

Dementia Depression

Dementia Others

SE

SE

SE

SE

SE

SE

SE

SE

SE

SE

SE

(14) (21)

Dementia Depression

SE

(41)

(55)

(47)

(16) (24)

75+ 75+

65+ 65+

59-86 65-85

21-92 21-92 70-79 70-79

-

56-75

65+ 65+

65+

65+

62-86 62-86

20+ 20+

78* 75*

65+ 65+

65+ 65+

(N)Age

Dementia Depression

Diagnosis

SE

t

Community Survey

Nursing Homes

Community Survey

Liaison Psychiatry Community Survey

Psychiatric Hospital

Community & Inpatients

Community Survey

Lialson Psychiatry

Liaison Psychiatry

Liaison Psychiatry

Psychiatric Hospital

Psychiatric Hospital

Setting

Cutoff

20Moderate

23Mild

20Moderate

23Mild

2 s Moderate

24Mild

.15

.40

15

23Mild

23Mild

23Mild

23Mild 15 22Mild

.35

.38

.I5 23Mild

Mild

.I5 23-

.20

.35

.20

.38

.20

Rev

.89

.80

54

.83

.52

.80

.94

1 .o

.53

.87

.30

70

87

96

87

76

80

85

55

10

82

10

52

.o

1

75

SPE

.94

SEN

.34

.80

.70

.53

.45

.71

.53

.28

1.0

.47

.87

.92

.97

75

.87

.72

1.0

.89

.92

.85

1 .o

.72

I .O

.98

PV-

.56

.48

PVC

AGECAT

DSM I11

NINCDSADRDA

Psychiatric Interview CAMDEX

DSM IIIR

CAMDEX

DSM I11

Consensus Diagnosis

Psychiatnc Interview

Consensus Diagnosis

Psychiatric Interview

Psychiatnc Interview

Criterion

TABLE 2 Recent Validation Studies of Some Bedside Screening Instruments

Depressed do not cluster in discrete low scoring group 24% of depressed still have scores 5 2 3 after treatment MMSE higher in depressed vs nondepressed dementia cases Delirious patients included in dementia sample 13% demented are depressed, 44% controls were depresed 5% of scores 5 2 3 diagnosed affectives; 13% neurotics 28% of controls diagnosed as depressed Incomplete protocols taken as failed (Fillenbaum 1988) MMSE correlates .71 with Missouri MSE Sample restricted to women from rural area 93% of cases as probable Alzheimer’s disease 22% of demented patients were depressed Five depressed patients scored below 18

Comment

Black (1990) Tables I & 2

Kafonek (1989)

Knopman (1989)

Brayne (1989)

Strain (1988)

Chandler (1988)

Roth (1986)

Folstein (1985) Table 3

Pfeffer (1984) Study 2

Anthony (1982)

Pfeffer (1981) Table 4

Folstein (1978) Figure 3

Folstein (1978) Figure 2

Reference

Downloaded by [New York University] at 01:49 19 March 2016

F

h) P

Liaison Psychiatry

78* 74*

Dementia Others

MSQ

MSQ

Community Survey Neurologic Clinic Community Survey Liaison Psychiatry Liaison Psychiatry Neurologic Clinic Liaison Psychiatry

75* 75"

22-77 22-77 65-84 65-84 65-98 65-98

-

56+ 56+ 65-88 65-84

OBS Others OBS Others OBS Functionals Dementia Others Dementia Others OMS Others OMS Others OMS Others OMS Healthy Organics Functionals

SPMSQ

SPMSQ

SPMSQ

SPMSQ

SPMSQ

CCSE

CCSE

CCSE

CCSE

CCSE

46* 46*

Nursing Homes

65+ 65+

OBS Others

SPMSQ

Dementia Clinic

Psychiatric Hospital

Nursing Home (?)

50+ 50+

SPMSQ

Community Survey (?) Dementia Clinic

-

Dementia Depression OBS Others

(19)

Community Survey

-

Dementia Others

MSQ

-

Psychiatric Hospital

SO*

81*

Dementia Others

MSQ

5Moderate

.20

7Mild

.56

.38

.35

.50

.35

.35

Mild

19-

19Mild

19Mild

19Mild

19Mild

6Mild

Mild

.I5 6-

.so

.08 (?)5Moderate

.40 6Mild

.28

8Mild 5Moderate

6Moderate

7Mild

73%Moderate

.35

.m

.15

.20

.49

1.o

.92

.62

.94

.86

.67

.27

.55

.80

.26

.68

.87

.36

.55

1.o

.89

33

.75

.90

.91

.99

1.0

.89

.98

.85

.98

.96

.95

I .o

.98

.54

Downloaded by [New York University] at 01:49 19 March 2016

.85

.78

.66

.86

.85

.58

1.0

.95

.70

.96

.93

.95

I .o

.98

.S5

.96

.86

.84

.89

.98

.86

.93

1.0

.71

.71

.78

.76

.87

.75

1.o

.83

.35

Neurologic Diagnosis

Psychiatric Interview

Psychiatric Interview

Neurologic Diagnosis

Psychiatric Interview

DSM I11

DSM III

Consensus Diagnosis

Psychiatric Interview

OARS

Psychiatric Interview

Psychiatric Interview

CARE

Consensus Diagnosis

Psychiatric Interview

Psychiatric Interview

Repeat MSQ six weeks later raised specificity to 95% Birthdate, name of previous President discriminate best 22% of controls diagnosed as depressed Modified to include new learning items No information on controls psychiabic staNs No information on controls psychiatric status Modified rating system seem more useful than original Day, birthdate, previous President discriminate best 1I % of Functionals diagnosed as Depressed Benign senescent patients considered non-impaired No correction for education in either study 35% of controls diagnosed as depressed Questionable OMS cases were considered positive 6% of scores

Primary or depressive dementia: mental status screening.

In the course of interviewing a patient, several aspects of everyday functioning must be covered to provide a range of observations necessary to sugge...
3MB Sizes 0 Downloads 0 Views