n-2018/36/1742-007802.00/0 "Toe Jounsat o» Nunvous avo Mena Disease Copii 6 1088 by The Willams & Wilkins Co Vol, 174,No.2 Printedin USA. Bedside Cognitive Screening Instruments A Critical Assessment AARON NELSON, PHD,’ BARRY S. FOGEL, MD, ano DAVID FAUST, Pa.D? Bedside cognitive screening instruments are used increasingly in clinical and research settings to detect cognitive impairment and to quantify its severity. The authors review the five most frequently cited bedside screening tests that use an interview format and require brief administration times: the Mini-Mental State Examination, the Cognitive Capacity Screening Examination, Mattis Dementia Rating Scale, Kahn's Mental Status Question- naire, and the Short Portable Mental Status Questionnaire. The tests all have adequate inter-rater reliability, and adequate test-retest reliability has been established for three of the tests. All of the tests show close correspondence dementia and are useful for the diagnosis and quantification of these syndromes. However, there is currently no evidence that the tests increase the level of diagnostic accuracy achieved through clinical examination alone. All of the tests have substantial falso-nogative rates, with false-negative errors frequent among patients with focal lesions, particularly of ‘the right hemisphere. False-positive errors may be more common among patients with less education and lower socioeconomic status. ‘The tests reviewed do not detect many types of cognitive deficit that may bear critically on differential diagnosis and case management. Suggestions are given for further research on the current measures and for the development of new screening tests that would meet a broader range of clinieal purposes. Brief, quantitative bedside cognitive screening in- struments have been used increasingly in clinical and research settings to assess the presence and severity of cognitive impairment. This paper critically reviews the five most frequently cited bedside cognitive sereens. The tests reviewed: a) use an interview for- ‘mat, b) require 30 minutes or less to administer, ¢) can be easily used at the bedside, and d) can’ be interpreted without formal neuropsychological train- ing. ‘Method of Review ‘To select the most popular tests fitting the above characteristics, we searched the Medline and Biblio- graphic Retrieval Service Psychological Abstracts Data Bases from 1975 and 1967, respectively, to the present, using various permutations of text and key words such as “screen,” “test,” “cognitive,” “organic,” “brain,” and “neuropsychology.” Identified publica- tions were reviewed and references to other bedside screening tests were obtained. For each sereening test identified by the first two steps, the test was used as the text/key word in a new search using the same data bases. Five tests appeared in at least 10 publications: Kahn’s Mental Status Questionnaire (MSQ) (Kahn et * New Bigland Sinei Hospital, 150 York Street, Stoughton, Mas- sachusetts 02072. Send reprint requests to Dr. Nelson. * Department of Peychiatry and Human Behavior, Brown Uni- versity, Providenee, Rhode Island 02912. n 4al., 1960), 48 publications; Mini-Mental State Exam- ination (MMS) (Folstein et al,, 1975), 35 publications; Short Portable Mental Status Questionnaire (SPMSQ) (Peiffer, 1975), 12 publications; Cognitive Capacity Screening Examination (CSE) (Jacobs et al., 197), 10 publications; Mattis Dementia Rating Scale (DRS) (Mattis, 1976), 10 publications. ‘The Tests Descriptions and reliability data for the five tests are presented in the following sections; published val- idation studies are abstracted in Table 1, listed by test, in chronological order of publication. Mini-Mental State Examination ‘The MMS (Folstein etal, 1975) is a general-purpose cognitive screening test consisting of 11 items and requiring 5 to 10 minutes to administer. Functions tested include orientation, registration (repetition), attention and calculation (serial sevens), recall, lan- guage (object naming, repetition, comprehension, reading and writing), a three-step praxis item, and a graphic copy of a geometric design. The score is a weighted sum of the items. The maximum score is 30 points. ‘The authors suggest that patients scoring be- low 24 are cognitively impaired. We found five tests of reliability: three in the orig- inal paper by Folstein et al. and two in a paper by Dick et al. (1984). Reliabilities ranged from .83 to 99ar NELSON et al. TABLE 1 Validation Studies of Cognitive Screening Tests "Tet No ‘Rathore ‘Subjece ‘Catala Rese ‘MMS 1 Folstein etal N= 192; 69 inpatients Correlation with Mean MMS = 9.6 for de- 0975) with depression or de- informal de- ‘mented patient; mean rmentia (mean aye scriptive diag- MMS = 19. for cognitively (6.4), 68 normal eon nosis impaired depressves; un- trols (mean age 73.9) complicated depresives and ‘ontrols scored 224 points. a Folstein etal 8 age-matched ubgroups Discrimination of MMS moans wore sigifi- as75) ‘of patients from study clinically ding: cantly diferent forthe subjects per group; nosed sub- three groups. age range 68-86 sroupe a Folstein etal. (V=14),un- Prediction ofim- Mean soores of both depressed (0975) ‘provement fl- groups improved signif lowing inpa ently; mean seores of de- ‘cognitively impaired tent treatment ments didnot “ Folstein et ot, Correlation with Mean scores of demented pa- a75) informal de- tients (N= 9) differed sig- hospital admissions sciptive diag- nificantly from mean scores (mean age 41.3) noses of other groups. Scores <20 ‘were seen only in dements, patients with “functional psyehosis,” and 1 of the 32 drug abusing patients 5 Folstein et ab Correlation with r= 78 for Verbal 1Q; r= 66 4978) ‘WAIS* {or Performance 10 6 DePaulo and N= 126; admissions to _—Coreelation with —_—_—Allpationts with exclusively Foletein ‘an adult neurology in peripheral lesions scored at cs78) patient service (mean least 24 points; 50% of pa ge 50.2) cally diagnosed tients with cerebral lesions swith cerebral scored 224 and 50% scored lesions; 26 pa- <4. tients clinically liagnosed as hhaving exclu. sively periph eral disorders 1 ‘Teal and N= 63; consecutive 1 _-Corzelation with ‘18 patients with diffuse atro- ‘Taoang, ferrals 0 radiology de radiographic phy scored 180+ 8.6; 10, 979) partment for CAT. cerebral lesions patients with focal lesions ‘alone scored 25.3 6.4; 32 patients with normal scans cored 26.4 5:6. Atrophies difered from normals sig- nificantly (p <.00). 8 Anthony et --N=9T;consecutive ad- Correlation with _—_-28 subject either delitious oF ‘at (1982), ‘missions to a medical DSM demented; 74 subjects nel ward 63% female, nosis of deli- ther delirious nor demented. 71% black, 60% age jum or demen- With scores <24 eslled ab ‘240, 75% with less tia normal, sensitivity = 87%, ‘than high choo edu specificity = 82.4%, false- cation, 4% with in- positives = 99.4%, fale- ‘come <37000/y2 negatives = 4%, All false- positives had =8 ys of eduBEDSIDE COGNITIVE SCREENING INSTRUMENTS "TABLE 1—continued 5 ‘Tet /Sty No atom Sect ‘Grae Reais 9 Cavenugh N= 280; medical inpe Correlation with Race and eociosconomie sta ‘and Wett tients (mean age 57); ‘BD? factori ts had significant main ef sein 153% female, 39% ally arranged {ects on MMS scores; sall (1983) black, 4% Hispanic, multiple positive relationship of BDI ‘mean rociosconomic gression on and MMS (7 =. p< status 44 demogzaphic 015), trend toward lower parameters (MMS scores in pationts 66 with high BDI scores. 0 Dik et at N= 125; consecutive Diagnosis eetab- 76% of patients with cognitive (1984) neurologicsl/neuro- lished clinically impairment scored <24;, surgical admissions with ad of 43% of normals scored <24 (mean age 49.9); 17 ‘computerized ‘Mean MMS scores of 14 pa: ‘ditional patients sxil tomogr tients with sight hemisphere ‘with known cognitive phy, angiogra- disease did not differ from impairment added to phy, and bio- controls; all of them seored tho sample ‘chemical tests; 2m. ‘correlation with lesion ste ‘and with new rologits clini eal diagnoses of cognitive impairment a Diek et a 37 patients from study Correlation with Correlation with Verbal IQ, 1984) 10 WAIS ccores 1r= 8; with Performance IQ,r= 86. ccse, 1 “cobs ota Correlation with Significant correlation be- as7) eyehiatrst’s tween clinical diagnosis of triste for organic disor isgnosis; no onganie mental disorder and der ‘operational ex score <20; one flae-poitive and one falte-negative. 2 Jacobs eta ychitrie inpa- 88% scored 220. 97) tients with funetions off expected in diagnoses this group 3 Jacobs etal N= 61; consecutively Coreelation with 7% scored <20; mean age of. as, admitted medial pa lab data; corre. these patients was 725.6 of tents; mean age 6 lation with patients with abnormal EEG in 10 Gs scored <20; 2 of 3 tients ‘with normal EEGs seored 320, Blood ure nitrogen ‘was significantly and nega- tively correlated with CCSE 4 Jacobs etal. «N= 25; hospital staff Scores above eut- Only one, a housekeeper, 97) off expected in scored 20. this group 5 Kaufman et -N-=59; neurology inpa-—-Corelation with 24 true-postives, 2 falae-nega- at (1978) tients neurologists? ‘tives, 18 true-negatives, 8 dingnoses of {alse-negatives. In 8, tho cognitive dys- history suggested cognitive function, based —_—_dyefunction but both the on history, (CCSE and neurologists ex- neurologic! amination were negative, ‘examination, All flse-negatives had gross snd clinical neurological disease. rental statue examination; the cutott score was 20