The Psychiatric Interview of Older Adults

DSM-IV and its text revision, DSM-IV-TR (American Psychiatric Association 1994, 2000), provide the clinician with a useful catalogue of symptoms and behaviors of psychiatric interest that are relevant to the diagnosis of the present illness. Symptoms are bits of data—the most visible part of the clinical picture and generally the part most easily agreed on among clinicians. Symptoms should be defined in such a way that, if clinicians each obtain equivalent information, minimal disagreement arises about the presence or absence of a symptom. The decision about whether those symptoms form a syndrome or derive from a particular etiology must be determined independently of the data collection on symptoms (see Chapter 2, “Demography and Epidemiology of Psychiatric Disorders in Late Life,” in The American Psychiatric Publishing Textbook of Geriatric Psychiatry, 3rd ed.).

Even so, the clinical interaction may be confounded by bias when a clinician communicates with an older adult about psychiatric symptoms. As many insightful clinicians, such as Eisenberg (1977), have recognized, physicians diagnose and treat diseases—that is, abnormalities in the structure and function of body organs and systems. Patients have illnesses—experiences of disvalued changes in states of being and in social function. Disease and illness do not maintain a one-to-one relationship. Factors that determine who becomes a patient and who does not can be understood only by expanding horizons beyond symptoms. In other words, patienthood is a social state (Eisenberg and Kleinman 1981). During the process of becoming a patient, the older adult, usually with the advice of others, forms a self-diagnosis of his or her problem and makes a judgment about the degree of ill-being perceived. For some, illness is perceived when a specific discomfort is experienced. For others, illness reflects a general perception of physical or social alienation and despair. Given that few uniform, satisfactory definitions of illness (or ill-being) exist, it is not surprising that terms for wellness (or well-being) also mean different things to different people. The historical background and the values of the older adult in a social class and culture contribute to the formation of constructs regarding the nature of the problem, the cause, and the possibility for recovery.

For this reason, the clinician must take care to avoid accepting the patient’s explanation for a given problem or set of problems. Statements such as “I guess I’m just getting old and there’s nothing really to worry about” or “Most people slow down when they get to be my age” can lull the clinician into complacency about what may be a treatable psychiatric disorder. On the other hand, the advent of new and disturbing symptoms in an older adult between each office visit can exhaust the clinician’s patience to the point at which adequate pursuit of the problem is derailed. For example, the older adult with hypochondria whose difficulty with awakenings during the night is increasing may insist that this symptom be treated with a sedative and plead with the clinician not to allow continual suffering. In the clinician’s view, however, the symptom is a normal accompaniment of old age and therefore should be accepted. Distress over changes in functioning, such as sexual functioning, may overwhelm the older adult patient and, especially if the clinician is perceived as unconcerned, may precipitate self-medication or even a suicide attempt.

To prevent attitudinal biases when eliciting reports by the older adult (which may result in missing the symptoms and signs of a treatable psychiatric disorder), the clinician must include in the initial interview a review of the more important psychiatric symptoms in a relatively structured format. Common symptoms that should be reviewed include excessive weakness or lethargy; depressed mood or the blues; memory problems; difficulty concentrating; feelings of helplessness, hopelessness, and uselessness; isolation; suspicion of others; anxiety and agitation; sleep problems; and appetite problems. Critical symptoms that should be reviewed include the presence or absence of suicidal thoughts, profound anhedonia, impulsive behavior (“I can’t control myself”), confusion, and delusions and hallucinations.

The review of symptoms is most valuable when it is considered in the context of symptom presentation. When did the symptoms begin? How long have they lasted? Has their severity changed over time? Are there physical or environmental events that precipitate the symptoms? What steps, if any, have been taken to try to correct the symptoms? Have any of these interventions proved successful? Do the symptoms vary during the day (diurnal variation)? Do they vary during the week or with seasons of the year? Do the symptoms form clusters—that is, are they associated with one another? Which symptoms appear ego-syntonic and which symptoms appear ego-dystonic? As symptoms are reviewed, a specific time frame facilitates focus on the present illness. Having a 1-month or 6-month window enables the patient to review symptoms and events temporally, an approach not usually taken by distressed elders, who tend to concentrate on immediate sufferings.

Critical to the assessment of the present illness is an assessment of function and change in function. The two parameters that are most important (and not included in usual assessments of physical and psychiatric illness) are social functioning and activities of daily living (ADLs). Questions should be asked about the social interaction of the older adult, such as the frequency of his or her visits outside the home, telephone calls, and visits from family and friends. Many scales have been developed to assess ADLs; however, in the interview the clinician can simply ask about ability to get around (for example, walk inside and outside the house), to perform certain physical activities independently (such as bathe, dress, shave, brush one’s teeth, and pick out one’s clothes) and to do instrumental activities (such as cook, keep one’s bank account, shop, and drive).

Next, the clinician must review the past history of symptoms and episodes. The patient should be asked if he or she has had a similar episode or episodes in the past. How long did the episodes last? When did they occur? How many times in the patient’s lifetime have such episodes occurred? Unfortunately, the older adult may not equate present distress with past episodes that are symptomatically similar, so the perspective of the family is especially valuable in the attempt to link current and past episodes. Other psychiatric and medical problems should be reviewed as well, especially medical illnesses that have led to hospitalization and the use of medication. Not infrequently, the older adult has experienced a major illness or trauma in childhood or as a younger adult, but he or she views this information as being of no relevance to the present episode and therefore dismisses it. Probes to elicit these data are essential. Older adults may ignore or even forget past psychiatric difficulties, especially if these difficulties were disguised. For example, mood swings in early or middle life may have occurred during periods of excessive and productive activity, episodes of excessive alcohol intake, or periods of vague, undiagnosed physical problems. Previous periods of overt disability in usual activities may flag those episodes. An older person sometimes becomes angry or irritated when the clinician continues to probe. Reassurance regarding the importance of obtaining this information will generally suffice, except when dealing with a patient who cannot tolerate the discomfort and distress, even for brief periods. Older persons who have chronic and moderately severe anxiety or a histrionic personality style, as well as distressed Alzheimer’s patients, tolerate their symptoms poorly.

The distribution of psychiatric symptoms and illnesses in the family should be determined next. The older person with symptoms consistent with senile dementia or primary degenerative dementia is highly likely to have a family history of dementia. The genogram remains one of the best means for evaluating the distribution of mental illness and other relevant behaviors throughout the family tree. This genogram should include both of the parents, blood-related aunts and uncles, brothers and sisters, spouse(s), children, grandchildren, and great-grandchildren. A history should be obtained about institutionalization, significant memory problems in family members, hospitalization for a nervous breakdown or depressive disorder, suicide, alcohol abuse and dependence, electroconvulsive therapy, long-term residence in a mental health facility (and possibly a diagnosis of schizophrenia), and use of mental health services by family members (Blazer 1984).

Of relevance to the pharmacological treatment of certain disorders in older adults—especially depression—is the tendency of individuals in a family to respond therapeutically to the same pharmacological agent. If the older adult has a depressive disorder and biological relatives have been treated effectively for depression, the clinician should determine what pharmacological agent was used to treat the depression. For example, a positive response to sertraline in a family member of the depressed older patient could make sertraline the drug of choice in treating that patient, assuming side effects are not at issue (Ayd 1975).

Mendlewicz and colleagues (1975) remind us that accurate genetic information can be better obtained when family members from more than one generation are interviewed. Many psychiatric disorders are characterized by a variety of symptoms, so asking the patient or one family member for a history of depression is insufficient. Research on the genetic expression of psychiatric disorders in families requires the psychiatric investigator to interview directly as many family members as possible to determine accurately the distribution of disorders throughout the family. Such detailed family assessment is not feasible for clinicians, yet a telephone call to a relative with permission from the patient may become a standard of clinical assessment as the genetics of psychiatric disorders are clarified.

Psychiatric disorders occur in a biomedical and psychosocial context. The clinician, although he or she will of course determine what medical problems the patient has experienced, might overlook a variation in the relative contribution of these medical disorders to psychopathology. The psychosocial contribution to the onset and continuance of the problem is just as likely to be overlooked. Has the spouse of the older adult undergone a change? Are the middle-aged children managing high stress, such as caring for an emotionally disturbed child and the loss of employment simultaneously? Are the grandchildren placing emotional stress on the elderly patient, perhaps requesting money? Has the economic status of the older adult deteriorated? Has the availability of medical care changed? Although many psychiatric disorders are biologically driven, they do not occur in a psychosocial vacuum. Environmental precipitants remain important in the web of causation leading to the onset of an episode of emotional distress and are critical to the assessment of the older adult.

Next, it is essential to evaluate the medication history of the older adult. A careful review of medications by the clinician is essential, although this may be done by a nurse or a physician’s assistant. The clinician should ask the older person to bring in all pill bottles as well as a list of medications taken and the dosage schedule. A double check between the written schedule and the pill containers will frequently expose some discrepancy. Both prescription and over-the-counter drugs, such as laxatives and vitamins, should be recorded. The clinician can then identify the medications that are potentially critical in terms of drug-drug interactions and ask about them during subsequent visits.

Most elderly persons take a variety of medicines simultaneously, and the potential for drug-drug interaction is high. For example, concomitant use of fluoxetine and warfarin has been associated with an increase in the half-life of warfarin, which could lead to severe bruising (although this finding is not well documented). Some medications prescribed for older persons—such as the beta-blocker propranolol and the antihypertensive drug alpha-methyldopa—can exacerbate or produce depressive symptoms. Antianxiety agents and sedative-hypnotics can precipitate episodes of confusion and depression. Antidepressants, such as the tricyclics (TCAs), may adversely interact with other drugs, including the antihypertensive agent clonidine. Simultaneous administration of clonidine and a TCA may lead to poorly controlled episodes of hypertension with confusional episodes and possibly an exacerbation of vascular (multi-infarct) dementia. The physician, a nurse, a social worker, or a paraprofessional should carefully determine present and past medication use through a historical inventory and a review of the patient’s medicine containers brought to the office.

Older persons are less likely than younger persons to abuse alcohol, but a careful history of alcohol intake is essential to the diagnostic workup. Older persons do not usually volunteer information about their alcohol intake, but they are generally forthcoming when asked about their drinking habits. Substance abuse beyond alcohol and prescription drugs is rare in older adults but not entirely absent.

A number of standardized assessment methods for delirium have emerged. Perhaps the best and the most easily used is the Confusion Assessment Method (CAM; Inouye 1990). This scale has incorporated DSM criteria for confusion into nine operationalized criteria, including acute onset (evidence of such onset), fluctuating course (behavior change during the day), inattention (trouble in focusing), disorganized thinking (presence of rambling or irrelevant conversations and illogical flow of ideas), and altered level of consciousness (rated from alert to comatose). Diagnosis requires both acute onset and fluctuating course, along with inattention and either disorganized thought or altered level of consciousness.

Two interviewer-administered cognitive screens for dementia have been popular in both clinical and community studies. The first is the Short Portable Mental Status Questionnaire (SPMSQ; Pfeiffer 1975), a derivative of the Mental Status Questionnaire developed by Kahn and colleagues (1960). The SPMSQ consists of 10 questions designed to assess orientation, memory, fund of knowledge, and calculation. For most community-dwelling older adults, two or fewer errors indicate intact functioning; three or four errors, mild impairment; five to seven errors, moderate impairment; and eight or more errors, severe impairment. The ease of administration of this instrument and its reliability as supported by accumulated epidemiological data make it useful for both clinical and community screens.

The Mini-Mental State Examination (Folstein et al. 1975) is a 30-item instrument that assesses orientation, registration, attention and calculation, recall, and language. It requires 5–10 minutes to administer and includes more items of clinical significance than does the SPMSQ. Seven to 12 errors suggest mild to moderate cognitive impairment and 13 or more errors severe impairment. This instrument is perhaps the most frequently used standardized screening instrument in clinical practice.

A number of clinical assessment procedures for dementia have emerged in recent years. The most widely used, and one of the first to appear, is the scale suggested by Blessed et al. (1968), usually referred to as the Blessed Dementia Index. In contrast to what can be gleaned by use of the screening scales described, clinical judgment is required in using the Blessed Dementia Index to assess changes in performance of everyday activities, such as handling money, household tasks, and shopping; changes in eating and dressing habits; changes in personality, interests, and drive; tests of information (orientation and recognition of persons); memory of past information, such as occupation, place of birth, and town where the individual worked; and concentration (calculation task). A score is assigned to each of these tasks, and a summary score is tabulated. The score has been shown to correlate well with the cerebral changes of primary degenerative dementia.

A more recent and comprehensive scale is the Alzheimer’s Disease Assessment Scale (Rosen et al. 1984). This clinical rating scale includes ratings of spoken language, language comprehension, recall, ability to follow commands, word-finding difficulty in spontaneous speech, ability to name objects, constructional praxis, ideational praxis, orientation, word recall, word recognition, and a series of noncognitive behaviors, such as tearfulness, distractibility, depression, and motor activity.

A dementia scale for assessing the probability that dementia is secondary to multiple infarcts was suggested by Hachinski et al. (1975). In the study, cerebral blood flow in patients with primary degenerative dementia was compared with those who had multi-infarct dementia. Certain clinical features were determined to be more associated with multi-infarct dementia, and each of these features was assigned a score. Those clinical features, along with their scores, are as follows: abrupt onset=2, stepwise deterioration=1, fluctuating course=2, nocturnal confusion=1, relative preservation of personality=1, depression=1, somatic complaints=1, emotional incontinence=1, history of hypertension=1, history of strokes=2, evidence of associated atherosclerosis=1, focal neurological symptoms=2, and focal neurological signs=2. A score of 7 or greater was highly suggestive of multi-infarct dementia. However, given the frequent overlap of multiple small infarcts and primary degenerative dementia, as well as the difficulty of assessing these items effectively, most investigators have ceased to rely on the Hachinski scale for clinical use.

A number of self-rating depression scales have been used to screen for depression in patients at all stages of the life cycle; most of these scales have been studied in older populations. The Zung Self-Rating Depression Scale (Zung 1965) was the most widely used until recent years. The initial popularity of the Zung scale was probably due to the availability of data for persons throughout the life cycle, especially elderly persons (Zung 1967). Few randomly sampled community populations have been surveyed with the scale; therefore, a deficit exists in normative community standards. In this 20-item scale, each of the 20 symptoms is ranked from 0 (“none”) to 3 (“all or about all the time”), according to severity. Most older adults can use the Zung scale, although the four choices may create problems for some elderly persons with mild cognitive impairment. Using the Zung scale, Freedman and colleagues (1982) found peak symptom levels in women ages 65–69 years and men ages 70–74 years.

The most widely used of the current instruments in community studies is the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977). This instrument, because of the normative population data available for it, has replaced the Zung scale in recent years as a common instrument for screening for depression. The CES-D scale is similar in format to the Zung scale. In a factor-analytic study of the CES-D in a community population, three factors were identified: enervation, positive affect, and interpersonal relationships (Ross and Mirowsky 1984). The disaggregation of these factors and the exploration of their interaction are significant steps forward in understanding the results derived from symptom scales such as the CES-D in older populations. For example, the enervation items (e.g., loss of interest, poor appetite) are more likely to be associated with a course of depressive episodes similar to that described for major depression with melancholia, and the positive-affect items more likely to be associated with life satisfaction scores.

A scale that has been widely used in clinical studies, although less studied in community populations, is the Beck Depression Inventory (BDI; Beck et al. 1961). The reliability of the BDI has been shown to be good in both depressed and nondepressed samples of older people (Gallagher et al. 1982). The instrument consists of 21 symptoms and attitudes, rated on a scale of 0 to 3 in terms of intensity. In another study by Gallagher et al. (1983), the BDI misclassified only 16.7% of subjects who had been diagnosed on the basis of the RDC as having major depression.

The Geriatric Depression Scale (GDS) was developed because the scales discussed above pre-sent problems for older persons who have difficulty in selecting one of four forced-response items (Yesavage et al. 1983). The GDS is a 30-item scale that permits patients to rate items as either present or absent; it includes questions about symptoms such as cognitive complaints, self-image, and losses. Items selected were thought to have relevance to late-life depression. The GDS has not been used extensively in community populations and is not as well standardized as the CES-D, but its yes/no format is preferred to the CES-D by many clinicians.

Of the interviewer-rated scales, the Hamilton Rating Scale for Depression (Ham-D; Hamilton 1960) is by far the most commonly used. The advantage of having ratings based on clinical judgment has made the Ham-D a popular instrument for rating outcome in clinical trials. For example, a reduction in the score to one-half the initial score or to a score below a certain value would indicate partial or complete recovery from an episode of depression.

A scale that has received considerable attention clinically, standardized in clinical but not community populations, is the Montgomery-Åsberg Rating Scale for Depression (Montgomery and Åsberg 1979). This scale follows the pattern of the Hamilton scale and concentrates on 10 symptoms of depression; the clinician rates each symptom on a scale of 0 to 6 (for a range of scores between 0 and 60). The symptoms include apparent sadness, reported sadness, inattention, reduced sleep, reduced appetite, concentration difficulties, lassitude, inability to feel, pessimistic thoughts, and suicidal thoughts. This scale, theoretically, is an improvement over the Hamilton scale in that it appears to better differentiate between responders and nonresponders to intervention for depression. The instrument does not include many somatic symptoms that tend to be more common in older adults, and therefore it may be of greater value in tracking the symptoms of depressive illness that would be expected to change with therapy.

A number of general assessment scales of psychiatric status (occasionally combined with functioning in other areas) have been found to be useful in both community and clinical populations.

One of the more frequently used scales is the Global Assessment of Functioning Scale (GAF; American Psychiatric Association 2000). Using this scale, the rater makes a single rating, ranging from 0 to 100, that best describes—on the basis of his or her clinical judgment—the lowest level of the subject’s functioning in the week before the rating. The scale has not been standardized for older adults, but its common use in psychiatric studies suggests the need for standardization. The scale was incorporated as Axis V in DSM-IV to measure overall functioning.

The Geriatric Mental State Schedule (Copeland et al. 1976), an adaptation of the Present State Exam (PSE; Wing et al. 1974) and the Psychiatric Status Schedule (Spitzer et al. 1968), is a semistructured interviewing guide that allows the rater to inventory symptoms associated with psychiatric disorders. More than 500 ratings are made on the basis of information obtained by a highly trained interviewer, who elicits reports of symptoms from the month preceding the evaluation. Data are computerized to derive psychiatric diagnoses (Copeland et al. 1986). The instrument measures depression, impaired memory, selected neurological symptoms such as aphasia, and disorientation.

The Comprehensive Assessment and Referral Evaluation (CARE; Gurland et al. 1977) is a hybridized assessment procedure developed for older adults. Dimensional scores are obtained in Memory–Disorientation, Depression–Anxiety, Immobility–Incapacity, Isolation, Physical–Perceptual Difficulty, and Poor Housing–Income. The goal of CARE is to provide a comprehensive assessment of the older adult that bridges the professional disciplines. The instrument has not been used extensively, although it has been used in cross-national studies. For example, Herbst and Humphrey (1980) used CARE in a study examining how hearing impairment relates to mental status. The investigators found a relationship between deafness and depression that was independent of age and socioeconomic status.

The Older Americans Resources and Services (OARS) Multidimensional Functional Assessment Questionnaire (Duke University Center for the Study of Aging and Human Development 1978), administered by a lay interviewer, produces functional impairment ratings in five dimensions: mental health, physical health, social functioning, economic functioning, and activities of daily living. In one community survey using OARS (Blazer 1978a), 13% of persons in the community were found to have mental health impairment. The OARS instrument was developed to integrate functional measures across a series of parameters relevant to older adults; it has been used widely in both community and clinical surveys. With the recent emphasis on discrete psychiatric disorders, however, the instrument has not been as widely used by mental health workers as it might otherwise have been.

Any discussion of clinical rating scales is not complete without a discussion of the Abnormal Involuntary Movement Scale (AIMS; National Institute of Mental Health 1975). There has been an increased incidence of tardive dyskinesia among older adults, coupled with the need for better documentation of this dreaded outcome of prolonged use of antipsychotic agents. Regular rating of patients on the AIMS by clinicians has therefore become essential to the practice of inpatient and outpatient geriatric psychiatry. The scale consists of seven movement disorders; the presence and severity of each is rated from “none” to “severe.” Three items require a global judgment: severity of abnormal movements, incapacitation due to abnormal movements, and the patient’s awareness of abnormal movements. Current problems with teeth or dentures are also assessed. Procedures are described to increase the reliability of this rating scale.

A number of structured interview schedules are now available for both clinical and community diagnosis. These interview schedules have allowed increased reliability of the identification of particular symptoms and psychiatric diagnoses. Unfortunately, if one adheres closely to the structured interview, the richness inherent in the unstructured interview tends to be lost. Comments made by the patient during the evaluation that could be used to trace relevant associations must be ignored to push through the interview schedule. Most of these interviews require more time than the traditional unstructured first session with the patient.

The oldest of the currently used interview schedules is the Present State Exam (PSE) (Wing et al. 1974). As noted above, the Geriatric Mental State Schedule is a variant of the PSE. The PSE is not an interview but a list of definitions of behaviors or symptoms of psychiatric interest, ranging from specific delusions to general changes in affect. The clinician scores whether the symptom is present, and a computer algorithm provides a diagnosis. Suggested questions for eliciting reports of the symptoms are available but not obligatory. Only 54 questions are required during the interview, although many additional probes are provided to track positive responses. The interview schedule provides an excellent education for many psychiatrists about the meaning of various symptoms relevant to work with older adults. Nevertheless, the focus on 1 month before the evaluation date and the association of the symptoms with the World Health Organization’s International Classification of Diseases (ICD)—instead of the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM)—make the PSE less popular with American investigators.

The most frequently used instrument in the United States is the Structured Clinical Interview for DSM-IV (SCID; First et al. 1997). This instrument is easily adaptable to the RDC, DSM-IV, and DSM-IV-TR. Although specific questions are suggested for probing most areas of interest, the interviewer using the SCID has the flexibility to ask additional questions and can use any available data to assign a diagnosis. The interviewer must have clinical training but does not have to be a psychiatrist. Many of the symptoms may not be relevant to older adults (especially the extensive probes for psychotic symptoms), and the interview frequently takes 2½–3 hours to administer. Nevertheless, the experience gained by the clinician in using this instrument can contribute to a more effective clinical practice.

A relatively recent addition to the schedules available is the Diagnostic Interview Schedule (DIS; Robins et al. 1981). This highly structured, computer-scored interview, which can be administered by a lay interviewer, allows psychiatric diagnoses to be made according to DSM-IV criteria, Feighner criteria (Feighner et al. 1972), and RDC. The DIS questions probe for the presence or absence of symptoms or behaviors relevant to a series of psychiatric disorders, the severity of the symptoms, and the putative cause of the symptoms. Diagnoses of cognitive impairment, schizophrenia or schizophreniform disorder, major depression, generalized anxiety disorder, panic disorder, agoraphobia, obsessive-compulsive disorder, dysthymic disorder, somatization disorder, alcohol abuse and/or dependence, and other substance abuse and/or dependence can be made from Axis I of DSM-IV. A diagnosis of antisocial personality disorder (Axis II) can also be made. The instrument has proved reasonably reliable in clinic populations for both current and lifetime diagnoses.

The range of disorders probed by the DIS questions, coupled with the instrument’s relative ease of administration (it generally takes 45–90 minutes to administer to an older adult), has made it popular for use in clinical studies. In addition, community-based comparative data are available on a large sample from the Epidemiologic Catchment Area study (Myers et al. 1984; Regier et al. 1984). The DIS can be supplemented with additional questions to probe for specific symptoms, such as melancholic symptoms and additional data on sleep disorders for depressed older adults. No problems have arisen when the instrument is used among older adults in the community. The memory decay that occurs in elderly persons in general is no more of a problem with this instrument than with others. Nevertheless, the DIS is of less value in the study of institutional populations and in reconstruction of lifetime history regardless of setting, because memory problems cannot be circumvented by clinical judgment. Supplementary data can be added to the instrument for developing a standardized diagnosis. A shortened version of the DIS, which has been used in recent epidemiological surveys, is the Composite International Diagnostic Interview (CIDI; World Health Organization 1989).