Pharmacologic treatment of major depression for elderly patients in residential care settings.

A significant drug-placebo difference was found in a double-blind, placebo-controlled study of nortriptyline for treatment of major depression among frail elderly patients living in an institutional setting. This finding confirms the validity of the DSM-III-R diagnosis of major depression and establishes the need for specific psychiatric services for the chronically ill elderly living in nursing homes and congregate housing facilities. The incidence of adverse events requiring early termination of treatment was 34%, demonstrating the vulnerability of these patients and their need for careful monitoring during treatment. High levels of self-care disability and low levels of serum albumin were both associated with decreased therapeutic responses, demonstrating the need for further research on psychopathology in these settings.

Pain complaints and cognitive status among elderly institution residents.

OBJECTIVE: To examine the association between self-reported pain and cognitive impairment among frail elderly institution residents. DESIGN: A cross-sectional correlational study. SETTING: A large urban nursing home and congregate apartment complex housing predominantly Jewish elderly. PARTICIPANTS: Seven hundred fifty-eight elderly institution residents (30% in the nursing home, 70% in congregate apartments). The sample was 70% female and averaged 83.3 years of age. MEASUREMENTS: Respondent self-reports tapped pain intensity, number of localized pain complaints, cognitive status, and disability in performance of activities of daily living. Attending physicians or physician assistants rated respondents' health status. MAIN RESULTS: Pain intensity and number of localized pain complaints bore small but significant negative relationships to cognitive impairment. Pain was positively associated with physician-rated ill health and functional disability. The association between pain and cognitive status remained significant even when controlled statistically for effects of physical health and functional disability. Item-by-item examination of localized pain complaints indicated that markedly cognitively impaired individuals were less likely to report pain in the back and joints. However, examination of possible physical causes of reported pain revealed no differences between pain reports of cognitively impaired versus intact individuals in either the presence or the absence of a likely physical cause. CONCLUSIONS: These data provide no evidence for the "masking" of pain complaints by cognitive impairment. They suggest instead that, although cognitively impaired elderly may slightly underreport experienced pain, their self-reports are generally no less valid that those of cognitively intact individuals. Limitations of the research are acknowledged and implications for treatment of cognitively impaired institution residents are discussed.

Psychometric characteristics of the minimum data set I: confirmatory factor analysis.

OBJECTIVE: To determine the structure and statistical reliability of the federally mandated Minimum Data Set (MDS). DESIGN: Confirmatory, hypothesis-testing factor analysis was performed on MDS protocols of 733 nursing home residents. SETTING: All participants were residents of the Philadelphia Geriatric Center. PARTICIPANTS: Participants represented consecutively admitted skilled and intermediate care residents and another pool of residents with probable dementia. MEASUREMENTS: MDS protocols were completed by nurse care coordinators. Item composites hypothesized represented the domains of cognition, activities of daily living, time use, social quality, depression, and problem behaviors. RESULTS: For higher functioning residents (n = 336) and for all residents together, all domain clusters except social quality were confirmed. None of the domain clusters were confirmed within the more impaired (n = 391) group. CONCLUSIONS: The MDS does provide usable indicators of five areas of basic competence of nursing home residents. Lack of reliability in rating many aspects of the behavior and states of cognitively impaired residents is evident, however. Improvement of such measures and rating procedures constitutes a major research priority.

Validation of the Cumulative Illness Rating Scale in a geriatric residential population.

OBJECTIVE: To evaluate the validity of the Cumulative Illness Rating Scale (CIRS) in a geriatric institutional population by examining its associations with mortality, hospitalization, medication usage, laboratory findings and disability. DESIGN: A validation of the CIRS using self- and physician-report surveys, with archival data drawn from medical charts and facility records. SETTING: Long-term care facility with skilled nursing and congregate apartments. PARTICIPANTS: Four hundred thirty-nine facility residents selected on the basis of completeness of self-report data and physician ratings. PRIMARY MEASURES: Composite measures of illness severity and comorbidity, based on physicians' CIRS ratings; time to death or acute hospitalization after assessment; medication use, drawn from pharmacy records; medical chart data on laboratory tests; self-reported functional disability. RESULTS: CIRS illness severity and comorbidity indices, as well as individual CIRS items, were significantly associated with mortality, acute hospitalization, medication usage, laboratory test results, and functional disability. The CIRS showed good divergent validity vis a vis functional disability in predicting mortality and hospitalization. CONCLUSIONS: The CIRS appears to be a valid indicator of health status among frail older institution residents. The illness severity and comorbidity composites performed equally well in predicting longitudinal outcomes. Item-level analyses suggest that the CIRS may be useful in developing differential illness profiles associated with mortality, hospitalization, and disability.

Psychometric characteristics of the minimum data set II: validity.

OBJECTIVE: To determine the validity of the Minimum Data Set (MDS). DESIGN: MDS domain scores were correlated with a variety of independently obtained measures of basic behavioral and mental health functions of 513 nursing home residents. SETTING: All participants were residents of the Philadelphia Geriatric Center. PARTICIPANTS: One group of residents (n = 260) represented consecutive admissions who were able to respond to formal testing. The other group of residents (n = 253) represented presumably cognitively impaired residents whose data did not depend on self-report. MEASUREMENTS: MDS item-composite scores based on a confirmatory factor analysis were derived for the domains of cognition, activities of daily living (ADL), time use, depression, and problem behaviors. Hypotheses stating how these MDS domains should be related to standard measures of cognitive function, ADL, depression, agitation, social behavior, and irritability were tested. CONCLUSIONS: The majority of the hypotheses were upheld, thus suggesting that the MDS is usable as a source of research data. The sizes of the validity coefficients were modest, however. Depression and problem behavior were less well affirmed than cognition, ADL, and Time Use. There is a clear need for improvement in training and probably in the form of MDS measurement in some areas.

Depression and medical illness in late life: report of a symposium.

The high comorbidity of medical illnesses and late life depression poses both challenges and opportunities. Challenges in assessment techniques, diagnosis, and specific prognosis affect clinical care and research methodology alike. However, investigations that turn this vexing "confound" into research questions may prove fruitful. For clinicians working with older persons, recognizing the prognostic import of comorbid medical illnesses in late-life depression is essential to treatment planning. This comorbidity also poses difficulties in diagnosing depression inasmuch as symptoms of the medical conditions may overlap with those of an affective disorder. Symptom assessments must strike a balance between overly inclusive (e.g., mistakenly treating the psychomotor slowing of Parkinson's disease as depression) and overly exclusive (e.g., erroneously dismissing the patient's mood symptoms as "understandable"). Clinicians also should be sensitive to the broad range of symptomatic presentations with varying severities of both mood and medical disorders, as exemplified by variability across treatment settings. For researchers, similar issues are of relevance in planning investigative strategies. Consideration should be given to the following: 1. Case identification is a crucial first step; the approach to depressive symptoms potentially confounded by medical illnesses must be defined explicitly. Choice of an inclusive approach avoids premature exclusion of relevant phenomena; exploratory analyses can examine the effects of other approaches to the relationships of interest. 2. The use of similar research instruments across sample sites would greatly facilitate comparisons of results. Each subject group offers its own "leverage" for answering particular questions. Psychiatric inpatients will highlight the contributions of severe psychopathology (useful, for example, in identifying biologic markers). Medical inpatients are well suited to studies examining validity of different approaches to case identification, investigating health service utilization, or highlighting the contribution of acute, severe, life-threatening medical disorders to affective illness. Long-term care residents lend themselves to issues that benefit from compression of health processes over time. Medical outpatients have many advantages regarding generalizability and public health significance. Community samples are needed to determine the biases of all the above groups, which are each defined by service utilization. 3. Study of the relationships between depression and medical illness may further understanding of pathogenic mechanisms in late life mood disorders. Research questions might be guided by the biopsychosocial conceptual context described above. On the one hand, this context demands multidimensional study methodology to identify the routes by which medical illness influences depression in particular patient groups. Multivariate models should examine direct and indirect effects of medical illness on depression while, at the same time, considering intervening variables such as functional disability, personality, and social support. Guided multiple regressions or structural equation modeling will allow for determination of strengths of associations. 4. At the same time, and of particular importance if complex multivariate analyses are used, specific theoretic models should help direct focused investigations. The development and testing of such models is a major challenge that should be addressed by current research. Finally, from a societal perspective, the comorbidity of depression and medical illness likely has a tremendous impact on both health and health care delivery for older adults. Further study is needed to identify more specific approaches to treatment. Yet existing data clearly support a policy of routine psychiatric assessment of older people in general medical settings...

The structure of depression among elderly institution residents: affective and somatic correlates of physical frailty.

BACKGROUND: Confounding of depression with somatic illness and anxiety, a problem in any age group, may be especially troublesome in frail older persons. This paper examined this problem in a factor analytic study of the structure of depressive symptomatology, identifying affective and somatic symptom clusters and relating those clusters to health and functional variables cross-sectionally and prospectively over a 1-year interval. METHODS: The factor structure of a DSM-IV symptom checklist was examined among 1,245 elderly long-term care residents. Regression analyses examined the association of resulting factors with cognition, functional disability, self- and physician-rated health, and pain at baseline and a year later. One-year mortality was also examined. RESULTS: Factor analysis revealed three unique symptom clusters: depressed mood, somatic symptoms, and psychic anxiety. Depressed mood and somatic symptoms were associated cross-sectionally with all functional health variables, but psychic anxiety was associated only with pain. Longitudinally, depressed mood was the only independent predictor of decline in cognition, functional ability, physician-rated health, and mortality; the last effect, however, did not withstand control for baseline health and functioning. Somatic symptoms at baseline predicted decrement in self-rated health a year later. Effects varied as a function of cognitive status. CONCLUSIONS: These data suggest that concerns about the confounding role of somatic symptoms in the association of depression with physical health are unfounded. Although somatic symptoms of depression and anxiety were associated with health and functional status cross-sectionally, depressed mood was by far the stronger predictor of health declines over time.

Association of antidepressants and other medications with mortality in the residential-care elderly.

To explore the extent to which treatment of depression affects survival, we evaluated the association between use of antidepressant medications and death rates among the residents of a large residential-care facility for the elderly using a retrospective record-review study (N = 624). One year survival, among those taking antidepressants (10.9%), was 11.8% compared to 11.1% among the remainder of the population. A second study followed a group of 32 patients in the same institution who had participated in a therapeutic trial of nortriptyline treatment for major depression. Patients who experienced adverse medical events during treatment exhibited significantly increased mortality; among treatment completers, there was no significant relationship between mortality and therapeutic response. These findings suggest that the inability to tolerate treatment with an antidepressant can be considered a manifestation of physiologic frailty and increased vulnerability to mortality from disease. The previously reported decrease in survival among residential-care patients with major depression is not paralleled by a similar effect in those taking antidepressants. This may reflect selection factors with respect to the ability to tolerate antidepressants, rather an effect of treatment.

Depression and cognitive change among institutionalized aged.

Interrelations between depression (Geriatric Depression Scale) and cognitive impairment (Blessed test) were examined among 201 nursing home and congregate apartment residents in a 2-year, 3-wave study. In structural equations models that controlled autocorrelations and within-occasion correlated residuals, introducing paths from depression to subsequent cognitive status significantly reduced unexplained variance, whereas paths from cognitive status to subsequent depression did not. Subsidiary analyses indicated that the relation of depression to subsequent cognitive status was strongest among persons with borderline (vs. impaired or intact) cognitive status, but only for the first time interval. Discussion addresses explanations for obtained results and implications for monitoring and treating depression among elderly long-term care residents.

Affective states in normal and depressed older people.

Ratings on a 10-item affect checklist yielding composite positive affect and negative affect scores were made daily for 30 days by older people in residential care: 19 were diagnosed as having major depression, 21 had minor depression, and 37 were without psychiatric diagnosis ("normal"). Mean levels of positive affect were highest in normal people and least in those with major depression; negative affect was lowest in normal ones and highest in those with a major depression. Variability was least among those with major depression in positive affect and among normal people in negative affect, while residents with minor depression showed some tendency, although inconsistent, toward greater day-to-day variability in positive affect. Patterns of invariance were such that those with major depression tended to be consistently lacking in positive affect but were variable in negative affect; normal people showed variability in positive affect but a relatively unvarying lack of negative affect. Clinical major depression was thus characterized less by "pervasive" depressive affect than by anhedonia.

Pain in cognitively impaired older persons.

To summarize, there has been shamefully little empirical research directly examining the prevalence and correlates of pain among cognitively impaired older people. Even less is known about techniques for assessing and managing pain in this group. Existing evidence suggests that cognitively impaired older persons may voice fewer complaints about pain, but there is no reason to believe that they are in fact at less risk of pain than their cognitively intact age-mates. Rather, for whatever reason, persons with cognitively deficits appear to be less inclined to report pain than are intact elders of similar health status. This reporting difference may account at least in part for the fact that pain is less likely to be treated aggressively among cognitively impaired individuals. Unfortunately, knowing the reason for this state of affairs does not mitigate its implication: cognitive deficits place frail older persons at risk of unnecessary pain simply because it is not properly identified. Data reviewed in this chapter suggest that accurate assessment of pain in cognitively impaired older persons, far from being impossible, may actually be only slightly more demanding than it is in intact individuals. Even among markedly impaired elders, self-reports should certainly be taken as valid indicators; early evidence suggests promising avenues for developing reliable, clear-cut guidelines for the nonverbal assessment of pain in very severely demented individuals. As the nation grows older and medical care advances, a growing proportion of individuals can expect to live well into their eighth and even ninth decades. Unfortunately, with this extended life span comes increased likelihood of both cognitive impairment and pain. Thus, expansion of our repertoire of techniques for assessing and managing pain among cognitively impaired older persons must be a central priority for research on pain in late life.

Use of a falls incident reporting system to improve care process documentation in nursing homes.

BACKGROUND: Falls are the most frequently reported adverse event among frail nursing home residents and are an important resident safety issue. Incident reporting systems have been successfully used to improve quality and safety in healthcare. The purpose of this study was to test the effect of a systematically guided menu-driven incident reporting system (MDIRS) on documentation of post-fall evaluation processes in nursing homes. METHODS: Six for-profit nursing homes in southeastern USA participated in the study. Over a 4-month period, MDIRS was used in three nursing homes matched with another three nursing homes which continued using their existing narrative incident report to document falls. Trained geriatric nurse practitioner auditors used a data collection audit tool to collect medical record documentation of the processes of care for residents who fell. Multivariate analysis of covariance was used to compare the post-fall nursing care processes documented in the medical records. RESULTS: 207 medical records of resident who fell were examined. Over 75% of the sample triggered at high risk for falls by the minimum data set. An adequate neurological assessment was documented for only 18.4% of residents who had experienced a fall. Although two-thirds of the sample had a diagnosis of incontinence, less than 20% of the records had incontinence-related interventions in the nursing care plan. Overall, there was more complete documentation of the post-fall evaluation process in the medical records in nursing homes using the MDIRS than in nursing homes using standard narrative incident reports (p<0.001). CONCLUSION: Further improvements are necessary in reporting mechanisms to improve the post-fall assessment in nursing home residents.

Spouse versus other family caregivers: psychological impact on impaired aged.

This correlational study examined negative affect among impaired aged who live with and receive care from either a spouse or a younger relative. Covariance analyses, controlling for group differences in age, health, and functional disability, soundly disconfirmed hypotheses that spouses' care would prove less stressful than would that of children. In fact, the reverse obtained for measures of number of expressed worries, negative affect toward caregivers, and perceptions of both negative affect and negative behavior of caregivers toward elderly respondents. The two care groups did not, however, differ in self-rated depression or in feelings of dependency upon care providers. Examination of social support indices revealed that Spouse and Other care groups differed only in number of other household residents who assist with health care activities; this variable was, however, unrelated to affect measures. Results are discussed in terms of differential stressors upon aged couples in contrast with recipients of second-generation relatives' assistance, and of kinship factors which may influence elders' perceptions and affect.

The relation of pain to depression among institutionalized aged.

Nursing home and congregate apartment residents (N = 598) were classified on the basis of a DSM-IIIR symptom checklist as suffering possible major, minor, or no depression; they also completed the Geriatric Depression Scale (GDS) and the Profile of Mood States (POMS). Possible major depressives reported more intense pain and a greater number of localized pain complaints than did minor depressives; nondepressed individuals reported the least intense pain and fewest localized complaints. The effect remained strong even when functional disability and health status were controlled statistically. Both pain intensity and number of localized complaints were correlated with GDS and POMS factor scores, but strength and direction of associations varied with level of depression. Item-by-item examination of localized complaints again indicated that more depressed individuals were more likely to report pain, particularly where physicians had identified a physical problem that might account for the pain. Results are compared with previous research on pain among younger individuals. Implications for treatment of depressed elderly are discussed.

Incidence of depression in long-term care settings.

A one-year longitudinal study examined incidence and persistence of depression among nursing home and congregate apartment residents. Of 868 persons interviewed at Time 1, 15.7% displayed research-diagnosed possible major depression, and 16.5% displayed minor depressive symptoms. A follow-up interview a year later (T2; n = 448) yielded an incidence rate for possible major depression of 6.6% among persons with only minor or no depression in the previous year. For T1 nondepressives (i.e., excluding minor depressives), the incidence of possible major depression was 5.6%, and of minor depression, 6.3%. More than 40% of T1 possible major depressives showed no remission of symptoms a year later. In contrast, more than half of T1 minor depressives showed no depression at T2; however, another 16.2% appeared at follow-up to suffer possible major depression. Change in depression was associated with cognitive status, functional disability, and physical health. In general, persistence of depression was associated with greater decline, although patterns differed somewhat for the three functional indicators. Clinical and research implications of findings are discussed.

Depression and mortality among institutionalized aged.

This study examined the association between depression and mortality among a group of nursing home and congregate apartment residents (initial n = 898) over a 30-month period. Baseline [Time 1 (T1)] and 1-year follow-up [Time 2 (T2)] assessments yielded research-based diagnoses of possible major, minor, or no depression, along with measures of functional disability, cognitive status, and physician-rated health. Event history analyses were used to assess differential mortality as a function of level of depression after T1 and of change in depressive status from T1 to T2. Significant effects for T1 depression at 6, 12, and 18 months after the interview reflected an increased death rate among possible major depressives as compared with other respondents. An effect of change in depressive status from T1 to T2 appeared to be caused by long-term negative effects of T1 depression. Finally, none of the observed associations remained significant when controlled for effects of physical health, functional disability, and cognitive status. Thus, the effects of depression on mortality among this sample appeared to be attributable strictly to the correlation of depression with ill health. However, cautious interpretation is recommended inasmuch as causal paths between depression, ill health, and death remain unclear.

Depression among institutionalized aged: assessment and prevalence estimation.

Aged nursing home and congregate apartment residents were screened for symptoms of depression and cognitive impairment. Of 708 survey respondents, 12.4% met DSM-IIIR criteria (33) for major depression; about half this group also displayed significant cognitive deficits. Another 30.5% of the total sample reported less severe but nonetheless marked depressive symptoms. Such "minor" depressive syndromes were much more common among congregate housing than nursing home residents. Possible major depression was more prevalent among newly admitted residents of both housing components. Comparison of cognitively impaired vs intact respondents revealed that the two groups' self-reports of depression were equally internally consistent, and bore equivalent correlations with observer ratings made by interviewers and direct care staff. Checks of medical records of a group of survey nonrespondents (n = 203) indicated that, excepting the extremely demented, the active sample of 708 accurately represents institution residents as a whole. Finally, comparison with clinical diagnoses made by facility psychology and psychiatry department staff indicated good concurrent validity of research screening measures and methods.

The factorial generality of brief positive and negative affect measures.

Confirmatory factor analysis was used to test the structure of 5-item affect rating scales designed to measure positive affect and negative affect. A proposed circumplex affect structure was the source of scales constructed to represent a cluster of positive terms, including pleasantness and activation; the negative terms represented anxiety, depression, and hostility. The hypothesized simple-structured positive and negative trait affect factors, with a moderate correlation between them, were found in all cases. Equivalent structure was confirmed for younger adults, middle-aged, and older adults of good health and above-average education. Although the hypothesized simple-structured positive and negative factors emerged for all other groups, three other tests of factor equivalence failed to be confirmed: trait and state factors in the older adult group were not identical. Factors derived from healthy and frail elders were structurally different. Variability among frail elders and variability over 30 days within the same person, when factored, also showed nonequivalence. Although the scales are extremely useful in assessing affect, comparisons across some subject groups should be made with caution.