Bedside Assessment of Executive Cognitive Impairment: The Executive Interview DonaldR. Royall, MD,*t$ Roderick K. Mahurin, PhD,t and Kevin F. Gray, M D t Objective: This study is a pilot validation of the Executive Interview (EXIT), a novel instrument designed to assess executive cognitive function (ECF) at the bedside. Design: Inter-rater reliability testing and validation using inter-group comparisons across levels of care and measures of cognition and behavior. Participants:Forty elderly subjects randomly selected across four levels of care. Setting: Settings ranged from independent living apartments to designated Alzheimer’s Special Care units in a single 537bed retirement community. Measurements: The EXIT: a 10-minute, 25-item interview scored from 0-50 (higher scores = greater executive dyscontrol) was administered by a physician. Subjects were also administered the Mini-Mental State Exam (MMSE) and traditional tests of “frontal”executive function by a neuropsy-
chologist, and the Nursing Home Behavior Problem Scale (NHBPS)by Licensed Vocational Nurses. Results: Interrater reliability was high ( I = .90). EXIT scores correlated well with other measures of ECF. The interview discriminated among residents at each level of care. In contrast, the MMSE did not discriminate apartment-dwelling from residential care residents, or residential care from nursing home residents. The EXIT was highly correlated with disruptive behaviors as measured by the NHBPS ( T = .79). Conclusions: These preliminary findings suggest that the EXIT is a valid and reliable instrument for the assessment of executive impairment at the bedside. It correlates well with level of care and problem behavior. It discriminatesresidents at earlier stages of cognitive impairment than the MMSE. J Am Geriatr SOC401221-1226,1992
xecutive cognitive functions (ECF) are those processes which orchestrate relatively simple ideas, movements, or actions into complex goaldirected behavior. Without them, behaviors important to independent living, such as cooking, dressing, or self-care can be expected to break down into their component parts.’ Patients become overdependent on environmental cues, easily distracted and perseverative.’ We believe that executive deficits undermine the independence of many patients, and lead directly to the expression of common behavior problems in the nursing home. The syndrome of executive dyscontrol has traditionally been associated with frontal lobe brain damage.3,4 This condition often leads to profound changes in behavior and personality such as apathy, diminished spontaneity, self-neglect, and irritabilit~.~ These behaviors are often encountered in the course of dementing illnesses‘-’ where they contribute greatly to patient morbidity, caregiver burden, and institutionalizati~n.’,’~ They are not limited to dementia, however, and can also be recognized in major depression, Parkinson’s disease, subcortical strokes, and schizophrenia with “negative features”.””2 All these illnesses are associated with either frontal lobe structural neuropsychological deficit^,'^-'^ or hypometabolism by Positron Emission Tomography or Single Photon Emission Tomography imaging.19.20 Executive dyscontrol has not been previously studied
as a determinant of problem behavior in the elderly. Nevertheless, our clinical experience with nursin home residents suggests that it may be quite common.25 Unfortunately, executive function is poorly assessed by routine mental status testing. Tests directed at more familiar cognitive skills may miss even gross ECF deficits.” Clinically, executive dysfunction can be inferred from the results of specific so-called “frontal lobe” tests, but this often requires extensive formal neuropsychological testing. Many of these tests are impractical for routine use in demented or medically ill populations. As a result, we were led to develop a clinically based bedside screen for ECF deficits, the Executive Interview (EXIT). In this paper we report our pilot validation of the EXIT in a comprehensive extended care community and its associations with problem behavior and functional status.
E
From the Departments of *Psychiatry and tMedicine, The University of Texas Health Science Center at San Antonio; and the SAudie L. Murphy Veterans Memorial Hospital and Geriatric Research Education and Clinical Center (GRECC), San Antonio, Texas. Address correspondence and reprint requests to Dr. Donald Royall, Dept. of Psvchiatrv. The Universitv of Texas Health Science Center at San Antonio, 7/03 Floid Curl Dr., San Antonio, TX 78284-7792.
JAGS 40:1221-1226, 1992 0 1992 by the American Geriatrics Society
METHOD Development In developing the EXIT, we sought to create a brief, clinically based instrument which could be administered at the bedside, preferably by non-psychiatrically trained personnel. Our goals were to (1) discriminate the presence and seventy of ECF deficits, (2) predict ECF-related impairments in selfcare and functional status, and (3) predict problem behaviors engendered by executive dyscontrol. An initial pool of 50 test items was developed after a review of the available literature on frontal/executive function and of our own clinical exuerience with the demented elderly. Items which cohd detect frontal motor Or cognitive perseveration*23 int r u s i o n ~disinhibition, ,~~ loss of suontaneitv, imitation beha~ior,’~ environmental depenhency,” and utilization behavior25were considered for inclusion. Certain items, such as the SNOUT REFLEX, elicit ,’
0002-8614/92/$3.50
1222
ROYALL ET AL
/AGS-DECEMBER 7992-VOL. 40, NO. 1 2
recognized signs of frontal lobe dysfunction. Others, such as WORD FLUENCY, are derived from established neuropsychological procedures previously correlated with executive dyscontrol. Some, such as GO/ NO-GO TASK and ECHOPRAXIA are adapted from literature on the assessment of executive deficits, especially that of Luria5 and L’Hermitte.’’”‘ Finally, items such as “GRIPTASK” are derived from our own experience with demented patients. EXIT items are designed in such a way that they elicit a response from the subject. This distinguishes the EXIT from observational measures or rating scales which are more vulnerable to subjective interpretation. Each item is scored on a three-point basis: 0 = intact performance; 1 =
specific partial error or equivocal response; 2 = specific incorrect response or failure to perform the task. Figure 1 presents a sample item from the EXIT, and the range of responses elicited. Consensus was reached among the authors regarding the final form of the EXIT. Potential items were tested among patients presenting to a geriatric assessment clinic. Items were retained based on their ease of use at the bedside, acceptability to the elderly, and unambiguity in scoring. Those with low clinical utility or item-class correlations were excluded after empirically derived inter-item correlations were calculated. We then standardized both the order and manner of administration for the remaining items; hence the
Patient
1
2
Age/Sex
86 I F
851 M
2/50
24150
20
3
EXIT Score
Verbal Fluency ( S p o M t l O O U S Wold8
aonomtd In 60 ..c) ~~~
~~
~
~~
Design Fluency
‘Look at these pictures. Each is made with only four (4) lines. I am going to give you one minute to draw as many DIFFERENT designs as you can. The only rules are that they must each be dinerent and be drawn with four lines. Now 901’
a@-10 or more unique drawings (no copier of exampier) 1
2
-
-
5 . 9 mique drawings l a r Ihm 5 unique drawings
I ~~~
~
@ Comments
Frontal hygroma
MMSE = 29
FIGURE 1. Range of responses generated by two sample EXIT items (verbal and design fluency).
IAGS-DECEMBER 1992-VOL. 40, NO. 12
BEDSIDE ASSESSMENT OF EXECUTIVE DYSCONTROL
EXIT’s designation as an “interview.” This provides some control over the interactive effects between the examiners’ own behavior and that of their subjects. In its present form, the EXIT contains 25 items and requires 10 minutes to administer. Possible scores range from zero to a maximum of 50 points, with higher scores indicating greater executive dyscontrol. The internal consistency of the final items selected for the EXIT is high (Chronbach’s alpha = 37). Subjects Forty subjects were recruited from among 537 residents of a large, charitably-operated, not-forprofit extended care community in San Antonio, Texas. This facility provides services in four clinical settings. These include Retirement Apartments (Level 1) in which essentially no services are provided, Residentical Care Apartments (Level 2) in which laundry, housecleaning, and meals are provided, Intermediate Care Nursing Units (Level 3) in which residents receive ADL assistance, some nursing care, and medication supervision, and designated Alzheimer’s Care Units (Level 4). Residents in Levels 1 and 2 were considered “noninstitutionalized.” Those in Levels 3 and 4 were considered ”institutionalized.” Ten residents at each level of care were randomly selected for enrollment. Informed consent was obtained from the patients themselves and/or their designated proxies. Only two residents refused to participate. These were replaced by other residents, randomly selected from the same level of care. Subjects were selected independently of their diagnoses or co-morbid medical conditions. Exclusion criteria included the inability to follow a simple command, the inability to speak English, or sensory/physical deficits which precluded cooperation with testing procedures. Two residents were dropped from the study because of these criteria and replaced. The subjects’ ages ranged from 71 to 96 years (M = 84.9; SD = 6.5). The institutionalized and non-institutionalized groups did not differ statistically by age, education, or gender (2-tailed t test; P > 0.05) (Table 1).None of these variables correlated significantly with the EXIT. Procedure The EXIT was administered to each subject by a physician. A neuropsychologist, blind to the patients’ EXIT scores, separately administered a brief behavioral and neuropsychological assessment. This included the Folstein Mini-Mental State Exam (MMSE),27the Wisconsin Card Sorting Task,” Trail Making Tests A and B,29 and the Serial Attention Test.30 Disruptive behavior was assessed with the Nursing Home Behavior Problem Scale (NHBPS).31This is a 29-
1223
item behavior checklist, adapted from the Nurses’ Observation Scale for Inpatient Evaluations (NOSIE)32for use in geriatric institutional settings. Raters are asked to report the frequency of each behavior, over the prior 3-day period, according to a 5-point frequency of occurrence scale (0 = never to 4 = always). The score is calculated as the sum on the individual items, higher scores indicating a greater level of problem behavior. Among elderly nursing home residents, the average score of those in physical restraints or receiving sedative medication was 21.2 in one study, versus 1.0 for those without these intervention^.^' The average score of those without mental impairment was 8.2. The NHBPS has been shown to correlate highly with both the NOSIE and the Cohen-Mansfield Agitation Invent o r ~The . ~ NHBPS ~ was filled out by each unit’s nursing supervisor, a licensed vocational nurse who was personally familiar with the subject. These raters were blind to the results of the subject’s EXIT and neuropsychological testing. The EXIT’s inter-rater reliability was determined in a separate study. A convenience sample of 30 subjects was tested independently by two physicians. These subjects’ levels of care again ranged from independently living controls to residents of designated Alzheimer’s units. One examiner was blind to the subject’s name, original EXIT score, diagnosis, and level of care. Inter-rater reliability was found to be high (Pearson’s r = .90).
RESULTS Institutionalized and non-institutionalized residents differed significantly with respect to both the MMSE (t(38) = 4.35, P < 0.001) and the EXIT (t(38) = 5.39, P < 0.001). If these groups are further broken down, however, the EXIT retains the ability to discriminate between residents at each level of care, despite the small number of residents sampled (Table 2: ANOVA, F(3,36) = 15.2, P < 0.001). In fact, no overlap at all was found between the apartments and the dementia units (Level 4)(NewmanKeuls post-hoc comparison, P < 0.001) (Figure 2). Furthermore, the EXIT could also discriminate between those living in Level 2 and Level 3 (P < 0.05); ie, those on the borderline between institutionalization and independent living. In contrast, the MMSE could only significantly discriminate between residents at the extremes of independence (Levels 1 and 4, and Levels 2 and 4), but not between non-institutionalized residents at different levels of care (Levels 1 and 2), or across
TABLE 1. CHARACTERISTICS OF INSTITUTIONALIZED AND NON-INSTITUTIONALIZED RESIDENTS Non-Institutionalized Institutionalized Extended Care Residents Nursing Home Residents Mean (SD) Mean (SD)
%Female Age (Yrs) Education (yrs) MMSE EXIT * Two-tailed t test; P < 0.001
85 % 85.6 (6.1) 12.8 (2.6) 25.0 (5.2) 14.2 (7.5)
90% 84.1 (7.0) 13.5 (3.3) 15.2 (8.6)* 27.9 (7.6)*
1224
IAGS-DECEMBER 1992-VOL. 40, NO. 12
ROYALL ET AL
TABLE 2. EXECUTIVE INTERVIEW (EXIT) AND MINI-MENTAL STATE EXAMINATION (MMSE) SCORES DISPLAYED BY LEVEL OF CARE.","" Level 1 Level 2 Level 3 Level 4
DISCUSSION These preliminary findings suggest that the EXIT is a valid and reliable instrument that can be used to assess ECF at the bedside. Significantly between-group
10.2 (3.7)* 16.8 (7.3)* 22.9 (7.2)** 31.1 (5.5) EXIT MMSE 27.2 (3.0) 22.9 (6.2) 21.2 (7.6)** 9.2 (4.4)
50T
0
0
50
T I
0
"t4
O"
I
0
0
30
f l
W
2 m
0
z
20
0
0
0
a
0
m
0
1 8
m '
0
0
0
-t
a
0
m o m &
0
o%o0
0
0 f
0
0
10
I
20
1
30
1 50
40
EXIT FIGURE 3. Relationship between EXIT and NHBPS
1
2 3 Level of Care
4
FIGURE 2. Relationship between EXIT and level of care. 1 = Retirement Apartments; 2 = Residential Care; 3 = Intermediate Care Nursing Unit; 4 = SNF/Alzheimers Unit.
the borderline of institutionalization (Levels 2 and 3) (Table 2). Inspection of the relationship between EXIT scores and problem behavior, as measured by the NHBPS, revealed that problem behaviors were rare below EXIT scores of 20. Clinically, however, we recognized the cognitive signs of executive impairment at even lower scores (eg, 10-20). If a more conservative EXIT cut-off between 15/16 points is used, essentially n o disruptive behavior is found in "normal" subjects (Figure 3). In contrast, EXIT scores >15 were strongly correlated with a variety of common disruptive behaviors ( Y = 0.79, P < 0.01). Table 3 presents the EXIT'S ability to discriminate individuals who were rated as demonstrating specific problem behaviors in the NHBPS. The establishment of a valid "cut point" for the EXIT awaits further testing in demonstrably normal subjects. The EXIT correlated inversely with the MMSE ( Y = -.85). It also correlated strongly with traditional measures of ECF, including Trail Making Part A ( Y = .73), Trail Making Test Part B (Y = .64), and the Test of Sustained Attention and Tracking (time, r = .82; errors, Y = 3 3 ) . It was moderately correlated with the Wisconsin Card Sorting Test ( Y = .52), even though it proved difficult to engage many demented subjects in that procedure.
TABLE 3. RELATIONSHIP BETWEEN NHBPS ITEMS AND EXIT SCORES EXIT 5 15 EXIT > 15 Behavior 70 % Resists care Becomes upset easily Enters others' rooms Awakens during night Talks to self Hurts self Refuses care Fights or is aggressive Fidgets, restless Difficulty falling asleep Bathroom in inappropriate places Says nonsensical things Damages things Screams or yells Argues or threatens Gets out of bed or chair Complains about health Inappropriate sexual behavior Sees or hears things not there Disturbs others at night Wanders Accuses others Asks for attention Uncooperative with staff Paces Tries to escape Complains or whines Does things over and over Does daneerous things
0 14 0 14
0 0 14
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0
0 0 0
38 69
15 50 42 4 38 27 31 23 0 58 8 27 38 8 27 8 12 8 8 46 27 46
8 4
35 8 4
IAGS-DECEMBER 1992-VOL. 40, NO. 12
BEDSIDE ASSESSMENT OF EXECUTIVE DYSCONTROL
findings were obtained despite the small number of subjects involved. EXIT items are internally consistent, and total EXIT scores correlate well with other measures of ECF. In addition, the interview appears to be predictive of both functional status and problem behavior across a broad range of functional ability. Confirmation of these findings in larger samples will be required to establish the EXIT's cut point and validate its utility as a predictive measure of functional outcomes. Advantages of the EXIT over traditional measures of ECF include its simplicity and clinical face validity. The EXIT can be successfully administered to frail elderly subjects, even by non-medical personnel. This is in contrast to traditional measures, such as the Wisconsin Card Sorting Test, which are difficult to use in a nursing home setting and require special training to administer and interpret. Many EXIT items are derived from routine clinical procedures (such as grip testing, checking for "cogwheel" rigidity, asking the patient to say "aaah. . .," etc.) and should benefit physicians of all disciplines by alerting them to previously unrecognized signs of executive dyscontrol. Moreover, the domains of executive dyscontrol elicited by the EXIT are easily recognizable in behavior of demented patients. Common examples include a patient wearing too many sweaters (perseveration), randomly filling pots of water in the kitchen (utilization behavior, environmental dependency), following along behind the caregiver (imitation behavior), and apparent apathy which improves with prompting (lack of spontaneity). The EXIT's strong correlation with the MMSE deserves further comment. Although our interview was not designed to assess the cognitive domains which the MMSE purports to measure, some of the EXIT's items are derived from our experience with that instrument. We feel that the MMSE provides opportunities to observe executive dyscontrol that are not captured in its conventional scoring. For example, a response of "world" in the delayed memory section of the MMSE (following the earlier request to spell the word world backwards) would normally be counted as a "recall" error. The EXIT, however, recognizes this as a verbal intr~sion.'~In a frail elderly sample such as ours, patients may be expected to develop executive impairments as part of more global cognitive disorders such as Alzheimer's disease. However, with more restricted samples it may be possible to divorce the ECF deficits measured by the EXIT from more general cognitive deficits measured by the MMSE. Still, the relationships among the EXIT, the MMSE, and ECF are likely to be complex and deserve further study. The ability of the EXIT to &scriminate between residents at different levels of functional impairment, and between those who exhibit a variety of problem behaviors, is potentially very useful. The wide range of scores it elicits provides sensitivity to executive dyscontrol across a broad spectrum of cognitive function. This could be useful, for example, in the assessment of very early cognitive impairment when deficits might not be apparent on other bedside instruments. It may also prove useful in the determination of an appropriate level of care for patients nearing discharge from
1225
the hospital, or admission to a nursing home. But beyond these clinical uses, the EXIT also offers a new perspective from which to view problem behavior. We believe that many behaviors of concern to the caregivers of the demented elderly can be construed as examples of disordered executive functions. Because their behavior is no longer directed by internal goals, these patients seem to lack initiative and spontaneity. Instead, they rely on environmental cues to direct much of their behaviorz6. This has important treatment implications. First, it provides a model to explain the power of environmental stimuli to influence the behavior of demented patients. Since patients with ECF deficits cannot plan well, behaviors such as "wandering" should not be viewed as "trying to escape." Instead, one can appreciate wandering as an environmental cue's (the door's) ability to elicit "door opening" behavior. Nursing homes may be able to reduce wandering to the extent that they either (1) change the stimulus to prevent unwanted habitual associations (eg, by painting the door to look like a window, fitting it with non-familiar locks or latches, etc.), or (2) elicit more appropriate behavior (eg, using a railroad sign to prevent "crossing"). The recognition of how other features of the environment influence the behavior of demented patients could lead to changes in nursing home construction and de~ign.'~-~~ Second, this model leads to an appreciation of the power of social cues to initiate and maintain problem behavior. Caregivers can be trained to recognize how their own behavior may be affecting that of their clients and to use social cues to produce behavioral change. As an example, the first author was able to encourage a recalcitrant patient into an exam room by simply offering a "handshake." The patient could not ignore such a powerful social cue and promptly complied. The use of distraction is a similar behavioral device. Finally, the recognition of ECF deficits could lead to specific behavioral treatment plans. These deficits greatly complicate efforts at rehabilitation. Patients may fail to respond in physical/occupational treatment settings because of problems initiating (apathy, loss of spontaneity) or maintaining (impersistence, perseveration, distractibility) complex behavioral chains. Classifying rehabilitation candidates on the basis of ECF would identify a subset of patients who might respond to alternative approaches targeted directly at their executive deficits, eg, modeling, prompting, sequencing, and verbal self-instruction.37-40 These insights would be generalizable to other treatment settings where the instruction of complex goal-directed behavior is involved (eg, insulin self injection, medication compliance, glucose monitoring, catheter and ostomy care). Thus, the recognition of executive dyscontrol has treatment implications which extend far beyond the assessment of dementia, and may lead to specific behavioral, environmental, and pharmacological interventions across a variety of disabling mental disorders.
ACKNOWLEDGMENTS The authors wish to acknowledge the cooperation and support they received from the Morningside
1226
1AG.S-DECEMBER 1992-VOL. 40, NO. 12
ROYALL ET AL
Manor Retirement Community and the Southwest Neuropsychiatric Institute.
REFERENCES I Shallice 5. Specific impairments of planning. Philos Trans R Soc Lond B 1982;298:199-209. 2 . Duncan 1 Disorganization of behavior after frontal lobe damage. Cognit Neuropsychol 1986;3:271-290 3 Shallice T. Burgess P.Higher order cognitive impairments and frontal lobe lesions in man. In: Levin HS, Eisenberg HM, Benton AL, eds. Frontal Lobc Function and Dysfunction. New York: Oxford Press, 1991, pp 125-138. 4 . Stuss DT, Benson DF. The frontal lobes: Control of cognition and memory. In: Perecmen E. cd. The Frontal Lobes Revisited. New York: IRBN, 1967, pp 141-15s. 5. Luria AR. Frontal lobe syndromes. In. Vinken PJ, Bruyn GW, ed. Handbook of Clinical Neurology. Amsterdam: North-Holland, 1969, pp 725757. 6. Burns A, Jacoby R, Levy R. Psychiatric phenomena in Alzheimer's disease 1V: Disorders of behaviour. Br J Psychiatry 3990;357:86-94. 7. Rubin EH, Morris JC, Storandt M, Berg L. Behavioral changes in patients with mild senile dementia of the Alzheimer's type. Psychiatry Res 1987a; 21:55-62. 8. Rubin EH. Morris JC, Berg L. The progression of personality changes in senile dementia of the Alzheimer's type. J Am Geriatr Soc 1987b;35: 72 1-725. Y. Rabins RV, Mace NL, Lucas MS. The impact of dementia on the family. JAMA 1982;248:333-335. 10. I'ruchno RA, Resch NC. Aberrant behaviors and Alzheimer's disease: Mental health effects on spouse caregivers. 1 Gerontol 1989;44: 5177-182. 11. Nestadt G, McHugh P. The frequency and specificity of some negative symptoms. In: Huber PG, ed. Proceedings of the 6th Schizophrenia Symposium of the University Psychiatry Clinic of Bonn Stuttgart. New York: Schattaner, 1984, p,p 183-190. 12 Stuss DT. Bcnsi)n DF I\leuropsycli~~logicai studies of the frontal lobes Psychol Bull 1984;95:3-28 13. I'earson RCA, Esiri MM, Hiorns RW et 51. Anatomical correlates of the distribution of the pathological changes in the neocortex in Alzheimer's disease. I h c Natl Acad Sci USA 1985;82:4531-4534. 14. Robinson RC, Szetele B. Mood change following left hemisphere brain injury. Ann Neurol 1981;9:447-453. 15. Brown RC, Marsden CD. Internal versus external cues and the control of attention in Parkinson's disease. Brain 1988a;111:323-345. 16. Goldberg TE, Weinberger DR, Berman KF et al. Further evidence of dementia of the prefrontal type in schizophrenia: A controlled study of teaching the Wisconsin Card Sorting Test. Arch Gen Psychiatry 1987;44: 1008-1014. 17. Gotham AM, Brown RG, Marsdan CD. 'Frontal' cognitive function in patients with Parkinson's disease 'on' and 'off' levodopa. Brain 1988;111: 299-321. 18. lshii N, Nishihara Y, lmamura T. Why d o frontal lobe symptoms predominate in vascular dementia with lacunes? Neurology 1986;36:340-345. 19. Baxter LR, Schwartz JM, Phelps ME et al. Reduction of pre-frontal cortex glucose metabolism common to three types of depression. Arch Gen
Psychiatry 1989;46:243-250. 20. Paulman RG, DeVous MD, Gregory RR et al. Hypofrontality and cognitive impairment in schizophrenia: Single-photon tomography and neuropsychological assessment of schizophrenic brain function. Biol Psychiatry 1990;27:377-399.
21. Royal1 DR, Mahurin RK. ECF deficits and anorectic behavior. J Am Geriatr Soc 1991;39:840-841. 2 2 . Lezak DM. The problem of assessing executive functions. Int J Psvchol 1982;17:281-297 23. Luria AR. Two kinds of motor perseveration in massive injury of the frontal lobes. Brain 1965;88:1-10.
24. Fuld PA, Katzman R, Davies P, Terry RD. Intrusions as a sign of Alzheimer's dementia. Chemical and pathological verification. Ann Neurol 1YR2;I 1: 155-1 59.
25 L'Hermitte F, Pillon B, Serdaru M. Human anatomy and the frontal lobes. Part I: Imitation and utilization behavior. A neuropsychological study of 75 patients. Ann Neurol 1986a;19:326-334. 26. L'Hermitte F. Human anatomy and the frontal lobes. Part 11: Patient behavior in complex and social situations: The environmental dependency syndrome. Ann Neurol 1986b;19:335-343. 27. Folstein M, Folstein S, McHugh PR. 'Mini-Mental State': A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;32: 189-198. 28. Grant DA, Berg EA. A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card sorting problem. J Exp Psychol 1948;38:404-411 29. Reitan RM, Davison LA. Clinical neuropsychology: Current status and applications. New York: Hemisphere, 1974. 30. Mahurin RK, Feher EP, Cooke N, Pirozzolo FJ. The serial order reversal test in assessment of dementia. Clin Neuropsychol 1990;3:275. 31. Ray WA, Taylor JA, Lichtenstein MJ, Meador K. The Nursing Home Behavior Problem Scale. J Gerontol Med Sci 1992;47:9-16. 32. Honigfeld G, Kleh CJ. The nurse's observation scale for inpatient evaluation: A new scale for measuring improvement in chronic schizophrenia. J Clin Psychol 3965;21:65-71 33. Cohen-Mansfield J. Agitated behaviors in the elderly: Preliminary results in the cognitively deteriorated. J Am Geriatr Soc 1986;34:722-727. 34. Moos RH. Conceptualization of human environments. Am Psychol 1987; 42:652-665. 35. Faletti MV. Environmental impact on mental health and functioning in
nursing homes: Implications for research and public policy. In: Harper MS, Lebowitz BD, eds. Mental Illness in Nursing Homes: Agenda for Research. Rockville, MD: NIHM, 1986, 181-192. 36 Gottesman LE. Milieu treatment of the aged in institutions. Gerontologist 1973;13:23-26. 37. Hover WJ. Application of operant techniques to the modification of elderly behavior. Gerontologist 1973;18:18-22. 38. Hoyer WJ, Mishara BL, Riebel RG. Problem behaviors as operants. Gerontologist 1975;20:452-459. 39. Meichenbaum D.'Self-instructional strategy training: A cognitive prosthesis for the aged. Hum Dev 1974;17:273-280.
40. Thompson LW, Wagner 8, Zeiss A, Gallagher D. Cognitive behavioral therapy with early stage Alzheimer's patients: An exploratory view of the utility of this approach. In: Light E, Lebowitz BD, eds. Alzheimer's Disease Treatment and Family Stress: Directions for Research. Rockville, MD: NIMH, 1989,383-397.
chologist, and the Nursing Home Behavior Problem Scale (NHBPS)by Licensed Vocational Nurses. Results: Interrater reliability was high ( I = .90). EXIT scores correlated well with other measures of ECF. The interview discriminated among residents at each level of care. In contrast, the MMSE did not discriminate apartment-dwelling from residential care residents, or residential care from nursing home residents. The EXIT was highly correlated with disruptive behaviors as measured by the NHBPS ( T = .79). Conclusions: These preliminary findings suggest that the EXIT is a valid and reliable instrument for the assessment of executive impairment at the bedside. It correlates well with level of care and problem behavior. It discriminatesresidents at earlier stages of cognitive impairment than the MMSE. J Am Geriatr SOC401221-1226,1992
xecutive cognitive functions (ECF) are those processes which orchestrate relatively simple ideas, movements, or actions into complex goaldirected behavior. Without them, behaviors important to independent living, such as cooking, dressing, or self-care can be expected to break down into their component parts.’ Patients become overdependent on environmental cues, easily distracted and perseverative.’ We believe that executive deficits undermine the independence of many patients, and lead directly to the expression of common behavior problems in the nursing home. The syndrome of executive dyscontrol has traditionally been associated with frontal lobe brain damage.3,4 This condition often leads to profound changes in behavior and personality such as apathy, diminished spontaneity, self-neglect, and irritabilit~.~ These behaviors are often encountered in the course of dementing illnesses‘-’ where they contribute greatly to patient morbidity, caregiver burden, and institutionalizati~n.’,’~ They are not limited to dementia, however, and can also be recognized in major depression, Parkinson’s disease, subcortical strokes, and schizophrenia with “negative features”.””2 All these illnesses are associated with either frontal lobe structural neuropsychological deficit^,'^-'^ or hypometabolism by Positron Emission Tomography or Single Photon Emission Tomography imaging.19.20 Executive dyscontrol has not been previously studied
as a determinant of problem behavior in the elderly. Nevertheless, our clinical experience with nursin home residents suggests that it may be quite common.25 Unfortunately, executive function is poorly assessed by routine mental status testing. Tests directed at more familiar cognitive skills may miss even gross ECF deficits.” Clinically, executive dysfunction can be inferred from the results of specific so-called “frontal lobe” tests, but this often requires extensive formal neuropsychological testing. Many of these tests are impractical for routine use in demented or medically ill populations. As a result, we were led to develop a clinically based bedside screen for ECF deficits, the Executive Interview (EXIT). In this paper we report our pilot validation of the EXIT in a comprehensive extended care community and its associations with problem behavior and functional status.
E
From the Departments of *Psychiatry and tMedicine, The University of Texas Health Science Center at San Antonio; and the SAudie L. Murphy Veterans Memorial Hospital and Geriatric Research Education and Clinical Center (GRECC), San Antonio, Texas. Address correspondence and reprint requests to Dr. Donald Royall, Dept. of Psvchiatrv. The Universitv of Texas Health Science Center at San Antonio, 7/03 Floid Curl Dr., San Antonio, TX 78284-7792.
JAGS 40:1221-1226, 1992 0 1992 by the American Geriatrics Society
METHOD Development In developing the EXIT, we sought to create a brief, clinically based instrument which could be administered at the bedside, preferably by non-psychiatrically trained personnel. Our goals were to (1) discriminate the presence and seventy of ECF deficits, (2) predict ECF-related impairments in selfcare and functional status, and (3) predict problem behaviors engendered by executive dyscontrol. An initial pool of 50 test items was developed after a review of the available literature on frontal/executive function and of our own clinical exuerience with the demented elderly. Items which cohd detect frontal motor Or cognitive perseveration*23 int r u s i o n ~disinhibition, ,~~ loss of suontaneitv, imitation beha~ior,’~ environmental depenhency,” and utilization behavior25were considered for inclusion. Certain items, such as the SNOUT REFLEX, elicit ,’
0002-8614/92/$3.50
1222
ROYALL ET AL
/AGS-DECEMBER 7992-VOL. 40, NO. 1 2
recognized signs of frontal lobe dysfunction. Others, such as WORD FLUENCY, are derived from established neuropsychological procedures previously correlated with executive dyscontrol. Some, such as GO/ NO-GO TASK and ECHOPRAXIA are adapted from literature on the assessment of executive deficits, especially that of Luria5 and L’Hermitte.’’”‘ Finally, items such as “GRIPTASK” are derived from our own experience with demented patients. EXIT items are designed in such a way that they elicit a response from the subject. This distinguishes the EXIT from observational measures or rating scales which are more vulnerable to subjective interpretation. Each item is scored on a three-point basis: 0 = intact performance; 1 =
specific partial error or equivocal response; 2 = specific incorrect response or failure to perform the task. Figure 1 presents a sample item from the EXIT, and the range of responses elicited. Consensus was reached among the authors regarding the final form of the EXIT. Potential items were tested among patients presenting to a geriatric assessment clinic. Items were retained based on their ease of use at the bedside, acceptability to the elderly, and unambiguity in scoring. Those with low clinical utility or item-class correlations were excluded after empirically derived inter-item correlations were calculated. We then standardized both the order and manner of administration for the remaining items; hence the
Patient
1
2
Age/Sex
86 I F
851 M
2/50
24150
20
3
EXIT Score
Verbal Fluency ( S p o M t l O O U S Wold8
aonomtd In 60 ..c) ~~~
~~
~
~~
Design Fluency
‘Look at these pictures. Each is made with only four (4) lines. I am going to give you one minute to draw as many DIFFERENT designs as you can. The only rules are that they must each be dinerent and be drawn with four lines. Now 901’
a@-10 or more unique drawings (no copier of exampier) 1
2
-
-
5 . 9 mique drawings l a r Ihm 5 unique drawings
I ~~~
~
@ Comments
Frontal hygroma
MMSE = 29
FIGURE 1. Range of responses generated by two sample EXIT items (verbal and design fluency).
IAGS-DECEMBER 1992-VOL. 40, NO. 12
BEDSIDE ASSESSMENT OF EXECUTIVE DYSCONTROL
EXIT’s designation as an “interview.” This provides some control over the interactive effects between the examiners’ own behavior and that of their subjects. In its present form, the EXIT contains 25 items and requires 10 minutes to administer. Possible scores range from zero to a maximum of 50 points, with higher scores indicating greater executive dyscontrol. The internal consistency of the final items selected for the EXIT is high (Chronbach’s alpha = 37). Subjects Forty subjects were recruited from among 537 residents of a large, charitably-operated, not-forprofit extended care community in San Antonio, Texas. This facility provides services in four clinical settings. These include Retirement Apartments (Level 1) in which essentially no services are provided, Residentical Care Apartments (Level 2) in which laundry, housecleaning, and meals are provided, Intermediate Care Nursing Units (Level 3) in which residents receive ADL assistance, some nursing care, and medication supervision, and designated Alzheimer’s Care Units (Level 4). Residents in Levels 1 and 2 were considered “noninstitutionalized.” Those in Levels 3 and 4 were considered ”institutionalized.” Ten residents at each level of care were randomly selected for enrollment. Informed consent was obtained from the patients themselves and/or their designated proxies. Only two residents refused to participate. These were replaced by other residents, randomly selected from the same level of care. Subjects were selected independently of their diagnoses or co-morbid medical conditions. Exclusion criteria included the inability to follow a simple command, the inability to speak English, or sensory/physical deficits which precluded cooperation with testing procedures. Two residents were dropped from the study because of these criteria and replaced. The subjects’ ages ranged from 71 to 96 years (M = 84.9; SD = 6.5). The institutionalized and non-institutionalized groups did not differ statistically by age, education, or gender (2-tailed t test; P > 0.05) (Table 1).None of these variables correlated significantly with the EXIT. Procedure The EXIT was administered to each subject by a physician. A neuropsychologist, blind to the patients’ EXIT scores, separately administered a brief behavioral and neuropsychological assessment. This included the Folstein Mini-Mental State Exam (MMSE),27the Wisconsin Card Sorting Task,” Trail Making Tests A and B,29 and the Serial Attention Test.30 Disruptive behavior was assessed with the Nursing Home Behavior Problem Scale (NHBPS).31This is a 29-
1223
item behavior checklist, adapted from the Nurses’ Observation Scale for Inpatient Evaluations (NOSIE)32for use in geriatric institutional settings. Raters are asked to report the frequency of each behavior, over the prior 3-day period, according to a 5-point frequency of occurrence scale (0 = never to 4 = always). The score is calculated as the sum on the individual items, higher scores indicating a greater level of problem behavior. Among elderly nursing home residents, the average score of those in physical restraints or receiving sedative medication was 21.2 in one study, versus 1.0 for those without these intervention^.^' The average score of those without mental impairment was 8.2. The NHBPS has been shown to correlate highly with both the NOSIE and the Cohen-Mansfield Agitation Invent o r ~The . ~ NHBPS ~ was filled out by each unit’s nursing supervisor, a licensed vocational nurse who was personally familiar with the subject. These raters were blind to the results of the subject’s EXIT and neuropsychological testing. The EXIT’s inter-rater reliability was determined in a separate study. A convenience sample of 30 subjects was tested independently by two physicians. These subjects’ levels of care again ranged from independently living controls to residents of designated Alzheimer’s units. One examiner was blind to the subject’s name, original EXIT score, diagnosis, and level of care. Inter-rater reliability was found to be high (Pearson’s r = .90).
RESULTS Institutionalized and non-institutionalized residents differed significantly with respect to both the MMSE (t(38) = 4.35, P < 0.001) and the EXIT (t(38) = 5.39, P < 0.001). If these groups are further broken down, however, the EXIT retains the ability to discriminate between residents at each level of care, despite the small number of residents sampled (Table 2: ANOVA, F(3,36) = 15.2, P < 0.001). In fact, no overlap at all was found between the apartments and the dementia units (Level 4)(NewmanKeuls post-hoc comparison, P < 0.001) (Figure 2). Furthermore, the EXIT could also discriminate between those living in Level 2 and Level 3 (P < 0.05); ie, those on the borderline between institutionalization and independent living. In contrast, the MMSE could only significantly discriminate between residents at the extremes of independence (Levels 1 and 4, and Levels 2 and 4), but not between non-institutionalized residents at different levels of care (Levels 1 and 2), or across
TABLE 1. CHARACTERISTICS OF INSTITUTIONALIZED AND NON-INSTITUTIONALIZED RESIDENTS Non-Institutionalized Institutionalized Extended Care Residents Nursing Home Residents Mean (SD) Mean (SD)
%Female Age (Yrs) Education (yrs) MMSE EXIT * Two-tailed t test; P < 0.001
85 % 85.6 (6.1) 12.8 (2.6) 25.0 (5.2) 14.2 (7.5)
90% 84.1 (7.0) 13.5 (3.3) 15.2 (8.6)* 27.9 (7.6)*
1224
IAGS-DECEMBER 1992-VOL. 40, NO. 12
ROYALL ET AL
TABLE 2. EXECUTIVE INTERVIEW (EXIT) AND MINI-MENTAL STATE EXAMINATION (MMSE) SCORES DISPLAYED BY LEVEL OF CARE.","" Level 1 Level 2 Level 3 Level 4
DISCUSSION These preliminary findings suggest that the EXIT is a valid and reliable instrument that can be used to assess ECF at the bedside. Significantly between-group
10.2 (3.7)* 16.8 (7.3)* 22.9 (7.2)** 31.1 (5.5) EXIT MMSE 27.2 (3.0) 22.9 (6.2) 21.2 (7.6)** 9.2 (4.4)
50T
0
0
50
T I
0
"t4
O"
I
0
0
30
f l
W
2 m
0
z
20
0
0
0
a
0
m
0
1 8
m '
0
0
0
-t
a
0
m o m &
0
o%o0
0
0 f
0
0
10
I
20
1
30
1 50
40
EXIT FIGURE 3. Relationship between EXIT and NHBPS
1
2 3 Level of Care
4
FIGURE 2. Relationship between EXIT and level of care. 1 = Retirement Apartments; 2 = Residential Care; 3 = Intermediate Care Nursing Unit; 4 = SNF/Alzheimers Unit.
the borderline of institutionalization (Levels 2 and 3) (Table 2). Inspection of the relationship between EXIT scores and problem behavior, as measured by the NHBPS, revealed that problem behaviors were rare below EXIT scores of 20. Clinically, however, we recognized the cognitive signs of executive impairment at even lower scores (eg, 10-20). If a more conservative EXIT cut-off between 15/16 points is used, essentially n o disruptive behavior is found in "normal" subjects (Figure 3). In contrast, EXIT scores >15 were strongly correlated with a variety of common disruptive behaviors ( Y = 0.79, P < 0.01). Table 3 presents the EXIT'S ability to discriminate individuals who were rated as demonstrating specific problem behaviors in the NHBPS. The establishment of a valid "cut point" for the EXIT awaits further testing in demonstrably normal subjects. The EXIT correlated inversely with the MMSE ( Y = -.85). It also correlated strongly with traditional measures of ECF, including Trail Making Part A ( Y = .73), Trail Making Test Part B (Y = .64), and the Test of Sustained Attention and Tracking (time, r = .82; errors, Y = 3 3 ) . It was moderately correlated with the Wisconsin Card Sorting Test ( Y = .52), even though it proved difficult to engage many demented subjects in that procedure.
TABLE 3. RELATIONSHIP BETWEEN NHBPS ITEMS AND EXIT SCORES EXIT 5 15 EXIT > 15 Behavior 70 % Resists care Becomes upset easily Enters others' rooms Awakens during night Talks to self Hurts self Refuses care Fights or is aggressive Fidgets, restless Difficulty falling asleep Bathroom in inappropriate places Says nonsensical things Damages things Screams or yells Argues or threatens Gets out of bed or chair Complains about health Inappropriate sexual behavior Sees or hears things not there Disturbs others at night Wanders Accuses others Asks for attention Uncooperative with staff Paces Tries to escape Complains or whines Does things over and over Does daneerous things
0 14 0 14
0 0 14
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0
0 0 0
38 69
15 50 42 4 38 27 31 23 0 58 8 27 38 8 27 8 12 8 8 46 27 46
8 4
35 8 4
IAGS-DECEMBER 1992-VOL. 40, NO. 12
BEDSIDE ASSESSMENT OF EXECUTIVE DYSCONTROL
findings were obtained despite the small number of subjects involved. EXIT items are internally consistent, and total EXIT scores correlate well with other measures of ECF. In addition, the interview appears to be predictive of both functional status and problem behavior across a broad range of functional ability. Confirmation of these findings in larger samples will be required to establish the EXIT's cut point and validate its utility as a predictive measure of functional outcomes. Advantages of the EXIT over traditional measures of ECF include its simplicity and clinical face validity. The EXIT can be successfully administered to frail elderly subjects, even by non-medical personnel. This is in contrast to traditional measures, such as the Wisconsin Card Sorting Test, which are difficult to use in a nursing home setting and require special training to administer and interpret. Many EXIT items are derived from routine clinical procedures (such as grip testing, checking for "cogwheel" rigidity, asking the patient to say "aaah. . .," etc.) and should benefit physicians of all disciplines by alerting them to previously unrecognized signs of executive dyscontrol. Moreover, the domains of executive dyscontrol elicited by the EXIT are easily recognizable in behavior of demented patients. Common examples include a patient wearing too many sweaters (perseveration), randomly filling pots of water in the kitchen (utilization behavior, environmental dependency), following along behind the caregiver (imitation behavior), and apparent apathy which improves with prompting (lack of spontaneity). The EXIT's strong correlation with the MMSE deserves further comment. Although our interview was not designed to assess the cognitive domains which the MMSE purports to measure, some of the EXIT's items are derived from our experience with that instrument. We feel that the MMSE provides opportunities to observe executive dyscontrol that are not captured in its conventional scoring. For example, a response of "world" in the delayed memory section of the MMSE (following the earlier request to spell the word world backwards) would normally be counted as a "recall" error. The EXIT, however, recognizes this as a verbal intr~sion.'~In a frail elderly sample such as ours, patients may be expected to develop executive impairments as part of more global cognitive disorders such as Alzheimer's disease. However, with more restricted samples it may be possible to divorce the ECF deficits measured by the EXIT from more general cognitive deficits measured by the MMSE. Still, the relationships among the EXIT, the MMSE, and ECF are likely to be complex and deserve further study. The ability of the EXIT to &scriminate between residents at different levels of functional impairment, and between those who exhibit a variety of problem behaviors, is potentially very useful. The wide range of scores it elicits provides sensitivity to executive dyscontrol across a broad spectrum of cognitive function. This could be useful, for example, in the assessment of very early cognitive impairment when deficits might not be apparent on other bedside instruments. It may also prove useful in the determination of an appropriate level of care for patients nearing discharge from
1225
the hospital, or admission to a nursing home. But beyond these clinical uses, the EXIT also offers a new perspective from which to view problem behavior. We believe that many behaviors of concern to the caregivers of the demented elderly can be construed as examples of disordered executive functions. Because their behavior is no longer directed by internal goals, these patients seem to lack initiative and spontaneity. Instead, they rely on environmental cues to direct much of their behaviorz6. This has important treatment implications. First, it provides a model to explain the power of environmental stimuli to influence the behavior of demented patients. Since patients with ECF deficits cannot plan well, behaviors such as "wandering" should not be viewed as "trying to escape." Instead, one can appreciate wandering as an environmental cue's (the door's) ability to elicit "door opening" behavior. Nursing homes may be able to reduce wandering to the extent that they either (1) change the stimulus to prevent unwanted habitual associations (eg, by painting the door to look like a window, fitting it with non-familiar locks or latches, etc.), or (2) elicit more appropriate behavior (eg, using a railroad sign to prevent "crossing"). The recognition of how other features of the environment influence the behavior of demented patients could lead to changes in nursing home construction and de~ign.'~-~~ Second, this model leads to an appreciation of the power of social cues to initiate and maintain problem behavior. Caregivers can be trained to recognize how their own behavior may be affecting that of their clients and to use social cues to produce behavioral change. As an example, the first author was able to encourage a recalcitrant patient into an exam room by simply offering a "handshake." The patient could not ignore such a powerful social cue and promptly complied. The use of distraction is a similar behavioral device. Finally, the recognition of ECF deficits could lead to specific behavioral treatment plans. These deficits greatly complicate efforts at rehabilitation. Patients may fail to respond in physical/occupational treatment settings because of problems initiating (apathy, loss of spontaneity) or maintaining (impersistence, perseveration, distractibility) complex behavioral chains. Classifying rehabilitation candidates on the basis of ECF would identify a subset of patients who might respond to alternative approaches targeted directly at their executive deficits, eg, modeling, prompting, sequencing, and verbal self-instruction.37-40 These insights would be generalizable to other treatment settings where the instruction of complex goal-directed behavior is involved (eg, insulin self injection, medication compliance, glucose monitoring, catheter and ostomy care). Thus, the recognition of executive dyscontrol has treatment implications which extend far beyond the assessment of dementia, and may lead to specific behavioral, environmental, and pharmacological interventions across a variety of disabling mental disorders.
ACKNOWLEDGMENTS The authors wish to acknowledge the cooperation and support they received from the Morningside
1226
1AG.S-DECEMBER 1992-VOL. 40, NO. 12
ROYALL ET AL
Manor Retirement Community and the Southwest Neuropsychiatric Institute.
REFERENCES I Shallice 5. Specific impairments of planning. Philos Trans R Soc Lond B 1982;298:199-209. 2 . Duncan 1 Disorganization of behavior after frontal lobe damage. Cognit Neuropsychol 1986;3:271-290 3 Shallice T. Burgess P.Higher order cognitive impairments and frontal lobe lesions in man. In: Levin HS, Eisenberg HM, Benton AL, eds. Frontal Lobc Function and Dysfunction. New York: Oxford Press, 1991, pp 125-138. 4 . Stuss DT, Benson DF. The frontal lobes: Control of cognition and memory. In: Perecmen E. cd. The Frontal Lobes Revisited. New York: IRBN, 1967, pp 141-15s. 5. Luria AR. Frontal lobe syndromes. In. Vinken PJ, Bruyn GW, ed. Handbook of Clinical Neurology. Amsterdam: North-Holland, 1969, pp 725757. 6. Burns A, Jacoby R, Levy R. Psychiatric phenomena in Alzheimer's disease 1V: Disorders of behaviour. Br J Psychiatry 3990;357:86-94. 7. Rubin EH, Morris JC, Storandt M, Berg L. Behavioral changes in patients with mild senile dementia of the Alzheimer's type. Psychiatry Res 1987a; 21:55-62. 8. Rubin EH. Morris JC, Berg L. The progression of personality changes in senile dementia of the Alzheimer's type. J Am Geriatr Soc 1987b;35: 72 1-725. Y. Rabins RV, Mace NL, Lucas MS. The impact of dementia on the family. JAMA 1982;248:333-335. 10. I'ruchno RA, Resch NC. Aberrant behaviors and Alzheimer's disease: Mental health effects on spouse caregivers. 1 Gerontol 1989;44: 5177-182. 11. Nestadt G, McHugh P. The frequency and specificity of some negative symptoms. In: Huber PG, ed. Proceedings of the 6th Schizophrenia Symposium of the University Psychiatry Clinic of Bonn Stuttgart. New York: Schattaner, 1984, p,p 183-190. 12 Stuss DT. Bcnsi)n DF I\leuropsycli~~logicai studies of the frontal lobes Psychol Bull 1984;95:3-28 13. I'earson RCA, Esiri MM, Hiorns RW et 51. Anatomical correlates of the distribution of the pathological changes in the neocortex in Alzheimer's disease. I h c Natl Acad Sci USA 1985;82:4531-4534. 14. Robinson RC, Szetele B. Mood change following left hemisphere brain injury. Ann Neurol 1981;9:447-453. 15. Brown RC, Marsden CD. Internal versus external cues and the control of attention in Parkinson's disease. Brain 1988a;111:323-345. 16. Goldberg TE, Weinberger DR, Berman KF et al. Further evidence of dementia of the prefrontal type in schizophrenia: A controlled study of teaching the Wisconsin Card Sorting Test. Arch Gen Psychiatry 1987;44: 1008-1014. 17. Gotham AM, Brown RG, Marsdan CD. 'Frontal' cognitive function in patients with Parkinson's disease 'on' and 'off' levodopa. Brain 1988;111: 299-321. 18. lshii N, Nishihara Y, lmamura T. Why d o frontal lobe symptoms predominate in vascular dementia with lacunes? Neurology 1986;36:340-345. 19. Baxter LR, Schwartz JM, Phelps ME et al. Reduction of pre-frontal cortex glucose metabolism common to three types of depression. Arch Gen
Psychiatry 1989;46:243-250. 20. Paulman RG, DeVous MD, Gregory RR et al. Hypofrontality and cognitive impairment in schizophrenia: Single-photon tomography and neuropsychological assessment of schizophrenic brain function. Biol Psychiatry 1990;27:377-399.
21. Royal1 DR, Mahurin RK. ECF deficits and anorectic behavior. J Am Geriatr Soc 1991;39:840-841. 2 2 . Lezak DM. The problem of assessing executive functions. Int J Psvchol 1982;17:281-297 23. Luria AR. Two kinds of motor perseveration in massive injury of the frontal lobes. Brain 1965;88:1-10.
24. Fuld PA, Katzman R, Davies P, Terry RD. Intrusions as a sign of Alzheimer's dementia. Chemical and pathological verification. Ann Neurol 1YR2;I 1: 155-1 59.
25 L'Hermitte F, Pillon B, Serdaru M. Human anatomy and the frontal lobes. Part I: Imitation and utilization behavior. A neuropsychological study of 75 patients. Ann Neurol 1986a;19:326-334. 26. L'Hermitte F. Human anatomy and the frontal lobes. Part 11: Patient behavior in complex and social situations: The environmental dependency syndrome. Ann Neurol 1986b;19:335-343. 27. Folstein M, Folstein S, McHugh PR. 'Mini-Mental State': A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;32: 189-198. 28. Grant DA, Berg EA. A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card sorting problem. J Exp Psychol 1948;38:404-411 29. Reitan RM, Davison LA. Clinical neuropsychology: Current status and applications. New York: Hemisphere, 1974. 30. Mahurin RK, Feher EP, Cooke N, Pirozzolo FJ. The serial order reversal test in assessment of dementia. Clin Neuropsychol 1990;3:275. 31. Ray WA, Taylor JA, Lichtenstein MJ, Meador K. The Nursing Home Behavior Problem Scale. J Gerontol Med Sci 1992;47:9-16. 32. Honigfeld G, Kleh CJ. The nurse's observation scale for inpatient evaluation: A new scale for measuring improvement in chronic schizophrenia. J Clin Psychol 3965;21:65-71 33. Cohen-Mansfield J. Agitated behaviors in the elderly: Preliminary results in the cognitively deteriorated. J Am Geriatr Soc 1986;34:722-727. 34. Moos RH. Conceptualization of human environments. Am Psychol 1987; 42:652-665. 35. Faletti MV. Environmental impact on mental health and functioning in
nursing homes: Implications for research and public policy. In: Harper MS, Lebowitz BD, eds. Mental Illness in Nursing Homes: Agenda for Research. Rockville, MD: NIHM, 1986, 181-192. 36 Gottesman LE. Milieu treatment of the aged in institutions. Gerontologist 1973;13:23-26. 37. Hover WJ. Application of operant techniques to the modification of elderly behavior. Gerontologist 1973;18:18-22. 38. Hoyer WJ, Mishara BL, Riebel RG. Problem behaviors as operants. Gerontologist 1975;20:452-459. 39. Meichenbaum D.'Self-instructional strategy training: A cognitive prosthesis for the aged. Hum Dev 1974;17:273-280.
40. Thompson LW, Wagner 8, Zeiss A, Gallagher D. Cognitive behavioral therapy with early stage Alzheimer's patients: An exploratory view of the utility of this approach. In: Light E, Lebowitz BD, eds. Alzheimer's Disease Treatment and Family Stress: Directions for Research. Rockville, MD: NIMH, 1989,383-397.
