- Email: [email protected]
Some Predictors of Psychiatric Consultation in Nursing Home Residents
Some Predictors of Psychiatric Consultation in Nursing Home Residents Joanne Fenton, M.D., Allen Raskin, Ph.D. Ann L. Gruber-Baldini, Ph.D., A. Srikumar Menon, M.D., Ph.D. Sheryl Zimmerman, Ph.D., Bruce Kaup, M.D. David Loreck, M.D., Paul E. Ruskin, M.D. Jay Magaziner, Ph.D.
Objective: Despite the high rate of psychiatric disorders in nursing homes, research indicates that psychiatric consultation is requested infrequently. The authors sought to determine the rate of psychiatric consultation in a nursing home population and to assess what factors were related to a consultation request. Methods: Subjects were recruited from a stratified random sample of 59 nursing homes across Maryland. All new admissions age 65 years and older from September 1992 through March 1995 were eligible for the study. A total of 2,285 subjects were included in the study. Variables examined were factor scores from the Cornell Scale for Depression in Dementia and the Psychogeriatric Dependency Rating Scale (Behavioral Subscale), nursing home characteristics, and whether the resident had a psychiatric consultation within 90 days of admission. Results: Twenty percent of the residents (N⳱404) had a psychiatric consultation. There was no significant association with demographic variables. Behaviors that triggered a psychiatric consultation included agitation, physical/ verbal abuse, wandering, and manic/destructive acts. A psychiatric consultation was also requested when residents displayed anxiety. Surprisingly, depression in retarded and psychotic residents did not trigger a psychiatric consult. Conclusion: As expected, behavioral problems and agitation are common reasons for a psychiatric consultation. However, the resident who is depressed, particularly the quiet or retarded depressed resident, may be overlooked. In this context, it is important for the nursing staff to recognize that lethargy and social withdrawal may be signs of depression, and a referral to a psychiatrist may be in order. (Am J Geriatr Psychiatry 2004; 12:297–304)
Received March 7, 2003; revised October 22, December 3, 2003; accepted January 6, 2004. From the Mental Health Service Line, Veterans Affairs Maryland Health Care Center (JF,ASM,BK,DL,PER); the Veterans Affairs Capitol Network Mental Illness Research Education and Clinical Center (PER); the Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine (ALGB,JM); the Department of Psychiatry, University of Maryland School of Medicine (AR); and the Cecil G. Sheps Center for Health Services Research, The University of North Carolina at Chapel Hill (SZ). Send correspondence to Paul Ruskin, M.D., Mental Health Service Line (116/MH), Baltimore VAMC, 10 N. Greene St., Baltimore, MD 21201. e-mail: [email protected] 䉷 2004 American Association for Geriatric Psychiatry
Am J Geriatr Psychiatry 12:3, May-June 2004
297
Psychiatric Consultation in Nursing Homes
S
killed residential facilities such as nursing homes have, to a great extent, replaced state psychiatric facilities as the primary source of residential care for elderly patients.1 Several studies have found the prevalence of psychiatric disorders in these facilities to exceed 70%.1,2 In spite of these high rates, fewer than 1% of nursing home residents receive explicit mental health interventions, and consultation is rarely requested, even for mental health problems identified as “serious.”3 Residents with serious problems were defined as those who had more than impaired judgment or physical-restraint orders. In one study, in which two-thirds of elderly residents had a mental disorder (including dementia), only 4.5% received any mental health treatment within a 1-month period.4 A recent study of skilled nursing home facilities in six states found that the perceived need for psychiatric service is far greater than the level of services provided.5 There are few reports describing rates and correlates of psychiatric consultation in the nursing home setting. Our objectives were to document the number of psychiatric consultations requested in a large and diverse sample of nursing homes and to examine the demographic, clinical, and nursing home characteristics associated with such consultations.
METHODS Subjects Subjects were recruited from a stratified random sample of 59 nursing homes across Maryland. Nursing homes were classified by number of beds and region. At the time of the study (1992), there were 221 licensed long-term care facilities in Maryland. They were broken into five regions (Eastern Shore, Western MD, Baltimore City, Baltimore Metropolitan, Washington Metropolitan), and three size categories (⬍50 beds, 50–150 beds, and ⬎150 beds); from these, a stratified random sample of facilities was selected. All new admissions age 65 years and older from September 1992 through March 1995, who had not been a resident in a nursing home in the previous year were eligible for the study. Informed consent was obtained from either the resident or a significant other. The Institutional Review Board of the University of Maryland approved the consent procedure.
298
There were a total of 3,283 eligible subjects, of whom 2,285 were enrolled in the study. Of the 2,285 who were included in the overall study, there were 132 residents for whom there was no follow-up data collection: 80 had a short stay (less than 8 days) and were excluded from follow-up; 40 had missing charts; and 12 denied permission to collect follow-up information. For an additional 138 residents, 25% or more of the chart information was missing; these residents are excluded from the calculation of event rates (N⳱2,015), but are included in the analyses of length of stay and final disposition. There were 998 who did not participate. The major reason for not participating was that the residents or their significant others refused consent. We had chart follow-up about physician contact on 2,015 subjects. Comparison of the age and sex distributions for enrolled and non-enrolled eligible cases indicated that those enrolled were slightly older (81.5 years versus 80.6 years; p ⬍0.001) and more often female (71.6% versus 68.6%; p ⬍0.05). Evaluations for residents with fever or acute medical illness were delayed until their condition was resolved. A more detailed description of study methods has been published elsewhere.6 Measures Data were collected from multiple sources, including interviews with residents, nursing staff, and significant others; medical records, and hospital discharge summaries. Trained lay interviewers conducted all interviews and medical chart reviews. The staff member who was most knowledgeable about a particular resident completed the evaluation for that resident. Given the large number of informants, interrater reliability was not assessed. Cognitive deficit. Cognitive deficits were assessed with the Mini-Mental State Exam (MMSE).7 The MMSE was administered to every resident who did not refuse, could communicate, and for whom testing could be done within 2 months of admission (N⳱1,446; 63% of the sample). Residents were classified by severity of cognitive impairment on the basis of MMSE scores: Severe (⬍10); Moderate (10–19); Mild (20–26), and Absent (27–30).7 Dementia. The determination of dementia was made in accordance with DSM-III-R criteria8 by an expert panel consisting of geriatric psychiatrists, neurologists, and a geriatrician, using detailed information collected by lay interviewers. Information on his-
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. tory of cognitive and functional decline, current cognitive and functional status, and information about affective, social, and behavioral status were considered in making a dementia diagnosis. The panel used information concerning cognitive status obtained from the MMSE, as well as the Blessed Dementia Scale, which was administered to a family member or friend. A detailed description of the dementia ascertainment process may be found elsewhere.9 Depression. Depression symptoms were ascertained with a modified version of the Cornell Scale for Depression in Dementia (CSDD) completed by a nurse-informant. The CSDD was modified in that we asked questions to the nursing home staff instead of having a rater directly assess the resident. The CSDD is an observer-rated scale designed to rate depression in residents with dementia. It contains 19 items, and total scores range from 0 to 38. On the basis of their scale scores, residents were categorized with No Depression (score 1–7); Mild-to-Moderate Depression (score 8–20); or Severe Depression (score 21Ⳮ). These categorizations follow the recommendations of Alexopoulos and associates,10 who developed this scale. This depression scale has been validated in both dementia and non-dementia conditions.11 Agitation. Behavioral functioning was assessed by use of the behavior subscale of the Psychogeriatric Dependency Rating Scale (PGDRS), which assessed the presence of aggression, passive hostility, attention-seeking, wandering, restlessness, and fearfulness.12 The instrument was developed to evaluate care needs of older adults. The PGDRS includes 17 items rated on a frequency-of-occurrence scale as 0: none; 1: 1–2 times a week; and 2: more often. Nursing home characteristics. The nursing home in which the resident resided was characterized in terms of locality (urban or rural), size (number of beds), chain affiliation (affiliated or not affiliated), Medicarecertified (certified or not), and for-profit status (forprofit or not-for-profit). Consultation. The presence or absence of a psychiatric consultation within the first 90 days was determined through chart review. Consultations were performed by psychiatrists. Data Analysis Previous factor-analytic studies of scales assessing agitation in elderly residents of nursing homes have
Am J Geriatr Psychiatry 12:3, May-June 2004
identified at least three independent factors.13–15 Based on these results, a factor analysis of the PGDRS Behavior subscale was undertaken. We performed a principal-components analysis, followed by a normal Varimax rotation, on all factors with eigenvalues of 1 or greater. Five factors met this criterion and were extracted. An exact factor scoring method was used that includes all items in a factor, with the items assigned weights on the basis of their loading on the factor. This approach maintains the orthogonality among factors. The Physical Aggression/ Verbal Abuse factor included items with highest loadings on behaviors interfering with and disturbing other residents and staff—hitting, biting, scratching, and deliberately refusing to obey commands. The Wandering factor referred to wandering in the nursing home and pacing. Items with high loadings on the third factor (Intrusiveness) were “demanding conversation” and “getting others to do things” that the resident was capable of doing. Psychotic Behavior Items included incoherent speech and hallucinations; and the last factor, Manic/Destructive or Inappropriate Behavior, included self-destructive acts, inappropriate elation, and destroying and damaging clothes and property (Table 1). A similar factor analysis was performed on the CSDD,16 and four independent factors were extracted (Table 2). The factors were named according to the items in the scale that had the highest loadings. Retarded/Depressed items were appetite loss, weight loss, and anergy. The Anxious/Agitated factor referred to irritability and agitation. Psychotic Depression included items about suicidal ideation and mood-congruent delusions; and the last factor, Sleep Disturbance, included items on early and middle insomnia. The residents who did and did not receive a psychiatric consultation were compared on demographic variables, nursing home characteristics, and clinical variables by use of chi-square and t-tests for these comparisons. Finally, a total of 14 variables, consisting of factor scores and demographics, were included in a logisticregression analysis to distinguish which residents did or did not receive a psychiatric consultation (Table 3). Clustering was not taken into account in this analysis because previous analyses with the same sample using the method suggested by Huber17 had little effect on the statistical results.
299
Psychiatric Consultation in Nursing Homes RESULTS The mean age of the sample population was 81.4 years. Female residents accounted for 70.9% of the sample, and 83% were white. The majority of the population were widowed (62.2%), and 23.9% were married. Fifty-one percent of the sample had at least a high school education. Eighty percent of the subjects did not have a psychiatric consultation; 9.4% had one consultation; and 10.6% had more than one consultation within the first 90 days. TABLE 1.
The majority of those who had a psychiatric visit stayed beyond 90 days. Of those, 64% who were only seen once in the first 90 days were seen again by 180 days. Comparing those residents who had a psychiatric consultation with those who did not, there was no significant difference in gender, marital status, race, age, or education (Table 4). Seven of the 12 clinical variables were significantly associated with referral for a psychiatric consultation (Table 5). These included the Total score on the CSDD (t[474.2]⳱6.05; p ⬍0.00) and the Cornell Anxious/Agitated subscale (t[245.3]⳱6.28; p ⬍0.00), Total score on the PGDRS
Factors of psychopathology extracted from the Psychogeriatric Dependency Rating Scale (PGDRS)
PGDRS Interferes with activity Uninvolved Wanders Objectionable behavior Demands conversation Slow to talk Makes noises Hits Refuses to obey Verbally abusive Repeated movements Self-destruct Destroys property Too happy Paranoid delusions Hallucinations Incoherent speech
1
2
3
4
5
0.462 0.103 0.091 0.278 0.114 0.077 0.440 0.735 0.644 0.794 0.083 0.248 0.116 0.099 0.590 0.060 0.085
0.423 0.059 0.813 0.483 0.156 0.356 0.084 0.216 0.294 0.043 0.804 0.257 0.049 0.082 0.138 0.057 0.305
0.266 0.751 0.074 0.010 0.745 0.371 0.361 0.096 0.191 0.133 0.097 0.097 0.046 0.168 0.132 0.083 0.008
0.167 0.026 0.020 0.356 0.082 0.326 0.306 0.097 0.042 0.015 0.048 0.256 0.225 0.127 0.351 0.640 0.669
0.201 0.044 0.100 0.120 0.116 0.065 0.156 0.125 0.102 0.094 0.072 0.406 0.689 0.765 0.033 0.205 0.060
Factor names: 1: Physical Aggression/Verbal Abuse; 2: Wandering; 3: Intrusiveness; 4: Psychotic Behavior; 5: Manic Destructiveness.
TABLE 2.
Factors of psychopathology extracted from the Cornell Scale for Depression in Dementia (CSDD)
CSDD Anxious Sadness Lack reaction Irritability Agitation Retardation Physical complaints Loss of interest Appetite loss Weight loss Lack of energy Diurnal mood change Hard to sleep Wake often Wake earlier Suicidal ideas Low self-esteem Pessimism Delusions
1
2
3
4
0.207 0.275 0.535 0.082 0.117 0.489 0.170 0.655 0.781 0.744 0.683 0.369 0.067 0.099 0.144 0.098 0.090 0.128 0.014
0.629 0.587 0.314 0.77 0.791 0.218 0.270 0.251 0.110 0.062 0.124 0.473 0.165 0.125 0.165 0.136 0.030 0.111 0.177
0.246 0.288 0.040 0.072 0.100 0.040 0.394 0.217 0.119 0.013 0.103 0.184 0.140 0.163 0.112 0.719 0.688 0.711 0.633
0.146 0.056 0.014 0.151 0.184 0.042 0.092 0.114 0.062 0.051 0.171 0.085 0.807 0.855 0.788 0.076 0.074 0.196 0.052
Factor names: 1: Retarded/Depressed; 2: Anxious/Agitated; 3: Psychotic Depression; 4: Sleep Disturbance.
300
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. (t[441.4]⳱6.37; p ⬍0.00), PGDRS factors Physical Aggression/Verbal Abuse (t[427.9]⳱3.61; p ⬍0.00), Wandering (t[444.6]⳱3.62; p ⬍0.00), and Inappropriate Behavior (t[362.2]⳱2.54; p ⬍0.01). The presence of dementia was also significant (v2[1]⳱47.84; p ⬍0.00), with more of these residents having a consultation. With regard to nursing home characteristics, residents who were in an urban rather than a rural location had a significantly higher rate of psychiatric consultations (22% versus 12%; v2[1]⳱18.79; p ⬍0.01). Those residing in a for-profit nursing home had a higher rate of consultation than those residing in a non-profit home (21% versus 18%; v2[1]⳱4.06; p ⬍0.05). The rate of psychiatric consultation was not influenced by whether the nursing home had a chain affiliation or was Medicare-certified. Also, the number of comprehensive-care beds and number of beds in the facility did not relate to the number of consultations. On the basis of these previous findings, 14 variables were included in logistic-regression analysis as possible correlates of residents’ having a psychiatric consultation (Table 3). These included the five factor scores of the PGDRS, the four factor scores of the CSDD, and a number of demographic variables and the presence or absence of dementia. Two variables, PGDRS Manic/Destructive and the Cornell factor Anxious/Agitated, significantly discriminated patients who did and did not receive a psychiatric consultation, whereas the PGDRS Physical Aggression/ Verbal Abuse and the presence or absence of TABLE 3.
Logistic regression: demographic and clinical predictors of psychiatric consultation
Variable Sex Marital status Race Education Physical Aggression/Verbal Abuse Wandering Intrusiveness Psychotic Behavior Manic/Destruction Retarded/Depressed Anxious/Agitated Psychotic Depression Sleep Disturbance Dementia Y/N
b
Wald v2
p
0.047 0.047 0.026 0.005 0.210 0.171 0.053 0.195 0.394 0.035 0.512 0.053 0.066 0.464
0.048 0.038 0.011 0.036 3.73 3.22 0.295 3.28 8.22 0.140 26.97 0.337 0.432 5.51
0.83 0.85 0.91 0.85 0.05* 0.07 0.59 0.07 0.00** 0.71 0.00** 0.56 0.51 0.02*
Overall model df: 14, individual comparisons df: 1. *Significant at the 0.05 level (two-tailed). **Significant at the 0.01 level (two-tailed).
Am J Geriatr Psychiatry 12:3, May-June 2004
dementia showed a trend toward significance (p ⬍0.05).
DISCUSSION This study is important in alerting caregivers to the clinical and delivery-service characteristics that are associated with a psychiatric consultation. To our knowledge, this article is the first to document the percentage of nursing home residents to have a psychiatric consultation within the first 90 days after admission. Our study looks at a large number of residents in a broad range of nursing-home settings in the state of Maryland. Twenty percent is a high percentage in comparison to other studies, which estimate that only 5% of nursing home residents received mental health treatment within a 1-month period,18 and another report of 2.3% of residents who had a mental health consultation in the previous month.4 In our study, there were a greater number of consultations within 3 months after admission, in comparison with a discrete 1-month period. Both previous studies reflect rates of consultation before the Omnibus Reconciliation Act (OBRA ’87), whereas our study may reflect the greater awareness of mental health needs after its implementation.19 There was no significant difference in the age, gender, or education of the residents who had a psychiatric consultation and those who did not. These results differ from those of two other studies, which found that older residents were less likely to receive mental health treatment.4,20 However, there was a restricted age range in this study, with admissions limited to those 65 years of age or older. Reasons for psychiatric referral of nursing home residents are diverse; these include behavior problems, mood-related problems, psychotic features, somatic complaints, and impaired activities of daily living.21 Staff are often concerned with the psychosocial aspects of patients’ problems. However, they have difficulty recognizing the legitimacy of their psychological complaints.23 Common reasons for referral in a continuing-care retirement community included depression and dementia. Residents in this setting often suffer emotional distress from the loss of independence.22 Front-line staff often have poor training and low pay; staff turnover is high, and those caring
301
Psychiatric Consultation in Nursing Homes for older residents may not be sensitive to the need for psychiatric consultation.23 Often, their view of the psychiatrist is primarily as a prescriber of medication, rather than an overall manager of patient care. An additional deterrent to psychiatric consultation in the nursing home is the low level of reimbursement to the psychiatrist. Clinically, a number of symptom syndromes were associated with a psychiatric consultation in this study. Both the PGDRS Behavioral subscale and the CSDD Total score were significant. When these scores were factor-analyzed, the factors were able to give a clearer picture of the specific symptoms, which prompted a psychiatric consultation. Behavioral disturbances, including Physical Aggression, Wandering, and Inappropriate Behavior were significantly correlated with getting a consultation. Similar findings have been documented by Goldman and Klugman.24 These findings suggest that the resident who TABLE 4.
Comparison of demographic variables between patients who did and did not have a psychiatric consultation
Age, years Education, years
ConsultationN (%)
476 (80) 1,135 (80)
120 (20) 284 (20)
0.00[1]
0.95
1,236 (81) 375 (77)
294 (19) 110 (23)
2.76[1]
0.10
304 (81) 1,305 (80)
72 (19) 332 (20)
0.24[1]
0.62
N 1,610 1,357
Mean (SD) 81.52 (7.67) 10.62 (3.94)
N 404 347
Mean (SD) 80.80 (7.36) 10.78 (3.86)
CSDD Total score Cornell Factor I: Retarded/Depression Cornell Factor II: Anxious/Agitated Cornell Factor III: Psychotic Depression Cornell Factor IV: Sleep Disturbance PGDRS Behavior Total score PGDRS Behavior Factor I: Physical Aggression/Verbal Abuse PGDRS Behavior Factor II: Wandering PGDRS Behavior Factor III: Intrusiveness PGDRS Behavior Factor IV: Psychotic Behavior PGDRS Behavior Factor V: Inappropriate Behavior
302
p
p 0.08 0.48
t [df] 1.71 [2,012] ⳮ0.70 [1,702]
Comparison of clinical variables between patients who did and did not have a psychiatric consultation No Consultation
Dementia Yes No
v2[df]
No ConsultationN (%) Gender Male Female Marital status Married Not married Race Non-white White
TABLE 5.
is loud and verbally abusive may get more attention than the resident who is quietly depressed (the “Retarded Depression” factor on the CSDD did not predict consultation). Residents with dementia were more likely to receive a psychiatric consultation (24%) than those without (14%). This finding also corresponds with those of other studies.20,24 Patients residing in urban and for-profit nursing homes had a higher rate of psychiatric consultation. Burns et al.4 found that residents living in rural nursing homes were more likely than those residing in an urban home to see a general practitioner rather than a psychiatrist when psychiatric problems developed. They also found that residents living in nonproprietary homes were more likely to see a general practitioner than a psychiatrist. Reichman et al.5 reported that the frequency of psychiatric consultation increased from suburban to urban facility locations and from small to medium to large facility sizes. Nursing
Consultation
N
Mean (SD)
N
Mean (SD)
1,345 854 854 854 854 1,282 1,240 1,240 1,240 1,240 1,240
4.25 (4.79) 0.001 (1.00) 0.12 (0.873) 0.02 (0.90) 0.022 (0.92) 1.79 (3.21) ⳮ0.06 (0.88) 0.07 (0.88) 0.01 (0.962) 0.02 (0.90) 0.05 (0.58)
353 202 202 202 202 355 343 343 343 343 343
6.34 (6.04) 0.006 (1.05) 0.48 (1.30) 0.036 (1.30) 0.138 (1.23) 3.55 (4.90) 0.21 (1.33) 0.19 (1.22) 0.09 (1.23) 0.06 (1.23) 0.20 (1.79)
N (%) 715 (76) 896 (86)
N (%) 257 (24) 147 (14)
t
p
[df]
6.05** 0.09 6.28** 0.61 1.73 6.37** 3.61** 3.62** 1.55 1.15 2.54*
0.00 0.92 0.00 0.54 0.08 0.00 0.00 0.00 0.123 0.251 0.01
[474.2] [1,054] [245.3] [250.6] [256.1] [441.4] [427.9] [444.0] [482.6] [448.5] [362.2]
v2 47.84**
p 0.00
[df] [1]
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. homes located in urban areas may allow for more convenient access by mental health staff than those in rural areas, where there is often a shortage of mental health professionals. This study has a number of limitations. As with many studies in a nursing home setting, this study had varying quality and quantity of information gathered. Data were collected from four sources. These sources were direct patient interviews, information from medical records (including the Minimum Data Set [MDS]), family informants, and nursing staff. The information varied from patient to patient and from section to section for the data collected. The intent was to get as complete a data set as possible. Several nurse-informants assessed the ratings for agitation and depression. Because of the large number of informants, we were unable to assess interrater reliability. The diagnosis of dementia was made by an expert panel, and not by direct evaluation. This method has been shown to have good reliability as compared with direct clinical assessment.10 Also, most residents were unable to provide their own information. In our study, only 63% were able to respond to the MMSE, and only 31% of those scored 24 or more. Medical comorbidity may affect the Cornell Scale and increase the Retardation factor. That is, some residents who have vegetative symptoms like sleep or appetite disturbance related to medical illness may be misclassified as depressed. In an attempt to minimize this misclassification, evaluations for residents with fever or acute medical illness were delayed until their condition was resolved. Another limitation was that we only looked at consultations that were performed within the first 90 days after admission. This may affect our findings, given that it does not take into account residents who may have been referred for a psychiatric consultation at a later date. Ninety days was chosen because we had clinical data available at baseline but not at subsequent time periods. Also, most residents were still in the nursing home at 90 days; over time the sample diminished because of death, hospitalization, or discharge. Also, the present study was carried out in a single state. Although the residents appear to be similar to those in national surveys, caution is required when attempting to generalize results beyond the state of Maryland.
Am J Geriatr Psychiatry 12:3, May-June 2004
An additional matter of importance is that the data are derived from a sample recruited in 1992–1995.The nursing home MDS, which had a care-planning component (including a trigger for depression), was implemented before the start of this study. It is unclear what impact this might have had on rates of psychiatric visits or depression detection. Since 1995, the MDS has not only become required for care planning and survey of resident demographics, but it is also being used for prospective payment and being assessed for quality indicators. Certainly the prevalence and treatment for depression in nursing homes may have changed since our assessment. Finally, only 70% of eligible subjects participated in this study. In order to improve response rates in research involving nursing home populations, it is important to build a mutually beneficial relationship between the nursing home and the research staff. By doing so, the researchers benefit from having the opportunity to collect important information while the nursing home staff benefit from periodic feedback on results. The response rate can increase when there is a solid relationship with the nursing home and there is mutual respect and exchange of information. An orderly and systematic approach to identifying residents and a direct, personal approach with staff and residents, with the support of the nursing home administration, maximized recruitment in this study.25 This study found that certain groups of patients, such as those with retarded depression, may be overlooked and may not receive psychiatric intervention. It is essential that nursing home staff be better educated about the broad range of mental health needs of the residents. Continued research and better education and support of staff may lead to better recognition of psychiatric problems and, thus, improved care. We appreciate the cooperation of the facilities, residents, and families participating in the University of Maryland Long-Term Care Project. This research was supported by a grant from the National Institute on Aging (RO1 AGO8211), by the Veterans Affairs Maryland Healthcare System, and by the Veterans Affairs Capitol Network Mental Illness Research Education and Clinical Center (MIRECC).
303
Psychiatric Consultation in Nursing Homes
References 1. Kim E, Rovner B: Epidemiology of psychiatric disturbances in nursing homes. Psychiatr Ann 1995; 25:409–412 2. Rovner B, Kafonek S, Filipp L: Prevalence of mental illness in the community nursing home. Am J Psychiatry 1986; 143: 1446–1449 3. Zimmer JG, Watson N, Treat A: Behavioral problems among patients in skilled nursing facilities. Am J Public Health 1984; 74:1118–1121 4. Burns W, Wagner HR, Taube JE, et al: Mental health service use by elderly in nursing homes. Am J Public Health 1993; 83:331– 337 5. Reichman WE, Coyne AC, Borson S, et al: Psychiatric consultation in the nursing home: a survey of six states. Am J Geriatr Psychiatry 1996; 6:320–327 6. Magaziner J, German P, Zimmerman SI, et al. The prevalence of dementia in a statewide sample of new nursing home admissions age 65 and older: diagnosis by expert panel. Gerontologist 2000; 40:663–672 7. Folstein MF, Folstein SE, McHugh PR: Mini-Mental State: a practical method for grading cognitive state of patients for the clinician. J Psychiatric Research 1975; 12:189–198 8. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, Revised (DSM III-R). Washington, DC, American Psychiatric Association, 1987 9. Magaziner J, Zimmerman S, German P, et al: Ascertaining dementia by expert panel in epidemiological studies of nursing home residents. Ann Epidemiol 1996; 6:431–437 10. Alexopoulos GS, Abrams RC, Young RC, et al: Cornell Scale for Depression in Dementia. Biol Psychiatry 1988; 23:271–284 11. Alexopoulos GS, Abrams RC, Young RC, et al: Use of the Cornell Scale in nondemented patients. J Am Geriatric Soc 1988; 36:230– 236 12. Wilkinson IM, Graham-White J: Psychogeriatric Dependency Scale (PGDRS): a method of assessment for use by nurses. Br J. Psychiatry 1980; 137:558–565
304
13. Bogner JA, Corrigan J, Strange M, et al: Reliability of the Agitated Behavior Scale. Head Trauma Rehabil 1999; 14:91–96 14. Cohen-Mansfield J: Assessment of agitation. Int Psychogeriatr 1996; 8:233–245 15. Tariot DM, Mack JL, Patterson MB, et al: The Behavior Rating Scale of the Consortium to Establish a Registry for Alzheimer’s Disease. Am J Psychiatry 1995; 152:1349–1357 16. Harwood DG, Ownby RL, Barker WW, et al: The factor structure of the Cornell Scale for Depression in Dementia among probable Alzheimer’s disease patients. Am J Geriatr Psychiatry 1998; 6:212–220 17. Huber PJ: The behavior of maximum-likelihood estimates under non-standard conditions, in Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, CA, University of California Press, 1967; 1:221–233 18. Shea D, Streit A: The provision and use of mental health services in nursing homes. Am J Public Health 1994; 82:284–287 19. Colenda C, Streim J, Greene JA, et al: The impact of OBRA ’87 on psychiatric services in nursing homes: joint testimony of the American Psychiatric Association and the American Association for Geriatric Psychiatry. Am J Geriatr Psychiatry 1999; 7:12–17 20. Shea DG, Streit A, Smyer MA: Determinants of the use of specialist mental health services by nursing home residents. Health Serv Res 1994; 29:169–185 21. Loebel JP, Borson S, Hyde T, et al: Relationships between requests for psychiatric consultations and psychiatric diagnoses in longterm care facilities. Am J Psychiatry 1991; 148:898–903 22. Rabins PV: Mental illness among nursing home patients. Hosp Community Psychiatry 1985; 36:119–128 23. Spore DL, Shyer MA, Cohn MD: Assessing nursing assistants’ knowledge of behavioral approaches to mental health problems. Gerontologist 1991; 31:309–317 24. Goldman LS, Klugman A: Psychiatric consultation in a teaching nursing home. Psychosomatics 1990; 31:277–281 25. Palumbo FB, Magaziner JS, Hyg MS, et al: Recruitment of long-term care facilities for research. J Am Geriatr Soc 1987; 35:154–158
Am J Geriatr Psychiatry 12:3, May-June 2004
Objective: Despite the high rate of psychiatric disorders in nursing homes, research indicates that psychiatric consultation is requested infrequently. The authors sought to determine the rate of psychiatric consultation in a nursing home population and to assess what factors were related to a consultation request. Methods: Subjects were recruited from a stratified random sample of 59 nursing homes across Maryland. All new admissions age 65 years and older from September 1992 through March 1995 were eligible for the study. A total of 2,285 subjects were included in the study. Variables examined were factor scores from the Cornell Scale for Depression in Dementia and the Psychogeriatric Dependency Rating Scale (Behavioral Subscale), nursing home characteristics, and whether the resident had a psychiatric consultation within 90 days of admission. Results: Twenty percent of the residents (N⳱404) had a psychiatric consultation. There was no significant association with demographic variables. Behaviors that triggered a psychiatric consultation included agitation, physical/ verbal abuse, wandering, and manic/destructive acts. A psychiatric consultation was also requested when residents displayed anxiety. Surprisingly, depression in retarded and psychotic residents did not trigger a psychiatric consult. Conclusion: As expected, behavioral problems and agitation are common reasons for a psychiatric consultation. However, the resident who is depressed, particularly the quiet or retarded depressed resident, may be overlooked. In this context, it is important for the nursing staff to recognize that lethargy and social withdrawal may be signs of depression, and a referral to a psychiatrist may be in order. (Am J Geriatr Psychiatry 2004; 12:297–304)
Received March 7, 2003; revised October 22, December 3, 2003; accepted January 6, 2004. From the Mental Health Service Line, Veterans Affairs Maryland Health Care Center (JF,ASM,BK,DL,PER); the Veterans Affairs Capitol Network Mental Illness Research Education and Clinical Center (PER); the Department of Epidemiology and Preventive Medicine, University of Maryland School of Medicine (ALGB,JM); the Department of Psychiatry, University of Maryland School of Medicine (AR); and the Cecil G. Sheps Center for Health Services Research, The University of North Carolina at Chapel Hill (SZ). Send correspondence to Paul Ruskin, M.D., Mental Health Service Line (116/MH), Baltimore VAMC, 10 N. Greene St., Baltimore, MD 21201. e-mail: [email protected] 䉷 2004 American Association for Geriatric Psychiatry
Am J Geriatr Psychiatry 12:3, May-June 2004
297
Psychiatric Consultation in Nursing Homes
S
killed residential facilities such as nursing homes have, to a great extent, replaced state psychiatric facilities as the primary source of residential care for elderly patients.1 Several studies have found the prevalence of psychiatric disorders in these facilities to exceed 70%.1,2 In spite of these high rates, fewer than 1% of nursing home residents receive explicit mental health interventions, and consultation is rarely requested, even for mental health problems identified as “serious.”3 Residents with serious problems were defined as those who had more than impaired judgment or physical-restraint orders. In one study, in which two-thirds of elderly residents had a mental disorder (including dementia), only 4.5% received any mental health treatment within a 1-month period.4 A recent study of skilled nursing home facilities in six states found that the perceived need for psychiatric service is far greater than the level of services provided.5 There are few reports describing rates and correlates of psychiatric consultation in the nursing home setting. Our objectives were to document the number of psychiatric consultations requested in a large and diverse sample of nursing homes and to examine the demographic, clinical, and nursing home characteristics associated with such consultations.
METHODS Subjects Subjects were recruited from a stratified random sample of 59 nursing homes across Maryland. Nursing homes were classified by number of beds and region. At the time of the study (1992), there were 221 licensed long-term care facilities in Maryland. They were broken into five regions (Eastern Shore, Western MD, Baltimore City, Baltimore Metropolitan, Washington Metropolitan), and three size categories (⬍50 beds, 50–150 beds, and ⬎150 beds); from these, a stratified random sample of facilities was selected. All new admissions age 65 years and older from September 1992 through March 1995, who had not been a resident in a nursing home in the previous year were eligible for the study. Informed consent was obtained from either the resident or a significant other. The Institutional Review Board of the University of Maryland approved the consent procedure.
298
There were a total of 3,283 eligible subjects, of whom 2,285 were enrolled in the study. Of the 2,285 who were included in the overall study, there were 132 residents for whom there was no follow-up data collection: 80 had a short stay (less than 8 days) and were excluded from follow-up; 40 had missing charts; and 12 denied permission to collect follow-up information. For an additional 138 residents, 25% or more of the chart information was missing; these residents are excluded from the calculation of event rates (N⳱2,015), but are included in the analyses of length of stay and final disposition. There were 998 who did not participate. The major reason for not participating was that the residents or their significant others refused consent. We had chart follow-up about physician contact on 2,015 subjects. Comparison of the age and sex distributions for enrolled and non-enrolled eligible cases indicated that those enrolled were slightly older (81.5 years versus 80.6 years; p ⬍0.001) and more often female (71.6% versus 68.6%; p ⬍0.05). Evaluations for residents with fever or acute medical illness were delayed until their condition was resolved. A more detailed description of study methods has been published elsewhere.6 Measures Data were collected from multiple sources, including interviews with residents, nursing staff, and significant others; medical records, and hospital discharge summaries. Trained lay interviewers conducted all interviews and medical chart reviews. The staff member who was most knowledgeable about a particular resident completed the evaluation for that resident. Given the large number of informants, interrater reliability was not assessed. Cognitive deficit. Cognitive deficits were assessed with the Mini-Mental State Exam (MMSE).7 The MMSE was administered to every resident who did not refuse, could communicate, and for whom testing could be done within 2 months of admission (N⳱1,446; 63% of the sample). Residents were classified by severity of cognitive impairment on the basis of MMSE scores: Severe (⬍10); Moderate (10–19); Mild (20–26), and Absent (27–30).7 Dementia. The determination of dementia was made in accordance with DSM-III-R criteria8 by an expert panel consisting of geriatric psychiatrists, neurologists, and a geriatrician, using detailed information collected by lay interviewers. Information on his-
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. tory of cognitive and functional decline, current cognitive and functional status, and information about affective, social, and behavioral status were considered in making a dementia diagnosis. The panel used information concerning cognitive status obtained from the MMSE, as well as the Blessed Dementia Scale, which was administered to a family member or friend. A detailed description of the dementia ascertainment process may be found elsewhere.9 Depression. Depression symptoms were ascertained with a modified version of the Cornell Scale for Depression in Dementia (CSDD) completed by a nurse-informant. The CSDD was modified in that we asked questions to the nursing home staff instead of having a rater directly assess the resident. The CSDD is an observer-rated scale designed to rate depression in residents with dementia. It contains 19 items, and total scores range from 0 to 38. On the basis of their scale scores, residents were categorized with No Depression (score 1–7); Mild-to-Moderate Depression (score 8–20); or Severe Depression (score 21Ⳮ). These categorizations follow the recommendations of Alexopoulos and associates,10 who developed this scale. This depression scale has been validated in both dementia and non-dementia conditions.11 Agitation. Behavioral functioning was assessed by use of the behavior subscale of the Psychogeriatric Dependency Rating Scale (PGDRS), which assessed the presence of aggression, passive hostility, attention-seeking, wandering, restlessness, and fearfulness.12 The instrument was developed to evaluate care needs of older adults. The PGDRS includes 17 items rated on a frequency-of-occurrence scale as 0: none; 1: 1–2 times a week; and 2: more often. Nursing home characteristics. The nursing home in which the resident resided was characterized in terms of locality (urban or rural), size (number of beds), chain affiliation (affiliated or not affiliated), Medicarecertified (certified or not), and for-profit status (forprofit or not-for-profit). Consultation. The presence or absence of a psychiatric consultation within the first 90 days was determined through chart review. Consultations were performed by psychiatrists. Data Analysis Previous factor-analytic studies of scales assessing agitation in elderly residents of nursing homes have
Am J Geriatr Psychiatry 12:3, May-June 2004
identified at least three independent factors.13–15 Based on these results, a factor analysis of the PGDRS Behavior subscale was undertaken. We performed a principal-components analysis, followed by a normal Varimax rotation, on all factors with eigenvalues of 1 or greater. Five factors met this criterion and were extracted. An exact factor scoring method was used that includes all items in a factor, with the items assigned weights on the basis of their loading on the factor. This approach maintains the orthogonality among factors. The Physical Aggression/ Verbal Abuse factor included items with highest loadings on behaviors interfering with and disturbing other residents and staff—hitting, biting, scratching, and deliberately refusing to obey commands. The Wandering factor referred to wandering in the nursing home and pacing. Items with high loadings on the third factor (Intrusiveness) were “demanding conversation” and “getting others to do things” that the resident was capable of doing. Psychotic Behavior Items included incoherent speech and hallucinations; and the last factor, Manic/Destructive or Inappropriate Behavior, included self-destructive acts, inappropriate elation, and destroying and damaging clothes and property (Table 1). A similar factor analysis was performed on the CSDD,16 and four independent factors were extracted (Table 2). The factors were named according to the items in the scale that had the highest loadings. Retarded/Depressed items were appetite loss, weight loss, and anergy. The Anxious/Agitated factor referred to irritability and agitation. Psychotic Depression included items about suicidal ideation and mood-congruent delusions; and the last factor, Sleep Disturbance, included items on early and middle insomnia. The residents who did and did not receive a psychiatric consultation were compared on demographic variables, nursing home characteristics, and clinical variables by use of chi-square and t-tests for these comparisons. Finally, a total of 14 variables, consisting of factor scores and demographics, were included in a logisticregression analysis to distinguish which residents did or did not receive a psychiatric consultation (Table 3). Clustering was not taken into account in this analysis because previous analyses with the same sample using the method suggested by Huber17 had little effect on the statistical results.
299
Psychiatric Consultation in Nursing Homes RESULTS The mean age of the sample population was 81.4 years. Female residents accounted for 70.9% of the sample, and 83% were white. The majority of the population were widowed (62.2%), and 23.9% were married. Fifty-one percent of the sample had at least a high school education. Eighty percent of the subjects did not have a psychiatric consultation; 9.4% had one consultation; and 10.6% had more than one consultation within the first 90 days. TABLE 1.
The majority of those who had a psychiatric visit stayed beyond 90 days. Of those, 64% who were only seen once in the first 90 days were seen again by 180 days. Comparing those residents who had a psychiatric consultation with those who did not, there was no significant difference in gender, marital status, race, age, or education (Table 4). Seven of the 12 clinical variables were significantly associated with referral for a psychiatric consultation (Table 5). These included the Total score on the CSDD (t[474.2]⳱6.05; p ⬍0.00) and the Cornell Anxious/Agitated subscale (t[245.3]⳱6.28; p ⬍0.00), Total score on the PGDRS
Factors of psychopathology extracted from the Psychogeriatric Dependency Rating Scale (PGDRS)
PGDRS Interferes with activity Uninvolved Wanders Objectionable behavior Demands conversation Slow to talk Makes noises Hits Refuses to obey Verbally abusive Repeated movements Self-destruct Destroys property Too happy Paranoid delusions Hallucinations Incoherent speech
1
2
3
4
5
0.462 0.103 0.091 0.278 0.114 0.077 0.440 0.735 0.644 0.794 0.083 0.248 0.116 0.099 0.590 0.060 0.085
0.423 0.059 0.813 0.483 0.156 0.356 0.084 0.216 0.294 0.043 0.804 0.257 0.049 0.082 0.138 0.057 0.305
0.266 0.751 0.074 0.010 0.745 0.371 0.361 0.096 0.191 0.133 0.097 0.097 0.046 0.168 0.132 0.083 0.008
0.167 0.026 0.020 0.356 0.082 0.326 0.306 0.097 0.042 0.015 0.048 0.256 0.225 0.127 0.351 0.640 0.669
0.201 0.044 0.100 0.120 0.116 0.065 0.156 0.125 0.102 0.094 0.072 0.406 0.689 0.765 0.033 0.205 0.060
Factor names: 1: Physical Aggression/Verbal Abuse; 2: Wandering; 3: Intrusiveness; 4: Psychotic Behavior; 5: Manic Destructiveness.
TABLE 2.
Factors of psychopathology extracted from the Cornell Scale for Depression in Dementia (CSDD)
CSDD Anxious Sadness Lack reaction Irritability Agitation Retardation Physical complaints Loss of interest Appetite loss Weight loss Lack of energy Diurnal mood change Hard to sleep Wake often Wake earlier Suicidal ideas Low self-esteem Pessimism Delusions
1
2
3
4
0.207 0.275 0.535 0.082 0.117 0.489 0.170 0.655 0.781 0.744 0.683 0.369 0.067 0.099 0.144 0.098 0.090 0.128 0.014
0.629 0.587 0.314 0.77 0.791 0.218 0.270 0.251 0.110 0.062 0.124 0.473 0.165 0.125 0.165 0.136 0.030 0.111 0.177
0.246 0.288 0.040 0.072 0.100 0.040 0.394 0.217 0.119 0.013 0.103 0.184 0.140 0.163 0.112 0.719 0.688 0.711 0.633
0.146 0.056 0.014 0.151 0.184 0.042 0.092 0.114 0.062 0.051 0.171 0.085 0.807 0.855 0.788 0.076 0.074 0.196 0.052
Factor names: 1: Retarded/Depressed; 2: Anxious/Agitated; 3: Psychotic Depression; 4: Sleep Disturbance.
300
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. (t[441.4]⳱6.37; p ⬍0.00), PGDRS factors Physical Aggression/Verbal Abuse (t[427.9]⳱3.61; p ⬍0.00), Wandering (t[444.6]⳱3.62; p ⬍0.00), and Inappropriate Behavior (t[362.2]⳱2.54; p ⬍0.01). The presence of dementia was also significant (v2[1]⳱47.84; p ⬍0.00), with more of these residents having a consultation. With regard to nursing home characteristics, residents who were in an urban rather than a rural location had a significantly higher rate of psychiatric consultations (22% versus 12%; v2[1]⳱18.79; p ⬍0.01). Those residing in a for-profit nursing home had a higher rate of consultation than those residing in a non-profit home (21% versus 18%; v2[1]⳱4.06; p ⬍0.05). The rate of psychiatric consultation was not influenced by whether the nursing home had a chain affiliation or was Medicare-certified. Also, the number of comprehensive-care beds and number of beds in the facility did not relate to the number of consultations. On the basis of these previous findings, 14 variables were included in logistic-regression analysis as possible correlates of residents’ having a psychiatric consultation (Table 3). These included the five factor scores of the PGDRS, the four factor scores of the CSDD, and a number of demographic variables and the presence or absence of dementia. Two variables, PGDRS Manic/Destructive and the Cornell factor Anxious/Agitated, significantly discriminated patients who did and did not receive a psychiatric consultation, whereas the PGDRS Physical Aggression/ Verbal Abuse and the presence or absence of TABLE 3.
Logistic regression: demographic and clinical predictors of psychiatric consultation
Variable Sex Marital status Race Education Physical Aggression/Verbal Abuse Wandering Intrusiveness Psychotic Behavior Manic/Destruction Retarded/Depressed Anxious/Agitated Psychotic Depression Sleep Disturbance Dementia Y/N
b
Wald v2
p
0.047 0.047 0.026 0.005 0.210 0.171 0.053 0.195 0.394 0.035 0.512 0.053 0.066 0.464
0.048 0.038 0.011 0.036 3.73 3.22 0.295 3.28 8.22 0.140 26.97 0.337 0.432 5.51
0.83 0.85 0.91 0.85 0.05* 0.07 0.59 0.07 0.00** 0.71 0.00** 0.56 0.51 0.02*
Overall model df: 14, individual comparisons df: 1. *Significant at the 0.05 level (two-tailed). **Significant at the 0.01 level (two-tailed).
Am J Geriatr Psychiatry 12:3, May-June 2004
dementia showed a trend toward significance (p ⬍0.05).
DISCUSSION This study is important in alerting caregivers to the clinical and delivery-service characteristics that are associated with a psychiatric consultation. To our knowledge, this article is the first to document the percentage of nursing home residents to have a psychiatric consultation within the first 90 days after admission. Our study looks at a large number of residents in a broad range of nursing-home settings in the state of Maryland. Twenty percent is a high percentage in comparison to other studies, which estimate that only 5% of nursing home residents received mental health treatment within a 1-month period,18 and another report of 2.3% of residents who had a mental health consultation in the previous month.4 In our study, there were a greater number of consultations within 3 months after admission, in comparison with a discrete 1-month period. Both previous studies reflect rates of consultation before the Omnibus Reconciliation Act (OBRA ’87), whereas our study may reflect the greater awareness of mental health needs after its implementation.19 There was no significant difference in the age, gender, or education of the residents who had a psychiatric consultation and those who did not. These results differ from those of two other studies, which found that older residents were less likely to receive mental health treatment.4,20 However, there was a restricted age range in this study, with admissions limited to those 65 years of age or older. Reasons for psychiatric referral of nursing home residents are diverse; these include behavior problems, mood-related problems, psychotic features, somatic complaints, and impaired activities of daily living.21 Staff are often concerned with the psychosocial aspects of patients’ problems. However, they have difficulty recognizing the legitimacy of their psychological complaints.23 Common reasons for referral in a continuing-care retirement community included depression and dementia. Residents in this setting often suffer emotional distress from the loss of independence.22 Front-line staff often have poor training and low pay; staff turnover is high, and those caring
301
Psychiatric Consultation in Nursing Homes for older residents may not be sensitive to the need for psychiatric consultation.23 Often, their view of the psychiatrist is primarily as a prescriber of medication, rather than an overall manager of patient care. An additional deterrent to psychiatric consultation in the nursing home is the low level of reimbursement to the psychiatrist. Clinically, a number of symptom syndromes were associated with a psychiatric consultation in this study. Both the PGDRS Behavioral subscale and the CSDD Total score were significant. When these scores were factor-analyzed, the factors were able to give a clearer picture of the specific symptoms, which prompted a psychiatric consultation. Behavioral disturbances, including Physical Aggression, Wandering, and Inappropriate Behavior were significantly correlated with getting a consultation. Similar findings have been documented by Goldman and Klugman.24 These findings suggest that the resident who TABLE 4.
Comparison of demographic variables between patients who did and did not have a psychiatric consultation
Age, years Education, years
ConsultationN (%)
476 (80) 1,135 (80)
120 (20) 284 (20)
0.00[1]
0.95
1,236 (81) 375 (77)
294 (19) 110 (23)
2.76[1]
0.10
304 (81) 1,305 (80)
72 (19) 332 (20)
0.24[1]
0.62
N 1,610 1,357
Mean (SD) 81.52 (7.67) 10.62 (3.94)
N 404 347
Mean (SD) 80.80 (7.36) 10.78 (3.86)
CSDD Total score Cornell Factor I: Retarded/Depression Cornell Factor II: Anxious/Agitated Cornell Factor III: Psychotic Depression Cornell Factor IV: Sleep Disturbance PGDRS Behavior Total score PGDRS Behavior Factor I: Physical Aggression/Verbal Abuse PGDRS Behavior Factor II: Wandering PGDRS Behavior Factor III: Intrusiveness PGDRS Behavior Factor IV: Psychotic Behavior PGDRS Behavior Factor V: Inappropriate Behavior
302
p
p 0.08 0.48
t [df] 1.71 [2,012] ⳮ0.70 [1,702]
Comparison of clinical variables between patients who did and did not have a psychiatric consultation No Consultation
Dementia Yes No
v2[df]
No ConsultationN (%) Gender Male Female Marital status Married Not married Race Non-white White
TABLE 5.
is loud and verbally abusive may get more attention than the resident who is quietly depressed (the “Retarded Depression” factor on the CSDD did not predict consultation). Residents with dementia were more likely to receive a psychiatric consultation (24%) than those without (14%). This finding also corresponds with those of other studies.20,24 Patients residing in urban and for-profit nursing homes had a higher rate of psychiatric consultation. Burns et al.4 found that residents living in rural nursing homes were more likely than those residing in an urban home to see a general practitioner rather than a psychiatrist when psychiatric problems developed. They also found that residents living in nonproprietary homes were more likely to see a general practitioner than a psychiatrist. Reichman et al.5 reported that the frequency of psychiatric consultation increased from suburban to urban facility locations and from small to medium to large facility sizes. Nursing
Consultation
N
Mean (SD)
N
Mean (SD)
1,345 854 854 854 854 1,282 1,240 1,240 1,240 1,240 1,240
4.25 (4.79) 0.001 (1.00) 0.12 (0.873) 0.02 (0.90) 0.022 (0.92) 1.79 (3.21) ⳮ0.06 (0.88) 0.07 (0.88) 0.01 (0.962) 0.02 (0.90) 0.05 (0.58)
353 202 202 202 202 355 343 343 343 343 343
6.34 (6.04) 0.006 (1.05) 0.48 (1.30) 0.036 (1.30) 0.138 (1.23) 3.55 (4.90) 0.21 (1.33) 0.19 (1.22) 0.09 (1.23) 0.06 (1.23) 0.20 (1.79)
N (%) 715 (76) 896 (86)
N (%) 257 (24) 147 (14)
t
p
[df]
6.05** 0.09 6.28** 0.61 1.73 6.37** 3.61** 3.62** 1.55 1.15 2.54*
0.00 0.92 0.00 0.54 0.08 0.00 0.00 0.00 0.123 0.251 0.01
[474.2] [1,054] [245.3] [250.6] [256.1] [441.4] [427.9] [444.0] [482.6] [448.5] [362.2]
v2 47.84**
p 0.00
[df] [1]
Am J Geriatr Psychiatry 12:3, May-June 2004
Fenton et al. homes located in urban areas may allow for more convenient access by mental health staff than those in rural areas, where there is often a shortage of mental health professionals. This study has a number of limitations. As with many studies in a nursing home setting, this study had varying quality and quantity of information gathered. Data were collected from four sources. These sources were direct patient interviews, information from medical records (including the Minimum Data Set [MDS]), family informants, and nursing staff. The information varied from patient to patient and from section to section for the data collected. The intent was to get as complete a data set as possible. Several nurse-informants assessed the ratings for agitation and depression. Because of the large number of informants, we were unable to assess interrater reliability. The diagnosis of dementia was made by an expert panel, and not by direct evaluation. This method has been shown to have good reliability as compared with direct clinical assessment.10 Also, most residents were unable to provide their own information. In our study, only 63% were able to respond to the MMSE, and only 31% of those scored 24 or more. Medical comorbidity may affect the Cornell Scale and increase the Retardation factor. That is, some residents who have vegetative symptoms like sleep or appetite disturbance related to medical illness may be misclassified as depressed. In an attempt to minimize this misclassification, evaluations for residents with fever or acute medical illness were delayed until their condition was resolved. Another limitation was that we only looked at consultations that were performed within the first 90 days after admission. This may affect our findings, given that it does not take into account residents who may have been referred for a psychiatric consultation at a later date. Ninety days was chosen because we had clinical data available at baseline but not at subsequent time periods. Also, most residents were still in the nursing home at 90 days; over time the sample diminished because of death, hospitalization, or discharge. Also, the present study was carried out in a single state. Although the residents appear to be similar to those in national surveys, caution is required when attempting to generalize results beyond the state of Maryland.
Am J Geriatr Psychiatry 12:3, May-June 2004
An additional matter of importance is that the data are derived from a sample recruited in 1992–1995.The nursing home MDS, which had a care-planning component (including a trigger for depression), was implemented before the start of this study. It is unclear what impact this might have had on rates of psychiatric visits or depression detection. Since 1995, the MDS has not only become required for care planning and survey of resident demographics, but it is also being used for prospective payment and being assessed for quality indicators. Certainly the prevalence and treatment for depression in nursing homes may have changed since our assessment. Finally, only 70% of eligible subjects participated in this study. In order to improve response rates in research involving nursing home populations, it is important to build a mutually beneficial relationship between the nursing home and the research staff. By doing so, the researchers benefit from having the opportunity to collect important information while the nursing home staff benefit from periodic feedback on results. The response rate can increase when there is a solid relationship with the nursing home and there is mutual respect and exchange of information. An orderly and systematic approach to identifying residents and a direct, personal approach with staff and residents, with the support of the nursing home administration, maximized recruitment in this study.25 This study found that certain groups of patients, such as those with retarded depression, may be overlooked and may not receive psychiatric intervention. It is essential that nursing home staff be better educated about the broad range of mental health needs of the residents. Continued research and better education and support of staff may lead to better recognition of psychiatric problems and, thus, improved care. We appreciate the cooperation of the facilities, residents, and families participating in the University of Maryland Long-Term Care Project. This research was supported by a grant from the National Institute on Aging (RO1 AGO8211), by the Veterans Affairs Maryland Healthcare System, and by the Veterans Affairs Capitol Network Mental Illness Research Education and Clinical Center (MIRECC).
303
Psychiatric Consultation in Nursing Homes
References 1. Kim E, Rovner B: Epidemiology of psychiatric disturbances in nursing homes. Psychiatr Ann 1995; 25:409–412 2. Rovner B, Kafonek S, Filipp L: Prevalence of mental illness in the community nursing home. Am J Psychiatry 1986; 143: 1446–1449 3. Zimmer JG, Watson N, Treat A: Behavioral problems among patients in skilled nursing facilities. Am J Public Health 1984; 74:1118–1121 4. Burns W, Wagner HR, Taube JE, et al: Mental health service use by elderly in nursing homes. Am J Public Health 1993; 83:331– 337 5. Reichman WE, Coyne AC, Borson S, et al: Psychiatric consultation in the nursing home: a survey of six states. Am J Geriatr Psychiatry 1996; 6:320–327 6. Magaziner J, German P, Zimmerman SI, et al. The prevalence of dementia in a statewide sample of new nursing home admissions age 65 and older: diagnosis by expert panel. Gerontologist 2000; 40:663–672 7. Folstein MF, Folstein SE, McHugh PR: Mini-Mental State: a practical method for grading cognitive state of patients for the clinician. J Psychiatric Research 1975; 12:189–198 8. American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, Revised (DSM III-R). Washington, DC, American Psychiatric Association, 1987 9. Magaziner J, Zimmerman S, German P, et al: Ascertaining dementia by expert panel in epidemiological studies of nursing home residents. Ann Epidemiol 1996; 6:431–437 10. Alexopoulos GS, Abrams RC, Young RC, et al: Cornell Scale for Depression in Dementia. Biol Psychiatry 1988; 23:271–284 11. Alexopoulos GS, Abrams RC, Young RC, et al: Use of the Cornell Scale in nondemented patients. J Am Geriatric Soc 1988; 36:230– 236 12. Wilkinson IM, Graham-White J: Psychogeriatric Dependency Scale (PGDRS): a method of assessment for use by nurses. Br J. Psychiatry 1980; 137:558–565
304
13. Bogner JA, Corrigan J, Strange M, et al: Reliability of the Agitated Behavior Scale. Head Trauma Rehabil 1999; 14:91–96 14. Cohen-Mansfield J: Assessment of agitation. Int Psychogeriatr 1996; 8:233–245 15. Tariot DM, Mack JL, Patterson MB, et al: The Behavior Rating Scale of the Consortium to Establish a Registry for Alzheimer’s Disease. Am J Psychiatry 1995; 152:1349–1357 16. Harwood DG, Ownby RL, Barker WW, et al: The factor structure of the Cornell Scale for Depression in Dementia among probable Alzheimer’s disease patients. Am J Geriatr Psychiatry 1998; 6:212–220 17. Huber PJ: The behavior of maximum-likelihood estimates under non-standard conditions, in Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, CA, University of California Press, 1967; 1:221–233 18. Shea D, Streit A: The provision and use of mental health services in nursing homes. Am J Public Health 1994; 82:284–287 19. Colenda C, Streim J, Greene JA, et al: The impact of OBRA ’87 on psychiatric services in nursing homes: joint testimony of the American Psychiatric Association and the American Association for Geriatric Psychiatry. Am J Geriatr Psychiatry 1999; 7:12–17 20. Shea DG, Streit A, Smyer MA: Determinants of the use of specialist mental health services by nursing home residents. Health Serv Res 1994; 29:169–185 21. Loebel JP, Borson S, Hyde T, et al: Relationships between requests for psychiatric consultations and psychiatric diagnoses in longterm care facilities. Am J Psychiatry 1991; 148:898–903 22. Rabins PV: Mental illness among nursing home patients. Hosp Community Psychiatry 1985; 36:119–128 23. Spore DL, Shyer MA, Cohn MD: Assessing nursing assistants’ knowledge of behavioral approaches to mental health problems. Gerontologist 1991; 31:309–317 24. Goldman LS, Klugman A: Psychiatric consultation in a teaching nursing home. Psychosomatics 1990; 31:277–281 25. Palumbo FB, Magaziner JS, Hyg MS, et al: Recruitment of long-term care facilities for research. J Am Geriatr Soc 1987; 35:154–158
Am J Geriatr Psychiatry 12:3, May-June 2004