Journal ojGerontology: MED/CAL SCIENCES In the Public Domain 1999, Vol. 54A, No. 3, M157-M161

Contribution of and Depression to Behavioral Disturbances in Geropsychiatric Inpatients With

Mark E. Kunik,!" A. Lynn Snow-Turek,' Nayyar Iqbal,' Victor A. Molinari,':' Downloaded from https://academic.oup.com/biomedgerontology/article/54/3/M157/564212 by guest on 02 October 2021 Claudia A. Orengo,':' Richard H. Workrnan,' and Stuart C. Yudofsky"

'Department of , Veterans Affairs Medical Center Hospital, Houston, Texas. 2Baylor College of Medicine, Houston. 'Department of Psychology, Texas A&M University, College Station.

Background. Specific behavioral disturbances in dementia may be associated with underlying disorders such as the presence of psychosis and depression. The objective of this study was to examine the association of depression and psychosis with behavioral disturbances in geropsychiatric inpatients with dementia.

Methods. All admissions between October 1993 and May 1995 were reviewed to identify those patients admitted to the Houston Veterans Affairs Geropsychiatry Unit with a diagnosis of dementia; 208 patients were inc1uded in the study. Hierarchical regression models were constructed to explore the contribution of depressive and psychotic symp­ toms, and depression and psychosis diagnoses to Cohen-Mansfield Agitation Inventory (CMAI) scores.

Results. Both depression and psychotic symptoms were significantly and positively correlated with behavioral dis­ turbances. Psychotic symptoms were associated with aggressive behavioral symptoms, and depressive symptoms were associated with constant requests for help, complaining, and negativism. Dementia severity accounted for significant variance in CMAI scores and was positively associated with behavioral disturbance; thought disorder symptoms accounted for more behavioral disturbance variance than did depressive symptoms.

Conclusions. Both depressive and psychotic symptoms were associated with overall behavioral disturbances in patients with dementia. Psychotic symptoms and depressive symptoms were associated with different types of behav­ ioral disturbances. Our findings support the contention that underlying depression or psychosis may partially account for different behavioral disturbances and that not all behavioral disturbances should be globally labeled "agitation," Future studies should address symptom-specific treatment of behaviorally disturbed patients.

HE behavioral disturbanees that oeeur in approximately those with , but not among those with delu­ T 80% of both community-based and nursing horne sions (10). However, the investigation used only limited patients with dementia (1-4) plaee these individuals and measures für physieal aggression, verbal outbursts, and agi­ their earegivers at physieal and psyehologieal risk. Addi­ tation. In a study of 33 outpatients with Alzheimer's dis­ tionally, these behaviors are associated with high health ease or multi-infaret dementia, the delusional patients were eare utilization and early institutionalization of the affeeted more aggressive, exhibited more severe aetivity distur­ individuals (5-7). Beeause faetors assoeiated with behav­ banee, and had more eognitive impairment than nondelu­ ioral disturbanees are poorly understood, their treatments sional patients (13). are often ill-conceived, Depression and psyehosis are asso­ From our review of the existing literature (8,10,13-21) ciated with behavioral disturbanees in older adults suffer­ and our preliminary study (9), we hypothesized that behav­ ing from dementia. A purpose of this study was to explore ioral disturbanees in dementia are assoeiated with specifie the relationship among depression, psyehosis, and behav­ faetors sueh as psyehosis or eomorbid depression. In this ioral disturbanee. study we examined the assoeiation of depression and psy­ Two studies have reported that depression is associated ehosis with speeifie behavioral disturbanees in geropsyehi­ with verbal agitation in nursing horne and geropsyehiatrie atrie inpatients with dementia. inpatient populations (8,9), although another study found no association between depression and aggression when METHOD aggression was measured with a single 3-point seale ques­ We reviewed all admissions (using eomputerized medi­ tion (10). Similarly, there is evidenee to suggest that psy­ eal reeords and the unit's database) to the Houston VA ehosis, which oeeurs in nearly 50% of patients with demen­ Medieal Center Geropsyehiatry Unit between Oetober tia (11-13), is also aeontributor to behavioral disturbanees. 1993 and May 1995 to identify those with a diagnosis of A reeent study of 75 patients with Alzheimer's disease dementia. Eaeh patient admitted to the Geropsyehiatry Unit reported that aggressive behavior was more frequent among reeeives a eomprehensive evaluation by a multidiseiplinary

M157 M158 KUNIKETAL.

team including two geriatric psychiatrists, a geropsycholo­ vorced or separated, and 7% were never married. The gist, psychiatric nurses, a social worker, and a physician majority of participants were Caucasian (70%) or African assistant. The diagnostic evaluation comprises a psychiatric American (24%). Thirty-three patients (16%) had vascular history and mental status evaluation, a social history, a dementia, 31 (15%) had Alzheimer's dementia, 10 (5%) complete medical history and physieal examination, blood had alcohol-induced persisting dementia, 42 (20%) had chemistry panel, ehest radiograph, electrocardiogram, elec­ other , and 92 (44%) had mixed dementias. troencephalogram, magnetie resonance imaging or com­ Forty-nine patients (24%) had diagnoses of coexistent puted tomographie brain scan, and other tests as indicated depression and dementia, and 36 (17%) had diagnoses of by the patient's history and physical examination. coexistent psychosis and dementia. Only a small number Downloaded from https://academic.oup.com/biomedgerontology/article/54/3/M157/564212 by guest on 02 October 2021 On admission to the Geropsychiatry Unit, the severity of (6%) of patients had a diagnosis of dementia coexistent cognitive impairrnent and behavioral symptoms of patients with both depression and psychosis. are rated with the following well-validated standardized in­ Table 1 shows the means, standard deviations, and ranges struments: (a) the Mini-Mental State Examination [MMSE; for the measures in this study. Patients were mostly at the (22)], which is a valid and reliable measure of cognitive moderately demented level and had an average of four med­ impairment; (b) the Hamilton Depression Rating Scale ical illnesses. The BPRS-THOT indicates low to moderate [HDRS; (23)], which has been shown to be a valid measure levels of psychotic symptomatology, and the CMAI mean of depression in cognitively impaired older adults (24); (c) indieates severe agitation at admission in the majority of the Thought Disorder subscale [BPRS-THOT; (25)] of the the sample; the majority of participants scored in the non­ Brief Psychiatrie Rating Scale (BPRS), which assesses psy­ depressed range ofthe HDRS. chosis (hallucinations, , thought disorder, and Table 2 shows the correlations of the HDRS, BPRS­ grandiosity); and (d) the Cohen-Mansfield Agitation Inven­ THOT, and MMSE with the CMAI subscales and total tory [CMAI; (26)], which includes subscales measuring score. Both depression and psychotic symptoms were sig­ physical aggression (CMAI-1), nonaggressive physical agi­ nificantly positively correlated with each type of behavioral tation (CMAI-2), and verbal agitation (CMAI-3). Interrater disturbance, except for the lack of association between reliability as measured by intraclass correlation coefficients was greater than 0.9 for the MMSE and HDRS; 0.76 for the CMAI; and 0.60 for the BPRS. Axis I psychiatric diagnoses Table 1. Means and Standard Deviations of Cognitive by DSM-III-R [and later DSM-IV; (27,28)] criteria were Impairment, Medical Burden, Thought Disturbance, established within 2 weeks of discharge at a consensus con­ and Agitation ference attended by two geriatric psychiatrists, a geropsy­ chologist, and other members of the research team. Diag­ Measure Mean (Standard Deviation) Range noses are established using the objective tests described, MMSE 17.64 (7.78) 0-30 information from clinical observation, and information Medical Burden 4.35 (2.67) 0-16 from clinical interviews of patient and caregiver. Those BPRS-THOT 8.01 (4.25) 0-22 patients diagnosed with coexistent depression met DSM-IV HDRS 14.03 (8.40) 0-36 criteria for depression after a diagnosis of dementia was es­ CMSUM 47.64 (21.30) 0-176 tablished (DSM-IV codes 290.13, 290.21, 290.43, 293.83). CMAII 12.49 (6.49) 0-57 CMAI2 15.45 (9.31) ü-44 A diagnosis of coexistent psychosis was made if a diagno­ CMAI3 9.24 (5.45) 0-29 sis of dementia was met and hallucinations or delusions were present (DSM-IV codes 290.12, 290.20, 290.42, Notes: MMSE =Mini-Mental State Exam; Medical Burden =Number of 293.82, 293.81). Finally, the number of active medieal medical diagnoses coded on Axis Ill; BPRS-THOT =Brief Psychiatrie Rat­ ing Scale Thought Disturbance subscale; HDRS =Hamilton Depression diagnoses listed on Axis III was included as a measure of Rating Scale; CMSUM =Cohen-Mansfield total score; CMAI1 =Cohen­ medical burden. Mansfield Agitation Inventory factor 1 subscale (physical aggression); Correlations were computed to explore the associations CMAI2 =Cohen-Mansfield factor 2 subscale (nonaggressive physical agita­ of depressive and psychotic symptoms with behavioral dis­ tion); CMAI3 =Cohen-Mansfield factor 3 subscale (verbal agitation). turbance symptoms. Correlations of individual CMAI items to the HDRS and BPRS-THOT scales were calculated to investigate the specific CMAI items associated with depres­ Table 2. Association of Depressive and Psychotic Symptoms With sive and psychotic symptoms. A Bonferroni correction of Behavioral Disturbance p < .001 controlled for error due to multiple comparisons. CMAII CMAI2 CMAI3 CMSuM Hierarchical regression models were constructed to explore the contribution of depressive and psychotic symptoms, and HDRS .26* .12 .29* .28* depression and psychotic diagnoses, to CMAI scores. BPRS-THOT .39* .40* .38* .44* MMSE -.38* -.58* -.23* -.49* RESULTS Notes: CMAII = Cohen-Mansfield Agitation Inventory factor 1 sub­ The final sample of 208 inpatients with dementia in­ scale (physical aggression); CMAI2 = Cohen-Mansfield factor 2 subscale cluded 204 men and 4 women. The mean age was 72 years (nonaggressive physical agitation); CMAI3 = Cohen-Mansfield factor 3 subscale (verbal agitation); CMSUM = Cohen-Mansfield total score; (SD = 6.35; range = 58-96 years), and mean MMSE score HDRS =Hamilton Depression Rating Scale; BPRS-THOT =Brief Psy­ was 17.6 (SD = 7.78). Forty-nine percent of the sample chiatrie Rating Scale Thought Disturbance subscale. were married, 16% were widowed, 28% were either di- *p< .001. PSYCHOSIS, DEPRESSION, AND BEHAVIOR M159

HDRS and CMAI-2. MMSE was the largest univariate pre­ that with 208 subjects, five predictor variables, and a con­ dictor and was significantly negatively correlated with the servative R2 estimation of 0.2, there was a 99% probability CMAI subscales and total score. of rejecting the null hypothesis when the null hypothesis Correlations between the individual CMAI items and the should be rejected (29). HDRS and BPRS-THOT scores were also calculated to All four regression equations were significant: CMSUM investigate the specific behavioral disturbance items associ­ [F(5,202) =21.465, R2 =.347, p < .001]; CM1 [F(5,202) = ated with depression and psychosis. Psychotic symptoms 13.013, R2 =.244, p < .001]; CM2 [F(5,202) =28.295, R2 = were associated with aggressive behavioral symptoms, .412, P < .001]; CM3 [F(5,202) = 10.708, R2 = .210, p < whereas depressive symptoms were associated with con­ .001 (see Table 4)]. As expected, dementia severity was Downloaded from https://academic.oup.com/biomedgerontology/article/54/3/M157/564212 by guest on 02 October 2021 stant requests for help, complaining, and negativism (see positively associated with behavioral disturbance and ac­ Table 3). counted for significant variance and significant incremental In order to evaluate the contribution of depressive and change in variance (5%-30%). Further, the addition of psychotic symptoms and depression and psychotic diag­ depression and thought disorder symptoms and depressive noses to behavioral disturbances while controlling for the and psychotic diagnostic variables accounted for a signifi­ possible effects of dementia severity and medication usage, cant amount of incremental change in variance (11%-18%) hierarchical regression equations were calculated using the in all models. Exarnination of the standardized regression behavioral disturbance measures as the criterion variables. weights reveals that, with the exception of the model for Separate equations were computed for each criterion vari­ CM2, the HDRS and BPRS-THOT scores accounted for the able: CMAI-1, CMAI-2, CMAI-3, and CMAI total score vast majority of the explained incremental variance in (CMSUM). Predictor variables were dementia severity, behavioral disturbance (ß = .21 to .33), whereas depressive depression and thought disorder symptoms, and depression and psychotic diagnoses accounted for little of the variance and psychotic diagnoses. MMSE score was entered on the (ß = .02 to -.10). BPRS-THOT was the only independent first step to control for the effects of dementia severity on variable with a large standardized regression weight in the behavioral disturbance. HDRS, BPRS-THOT, and depres­ CM2 model. The standardized regression weights describe sion and psychotic diagnoses (dummy coded following a positive relationship between behavioral disturbance and convention for use of categorical variables in linear regres­ increase in depression or psychosis symptoms, but show sion) were entered together on the next step; the last step that having a diagnosis of depression or psychotic distur­ was an interaction between HDRS and BPRS-THOT. bance was associated with less behavioral disturbance-an Following convention for hierarchical regression analy­ unexpected direction of association. sis, if a step failed to contribute significant variance to the model, the step was then deleted from further analyses. DISCUSSION Medication usage was nonsignificant, and therefore omitted This study found that depressive and psychotic symp­ from all models. In addition, the interaction variable was toms were associated with the same general clusters of nonsignificant in all models and therefore omitted. behavioral disturbance symptoms in dementia patients, but A power analysis for the regression analyses determined

Table 4. Regression of Depression and Psychosis on Behavioral Disturbance Table 3. Cohen-Mansfield Items Significantly Associated With Depression and Psychosis Step 1: Step 2: -- CMAIItem BPRS-THOT HDRS MMSE HRDS BPRS-THOT DEP PSYC 1. Pacing and aimless wandering .29 CMSUM 2. Inappropriate robing or disrobing .27 ß -.329 .208 .326 -.059 -.103 4. Cursing or verbal aggression .32 R28 .165* .182* 5. Constant unwarranted requests for CMAIl or help .23 ß -.247 .211 .277 -.060 -.045 7. Hitting .35 R28 .097* .147* 8. Kicking .25 9. Grabbing .27 CMAI2 10. Pushing .36 ß -.470 .075 .285 -.102 -.140 28 11. Throwing things .27 R .301* .111 * 12. Making strange noises .24 CMAI3 13. Screaming .23 ß -.154 .196 .309 .024 -.075 16. Trying to get to a different place .26 R28 .047t .162* 17. Intentional falling .22 18. Complaining .23 Notes: MMSE =Mini Mental State Exam; HDRS =Hamilton Depres­ sion Rating Scale; BPRS-THOT Brief Psychiatrie Rating Scale Thought 19. Negativism .31 .24 = Disturbance subscale; DEP Depression diagnosis; PSYC Psychosis 21. Hurting self or other .45 = = diagnosis: CMSUM = Cohen-Mansfield Agitation Inventory total score; Notes: CMAI =Cohen-Mansfield Agitation Inventory; HDRS =Hamil­ CMAIl-CMAI3: Cohen-Mansfield Agitation Inventory subscales 1-3. ton Depression Rating Scale; BPRS-THOT = Brief Psychiatrie Rating *p < .001; tp< .005. Scale Thought Disturbance subscale. All correlations p < .001. R28 for Step 2 includes HDRS, BPRS-THOT, DEP, and PSYC. M160 KUNIKETAL.

that the individual behavioral items accounting for these produces more active symptoms whereas depression is associations differed. Psychosis and depression were each more likely to produce passive symptoms. Behavioral associated with the construct of agitation, but psychotic scales such as the CMAI place more emphasis on these symptoms were associated with aggressive behavioral symp­ active symptoms. toms, whereas depressive symptoms were associated with Surprisingly, psychosis and depression diagnoses were constant requests for help, complaining, and negativism. much less important than symptoms in accounting for all These findings support the contention that underlying de­ types of behavioral disturbanees. Further, the association pression or psychosis may partially account for different between the diagnoses and behavioral disturbances was in behavioral disturbanees, and that not all behavioral distur­ the direction of diagnoses being associated with less behav­ Downloaded from https://academic.oup.com/biomedgerontology/article/54/3/M157/564212 by guest on 02 October 2021 bances should be conceptually combined under the global ioral disturbance. This finding appears contradictory to the rubric of "agitation," We conclude that patients with de­ results concerning psychiatrie symptoms, but may be an mentia and aggressive symptoms warrant in-depth assess­ artifact of the DSM-IV diagnostic system used in the con­ ment for psychosis, and patients presenting with frequent sensus conference. Delusions or hallucinations must be requests for help and negativism warrant in-depth assess­ prominent and predominant in order to meet diagnostic cri­ ment for depression. teria for dementia with delusions or dementia with a coex­ To our knowledge, this is the first study to examine the istent psychotic disorder; similarly, depressive symptoms associations between both depressive and psychotic symp­ must be severe enough to meet the criteria for a major toms in geropsychiatric inpatients with dementia. In con­ depressive episode, and be predominant, in order to meet trast to earlier studies reporting depression to be associated diagnostic criteria for dementia with depression. In con­ only with verbal agitation (8,9), we found that depressive trast, the continuous nature of the depressive and psychotic symptoms were correlated with physical as weIl as verbal symptom rating scales allows for the detection of a positive agitation. This may be a result of our larger sample size or association between behavioral disturbances and depressive our sample of mostly male geropsychiatric inpatients. Psy­ and psychotic symptoms that are not severe enough to meet chosis was moderately positively correlated with physical criteria for depression and psychosis diagnoses. Thus, the aggression, nonaggressive physical agitation, and verbal results may reflect an association between behavioral dis­ agitation. This result agrees with several studies that have turbances and psychosis and depression, but at lower levels reported psychosis to be associated with physical aggres­ of severity than warrant a DSM-IV diagnosis of psychosis sion (13,16), or with both physical and verbal aggression or depression. (30,31). This study had severallimitations that must be acknowl­ The literature contains conflicting reports of the associa­ edged. First, the generalizability of this study is probably tion between dementia severity and behavioral distur­ limited. The sample consisted only of inpatient veterans, banees. Several studies (8,17-21) reported a strong associa­ the majority of whom were male. In addition, the relative tion between the two variables, whereas Aarsland and prevalences of dementia subtypes differ from those reported colleagues (10) did not find significant differences in be­ for the general population. This is probably due to the havioral disturbances between severity levels of dementia higher prevalences of alcohol abuse, strokes, and heart dis­ as measured by the Global Deterioration Scale. We found ease in our population. In addition, this may be due to the dementia severity to be moderately positively correlated investigators' policy of systematically including more than with all types of behavioral disturbance, and regression one DSM-IV dementia diagnosis if there was substantial analyses found dementia severity to account for a large evidence to support multiple presumed etiologies. Sec­ amount of the variance in behavioral disturbance. These ondly, this study utilized a mixed sample of dementia diag­ results suggest that future researchers should consider con­ noses; however, there is no consensus in the literature about trolling for dementia severity in their analyses of other con­ the importance of pure dementia samples. Too few patients tributing factors to behavioral disturbance. Interestingly, the in these pure subgroups prevented separate analyses. Finally, importance of dementia severity appears to vary with the although our measure of psychotic symptoms is reliable type of behavioral disturbance. The MMSE accounted for and valid, more comprehensive scales exist (32). Although 30% of the variance in physical nonaggressive agitation, our models accounted for significant amounts of variance, 16% of the variance in physical aggression, and only 5% of even our best model failed to account for 59% of the vari­ the variance in verbal aggression. Perhaps this discrepancy ance in behavioral disturbanees. The limitations of ~is between types of behavioral disturbance explains Aarsland study may have contributed to the amount of variance un­ and colleagues' (10) nonsignificant finding, as their mea­ explained. It may also be that specific cognitive deficits, sure of aggressive behavior combines both verbal and phys­ such as severity of frontal lobe deficits, may be modulating ical aggression. the association between behavioral disturbances and de­ After controlling for the effect of dementia severity, pression and psychosis. Thus, future studies should sample depressive and psychotic symptoms and diagnoses together both men and women, examine different types of demen­ accounted for a modest amount of variance in all types of tias, examine specific cognitive deficits, and utilize more in­ behavioral disturbanees. Psychotic symptoms consistently depth scales for the assessment of psychiatrie symptoms. accounted for relatively more variance than depressive The results of this study underscore the importance of symptoms; in particular, depressive symptoms contributed careful screening for depressive and psychotic symptoms in relatively little to the nonaggressive physical agitation score. patients with dementia and behavioral disturbanees. Atten­ As reflected in individual items, it may be that psychosis tion to specific psychiatrie symptoms may have important PSYCHOSIS, DEPRESSION, AND BEHAVIOR Ml61

treatment implications. In the absence of any gold standard chosis and physical aggression in probable Alzheimer's disease. Am J for treating behavioral disturbances in patients with demen­ Psyehiatry. 1991;148:1159-1163. 14. Nyth AL, Gottfries CG. The clinical efficacy of citalopram in treat­ tia, this study adds credence to treating patients with ment of emotional disturbances in dementia disorders. A Nordic mul­ depressive symptoms with antidepressants and those with ticentre study. Br J Psyehiatry. 1990;157:894-901. psychotic symptoms with . Future studies 15. Schneider LS, Pollock VE, Lyness SA. A metaanalysis of controlled should address the effect of treating behaviorally disturbed trials of neuroleptic treatment in dementia. J Am Geriatr Soe. 1990; 38:553-563. dementia patients experiencing depressive symptoms with 16. Flynn FG, Cummings JL, Gornbein J. Delusions in dementia syn­ antidepressant agents and those experiencing thought disor­ dromes: investigation of behavioral and neuropsychological corre­ der symptoms with agents. lates. J Neuropsyehiatry Clin Neurosei. 1991;3:364-370. Downloaded from https://academic.oup.com/biomedgerontology/article/54/3/M157/564212 by guest on 02 October 2021 17. Cohen-Mansfield J, Marx MS, Rosenthai AS. Dementia and agitation in nursing horne residents: how are they related? Psyehol Aging. ACKNOWLEDGMENTS 1990;5:3-8. 18. Cooper JK, Mungas D, Weiler PG. Relation of cognitive status and The authors acknowledge the compassionate care of the entire Veterans abnormal behaviors in Alzheimer's disease. J Am Geriatr Soe. Affairs Medical Center geropsychiatric staff and the careful and thoughtful 1990;38:867-870. review of the manuscript by Dr. J. D. Hamilton, Chief of Psychiatry at the 19. Gilley D, Wilson R, Bennett D, Bernard B, Fox J. Predictors of Houston VAMCHospital. behavioral disturbance in Alzheimer's disease. J Gerontol Med Sei. Address correspondence to Dr. Mark E. Kunik, Department of Psychia­ 1991;46:M362-M371. try, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. 20. Kurita A, Blass JP, Nolan KA, Black RS, Thaler HT. Relationship E-mail: [email protected] between cognitive status and behavioral symptoms in Alzheimer's dis­ ease and mixed dementia. J Am Geriatr Soe. 1993;41:732-736. 21. Miller T, Tinklenberg J, Brooks J, Fenn H, Yesavage J. Selected psy­ REFERENCES chiatrie symptoms associated with rate of cognitive decline in patients 1. Wragg RE, Jeste DV. Neuroleptics and alternative treatments: man­ with Alzheimer's disease. J Geriatr Psyehiatry Neurol. 1993;6: agement of behavioral symptoms and psychosis in Alzheimer's dis­ 235-238. ease and related conditions. Psyehiatr Clin N Am. 1988;11:195-213. 22. Folstein MF, Folstein SE, McHugh PR. "Mini-mental state." A practi­ 2. Swearer JM, Drachman DA, O'Donnel BA, Mitchell AL. Trouble­ cal method for grading the cognitive state of patients for the clinician. some and disruptive behaviors in dementia. Relationships to diagnosis J Psyehiatr Res. 1975;12:189-198. and disease severity. J Am Geriatr Soe. 1988;36:784-790. 23. Hamilton M. A rating scale for depression: J Neurol Neurosurg Psy­ 3. Cohen-Mansfield J, Billig N. Agitated behaviors in the elderly. I. A chiatry. 1960;23:56-62. conceptual review. J Am Geriatr Soe. 1986;34:711-721. 24. Mulsant BH, Sweet R, Rifai AH, Pasternak RE, McEachran A, 4. Rovner BW, Kafonek SD, Filipp L, Lucas MJ, Folstein MF. Preva­ Zubenko GS. The use of the Hamilton Rating Scale for Depression in lence of mental illness in a community nursing horne. Am J Psyehia­ elderly patients with cognitive impairment and physical illness. Am J try. 1986;143:1446-1449. Geriatr Psyehiatry. 1994;3:220-229. 5. Colenda CC, Hamer RM. Antecedents and interventions for aggres­ 25. Overall JE, Gorharn DR. The brief psychiatrie rating scale. Psyehol sive behavior of patients at a geropsychiatric state hospital. Hosp Res. 1962;10:799-812. Commun Psyehiatry. 1991;42:287-292. 26. Cohen-Mansfield J, Marx MS, Rosenthal AS. A description of agita­ 6. Moak GS. Characteristics of demented and nondemented geriatrie tion in a nursing horne. J Gerontol Med Sei. 1989;44:M77-M84. admissions to astate hospital. Hosp Commun Psyehiatry. 1990;41: 27. American Psychiatrie Association. Diagnostie and Statistieal Manual 799-801. ofMental Disorders (3rd ed., rev.). Washington, DC: American Psy­ 7. Steele C, Rovner B, Chase GA, Folstein M. Psychiatrie symptoms chiatric Press; 1987. and nursing horne placement of patients with Alzheimer's disease. Am 28. American Psychiatrie Association. Diagnostie and Statistical Manual J Psyehiatry. 1990;147:1049-1051. ofMental Disorders (4th ed.). Washington, DC: American Psychiatrie 8. Cohen-Mansfield J, Marx MS, Werner P. Agitation in elderly persons: Press; 1995. an integrative report of findings in a nursing horne. Int Psyehogeri­ 29. Cohen JC, Cohen, P. Applied Multiple Regression/Correlation Analy­ atries. 1992;4(SuppI2):221-240. sis for the Behavioral Seienees (2nd ed.). Hillsdale, NJ: Lawrence 9. Rashti DK, Kunik ME, Molinari V, Orengo CA, Workman R. Agita­ Erlbaum Associates; 1983. tion in dementia of the Alzheimer's type in a geropsychiatric inpatient 30. Kotrla KJ, Chacko RC, Harper RG, Doody R. Clinical variables asso­ unit. Clin Gerontol. 1996;16:57-62. ciated with psychosis in Alzheimer's disease. Am J Psyehiatry. 10. Aarsland D, Cummings JL, Yenner G, Miller B. Relationship of 1995;152:1377-1379. aggressive behavior to other neuropsychiatrie symptoms in patients 31. Lopez OL, Becker JT, Brenner RP, Rosen J, Bajulaiye 01, Reynolds withAlzheimer's disease. Am J Psyehiatry. 1996;153:243-247. CF. Alzheimer's disease with delusions and hallucinations: neuropsy­ 11. Drevets WC, Rubin EH. Psychotic symptoms and the longitudinal chological and electroencephalographic correlates. Neurology. 1991; course of senile dementia of the Alzheimer type. Biol Psyehiatry. 41:906-912. 1989;25:39-48. 32. Kay SR, Opler LA, Fiszbein A. Signnificance of positive and negative 12. Zubenko GS, Moossy J, Martinez AJ, et al. Neuropathologie and neu­ syndromes in chronic . Br J Psychiatry. 1986;149: rochemical correlates of psychosis in primary dementia [published 439-448. erratum appears in Areh Neurol. 1992;49:1064]. Areh Neurol. 1991;48:619-624. Reeeived June 16, 1997 13. Deutsch LH, Bylsma FW, Rovner BW, Steele C, Folstein MF. Psy- AeeeptedAugust 14, 1998