Chapter 5
Older Adults and Mental Health

Chapter Overview

Overview of Mental Disorders in Older Adults

Depression in Older Adults

Alzheimer’s Disease

Other Mental Disorders in Older Adults

Service Delivery

Other Services and Supports



Other Mental Disorders in Older Adults

Anxiety Disorders

Prevalence of Anxiety
Anxiety symptoms and syndromes are important but understudied conditions in older adults. Overall, community-based prevalence estimates indicate that about 11.4 percent of adults aged 55 years and older meet criteria for an anxiety disorder in 1 year (Flint, 1994; Table 5-1). Phobic anxiety disorders are among the most common mental disturbances in late life according to the ECA study (Table 5-1). Prevalence studies of panic disorder (0.5 percent) and obsessive-compulsive disorder (1.5 percent) in older samples reveal low rates (Table 5-1) (Copeland et al., 1987a; Copeland et al., 1987b; Bland et al., 1988; Lindesay et al., 1989). Although the National Comorbidity Study did not cover this age range, and the ECA did not include this disorder, other studies showed a prevalence of generalized anxiety disorder in older adults ranging from 1.1 percent to 17.3 percent higher than that reported for panic disorder or obsessive-compulsive disorder (Copeland et al., 1987a; Skoog, 1993). Worry or “nervous tension,” rather than specific anxiety syndromes, may be more important in older people. Anxiety symptoms that do not fulfill the criteria for specific syndromes are reported in up to 17 percent of older men and 21 percent of older women (Himmelfarb & Murrell, 1984).

In addition, some disorders that have received less study in older adults may become more important in the near future. For example, post-traumatic stress disorder (PTSD) is expected to assume increasing importance as Vietnam veterans age. At 19 years after combat exposure, this cohort of veterans has been found to have a PTSD prevalence of 15 percent (cited in McFarlane & Yehuda, 1996). As affected patients age, there is a continuing need for services. In addition, research has shown that PTSD can manifest for the first time long after the traumatic event (Aarts & Op den Velde, 1996), raising the specter that even more patients will be identified in the future.

Treatment of Anxiety
The effectiveness of benzodiazepines in reducing acute anxiety has been demonstrated in younger and older patients, and no differences in the effectiveness have been documented among the various benzodiazepines. Some research suggests that benzodiazepines are marginally effective at best in treating chronic anxiety in older patients (Smith et al., 1995).

The half-life of certain benzodiazepines and their metabolites may be significantly extended in older patients (particularly for the compounds with long half-life). If taken over extended periods, even short-acting benzodiazepines tend to accumulate in older individuals. Thus, it is generally recommended that any use of benzodiazepines be limited to discrete periods (less than 6 months) and that long-acting compounds be avoided in this population. On the other hand, use of short-acting compounds may predispose older patients to withdrawal symptoms (Salzman, 1991).

Side effects of benzodiazepines may include drowsiness, fatigue, psychomotor impairment, memory or other cognitive impairment, confusion, paradoxical reactions, depression, respiratory problems, abuse or dependence problems, and withdrawal reactions. Benzodiazepine toxicity in older patients includes sedation, cerebellar impairment (manifested by ataxia, dysarthria, incoordination, or unsteadiness), cognitive impairment, and psychomotor impairment (Salzman, 1991). Psychomotor impairment from benzodiazepines can have severe consequences, leading to impaired driver skills and motor vehicle crashes (Barbone et al., 1998) and falls (Caramel et al., 1998).

Buspirone is an anxiolytic (antianxiety) agent that is chemically and pharmacologically distinct from benzodiazepines. Controlled studies with younger patients suggest that the efficacy of buspirone is comparable to that of the benzodiazepines. It also has proven effective in studies of older patients (Napoliello, 1986; Robinson et al., 1988; Bohm et al., 1990). On the other hand, buspirone may require up to 4 weeks to take effect, so initial augmentation with another antianxiety medication may be necessary for some acutely anxious patients (Sheikh, 1994). Significant adverse reactions to buspirone are found in 20 to 30 percent of anxious older patients (Napoliello, 1986; Robinson et al., 1988). The most frequent side effects include gastrointestinal symptoms, dizziness, headache, sleep disturbance, nausea/vomiting, uneasiness, fatigue, and diarrhea. Still, buspirone may be less sedating than benzodiazepines (Salzman, 1991; Seidel et al., 1995).

Although the efficacy of antidepressants for the treatment of anxiety disorders in late life has not been studied, current patterns of practice are informed by the efficacy literature in adults in midlife (see Chapter 4).

Schizophrenia in Late Life

Although schizophrenia is commonly thought of as an illness of young adulthood, it can both extend into and first appear in later life. Diagnostic criteria for schizophrenia are the same across the life span, and DSM-IV places no restrictions on age of onset for a diagnosis to be made. Symptoms include delusions, hallucinations, disorganized speech, disorganized or catatonic behavior (the so-called “positive” symptoms), as well as affective flattening, alogia, or avolition5 (the so-called “negative” symptoms). Symptoms must cause significant social or occupational dysfunction, must not be accompanied by prominent mood symptoms, and must not be uniquely associated with substance use.

Prevalence and Cost
One-year prevalence of schizophrenia among those 65 years or older is reportedly only around 0.6 percent, about one-half the 1-year prevalence of the 1.3 percent that is estimated for the population aged 18 to 54 (Tables 5-1 and 4-1).

The economic burden of late-life schizophrenia is high. A study using records from a large California county found the mean cost of mental health service for schizophrenia to be significantly higher than that for other mental disorders (Cuffel et al., 1996); the mean expenditure among the oldest patients with schizophrenia (> 74 years old) was comparable to that among the youngest patients (age 18 to 29). While long-term studies have shown that use of nursing homes, state hospitals, and general hospital care by patients with all mental disorder diagnoses has declined in recent decades, the rate of decline is lower for older patients with schizophrenia (Kramer et al., 1973; Redick et al., 1977). The high cost of these settings contributes to the greater economic burden associated with late-life schizophrenia.

Late-Onset Schizophrenia
Studies have compared patients with late onset (age at onset 45 years or older) and similarly aged patients with earlier onset of schizophrenia (Jeste et al., 1997); both were very similar in terms of genetic risk, clinical presentation, treatment response, and course.

Among key differences between the groups, patients with late-onset schizophrenia were more likely to be women in whom paranoia was a predominant feature of the illness. Patients with late-onset schizophrenia had less impairment in the specific neurocognitive areas of learning and abstraction/cognitive flexibility and required lower doses of neuroleptic medications for management of their psychotic symptoms. These and other differences between patients with early- and late-onset illness suggest that there might be neurobiologic differences mediating the onset of symptoms (DeLisi, 1992; Jeste et al., in press).

Course and Recovery
The original conception of “dementia praecox,” the early term for schizophrenia, emphasized progressive decline (Kraepelin, 1971); however, it now appears that Kraepelin’s picture captures the outcome for a small percentage of patients, while one-half to two-thirds significantly improve or recover with treatment and psychosocial rehabilitation (Chapter 4). Although the rates of full remission remain unclear, some patients with schizophrenia demonstrate remarkable recovery after many years of chronic dysfunction (Nasar, 1998). Research suggests that a factor in better long-term outcome is early intervention with antipsychotic medications during a patient’s first psychotic episode (see Chapter 4).

A recent cross-sectional study that compared middle-aged with older patients, all of whom lived in community settings, found some similarities and differences (Eyler-Zorrilla et al., 1999). The older patients experienced less severe symptoms overall and were on lower daily doses of neuroleptics than middle-aged patients who were similar in demographic, clinical, functional, and broad cognitive measures. In addition, positive symptoms were less prominent (or equivalent) in the older group, depending on the measure used. Negative symptoms were more prominent (or equivalent) in the older group, and older patients scored more poorly on severity of dyskinesia. Older patients were impaired relative to middle-aged ones on two measures of global cognitive function. This finding, however, appeared to reflect a normal degree of decline from an impaired baseline, as the degree of change in cognitive function with age in the patient group was equivalent to that seen in the comparison group.

A recent study used the Direct Assessment of Functional Status scale (DAFS) (Loewenstein et al., 1989) to compare the everyday living skills of middle-aged and older adults with schizophrenia with those of people without schizophrenia of similar ages (Klapow et al., 1997). The patients exhibited significantly more functional limitations than the controls did across most DAFS subscales. In another recent study that used a measure of overall disease impact, the Quality of Well-Being Scale, older outpatients with schizophrenia manifested significantly lower quality of well-being than did comparison subjects, and their scores were slightly worse than those of ambulatory AIDS patients (Patterson et al., 1996).

Thus, while schizophrenia may be less universally deteriorating than previously has been assumed, older patients with the disorder continue nonetheless to exhibit functional deficits that warrant research and clinical attention.

Etiology of Late-Onset Schizophrenia
Recent studies support a neurodevelopmental view of late-onset schizophrenia (Jeste et al., 1997). Equivalent degrees of childhood maladjustment have been found in patients with late-onset schizophrenia and early-onset schizophrenia, for example, suggesting that some liability for the disorder exists early in life. Equivalent degrees of minor physical anomalies in patients with late-onset schizophrenia and early-onset schizophrenia suggest the presence of developmental defects in both groups (Lohr et al., 1997). The presence of a genetic contribution to late-onset and early-onset schizophrenia is evident in increased rates of schizophrenia among first-degree relatives (Rokhlina, 1975; Castle & Howard, 1992; Castle et al., 1997).

If late-onset schizophrenia is neurodevelopmental in origin, an explanation for the delayed onset may be that late-onset schizophrenia is a less severe form of the disorder and, as such, is less likely to manifest early in life. Recent research suggests that in several arenas—for example, neuropsychological impairments in learning, retrieval, abstraction, and semantic memory as well as electroencephalogram abnormalities—the deficits of patients with late-onset schizophrenia are less severe (Heaton et al., 1994; Jeste et al., 1995b; Olichney et al., 1995, 1996; Paulsen et al., 1995, 1996). Also, negative symptoms are less pronounced and neuroleptic doses are lower in patients with late-onset schizophrenia (Jeste et al., 1995b). The etiology and onset of schizophrenia in younger adults often are explained by a diathesis-stress model in which there is a genetic vulnerability in combination with an environmental insult (such as obstetric complications), with onset triggered by maturational changes or life events that stress a developmentally damaged brain (Feinberg, 1983; Weinberger, 1987; Wyatt, 1996). Under this multiple insult model, patients with late-onset schizophrenia may have had fewer insults and thus have a delayed onset. An alternative or complementary explanation for the delayed onset in late-onset schizophrenia is the possibility that these patients possess protective features that cushion the blow of any additional insults. The preponderance of women among patients with late-onset schizophrenia has fueled hypotheses that estrogen plays a protective role.

The view of late-onset schizophrenia as a less severe form of schizophrenia, in which the delayed onset results from fewer detrimental insults or the presence of protective factors, suggests a continuous relationship between age at onset and severity of liability. An alternative view is that late-onset schizophrenia is a distinct neurobiological subtype of schizophrenia. The preponderance of women and of paranoid subtype patients seen in late-onset schizophrenia supports this view. These two etiologic theories of late-onset schizophrenia call for further research.

Treatment of Schizophrenia in Late Life
Pharmacological treatment of schizophrenia in late life presents some unique challenges. Conventional neuroleptic agents, such as haloperidol, have proven effective in managing the “positive symptoms” (such as delusions and hallucinations) of many older patients, but these medications have a high risk of potentially disabling and persistent side effects, such as tardive dyskinesia (Jeste et al., in press). The cumulative annual incidence of tardive dyskinesia among older outpatients (29 percent) treated with relatively low daily doses of conventional antipsychotic medications is higher than that reported in younger adults (Jeste et al., in press).

Recent years have witnessed promising advances in the management of schizophrenia. Studies with mostly younger schizophrenia patients suggest that the newer “atypical” antipsychotics, such as clozapine, risperidone, olanzapine, and quetiapine, may be effective in treating those patients previously unresponsive to traditional neuroleptics. They also are associated with a lower risk of extrapyramidal symptoms and tardive dyskinesia (Jeste et al., in press). Moreover, the newer medications may be more effective in treating negative symptoms and may even yield partial improvement in certain neurocognitive deficits associated with this disorder (Green et al., 1997).

The foremost barriers to the widespread use of atypical antipsychotic medications in older adults are (1) the lack of large-scale studies to demonstrate the effectiveness and safety of these medications in older patients with multiple medical conditions, and (2) the higher cost of these medications relative to traditional neuroleptics (Thomas & Lewis, 1998).

Alcohol and Substance Use Disorders in Older Adults

Older people are not immune to the problems associated with improper use of alcohol and drugs, but as a rule, misuse of alcohol and prescription medications appears to be a more common problem among older adults than abuse of illicit drugs. Still, because few studies of the incidence and prevalence of substance abuse have focused on older adults—and because those few were beset by methodological problems—the popular perception may be misleading.

A persistent research problem has been that diagnostic criteria for substance abuse were developed and validated on young and middle-aged adults. For example, DSM-IV criteria include increased tolerance to the effects of the substance, which results in increased consumption over time; yet, changes in pharmacokinetics and physiology may alter drug tolerance in older adults. Decreased tolerance to alcohol among older individuals may lead to decreased consumption of alcohol with no apparent reduction in intoxication. Criteria that relate to the impact of drug use on typical tasks of young and middle adulthood, such as school and work performance or child rearing, may be largely irrelevant to older adults, who often live alone and are retired. Thus, abuse and dependence among older adults may be underestimated (Ellor & Kurz, 1982; Miller et al., 1991; King et al., 1994).


Alcohol Abuse and Dependence
The prevalence of heavy drinking (12 to 21 drinks per week) in older adults is estimated at 3 to 9 percent (Liberto et al., 1992). One-month prevalence estimates of alcohol abuse and dependence in this group are much lower, ranging from 0.9 percent (Regier et al., 1988) to 2.2 percent (Bailey et al., 1965). Alcohol abuse and dependence are approximately four times more common among men than women (1.2 percent vs. 0.3 percent) ages 65 and older (Grant et al., 1994). Although lifetime prevalence rates for alcoholism are higher for white men and women between ages 18 and 29, African American men and women have higher rates among those 65 years and older. For Hispanics, men had rates between those of whites and African Americans. Hispanic females had a much lower rate than that for whites and African Americans (Helzer et al., 1991). Longitudinal studies suggest variously that alcohol consumption decreases with age (Temple & Leino, 1989; Adams et al., 1990), remains stable (Ekerdt et al., 1989), or increases (Gordon & Kannel, 1983), but it is anticipated that alcohol abuse or dependence will increase as the baby boomers age, since that cohort has a greater history of alcohol consumption than current cohorts of older adults (Reid & Anderson, 1997).

Misuse of Prescription and Over-the-Counter Medications
Older persons use prescription drugs approximately three times as frequently as the general population (Special Committee on Aging, 1987), and the use of over-the-counter medications by this group is even more extensive (Kofoed, 1984). Annual estimated expenditures on prescription drugs by older adults in the United States are $15 billion annually, a fourfold greater per capita expenditure on medications compared with that of younger individuals (Anderson et al., 1993; Jeste & Palmer, 1998). Not surprisingly, substance abuse problems in older adults frequently may result from the misuse—that is, underuse, overuse, or erratic use—of such medications; such patterns of use may be due partly to difficulties older individuals have with following and reading prescriptions (Devor et al., 1994). In its extreme form, such misuse of drugs may become drug abuse (Ellor & Kurz, 1982; DSM-IV).

Research studies that have relied on medical records review show consistently that alcohol abuse and dependence are significantly more common than other forms of substance abuse and dependence (Finlayson & Davis, 1994; Moos et al., 1994). Yet prescription drug dependence is not uncommon and, as Finlayson and Davis (1994) found, the greatest risk factor for abuse of prescription medication was being female. This finding is supported by other studies showing that older women are more likely than men to visit physicians and to be prescribed psychoactive drugs (Cafferata et al., 1983; Baum et al., 1984; Mossey & Shapiro, 1985; Adams et al., 1990). In contrast, an analysis of data from the National Household Survey on Drug Abuse concluded that older men were more likely than women to report use of sedatives, tranquilizers, and stimulants (Robins & Clayton, 1989). Older adults of both sexes are at risk for analgesic abuse, which can culminate in various nephropathies (Elseviers & De Broe, 1998).

Benzodiazepine use represents an area of particular concern for older adults given the frequency with which these medications are prescribed at inappropriately high doses (Shorr et al., 1990) and for excessive periods of time. A national survey of approximately 3,000 community-dwelling persons found that older persons were overrepresented among the 1.6 percent who had taken benzodiazepines daily for 1 year or longer (71 percent > 50 years; 33 percent > 65 years of age) (Mellinger et al., 1984). Benzodiazepine users were more likely to be older, white, female, less educated, separated/divorced, to have experienced increased stressful life events, and to have a psychiatric diagnosis (Swartz et al., 1991).

Illicit Drug Abuse and Dependence
In contrast to alcohol and licit medications, older adults infrequently use illicit drugs. Less than 0.1 percent of older individuals in the Epidemiologic Catchment Area study met DSM-III (American Psychiatric Association, 1980) criteria for drug abuse/dependence during the previous month (Regier et al., 1988). This compared with a 1-month prevalence rate of 3.5 percent among 18- to 24-year-olds. ECA data further suggest a lifetime prevalence of illegal drug use of 1.6 percent for persons older than 65 years (Anthony & Helzer, 1991).

The development of addiction to illicit drugs after young adulthood is rare, while mortality is high (Atkinson et al., 1992). For example, over 27 percent of heroin addicts died during a 24-year period (Hser et al., 1993), and 5.6 percent of deaths associated with heroin or morphine use were among persons older than 55 (National Institute on Drug Abuse, 1992).

As is projected to occur with trends in alcohol consumption, the low prevalence of older adults’ drug use and abuse in the late 1990s may change as the baby boomers age. Annual “snapshot” data extrapolated from the National Household Survey on Drug Abuse, which has been conducted since 1971, afford a glimpse of trends. Patterson and Jeste (1999) recently compared prevalence estimates of those born during the baby boom with an older (> 35 years) non-baby-boomer cohort. The difference between baby boomers and the previous cohort translated in 1996 into an excess of approximately 1.1 million individuals using drugs. Their excess drug use, combined with their sheer numbers, means that more drug use is expected as this cohort ages, placing greater pressures on treatment programs and other resources.

Projections also suggest that the costs of alcohol and substance abuse are likely to rise in the near future. Across age ranges, drug abuse and alcohol abuse have been estimated to cost over $109.8 billion and $166.5 billion, respectively (Harwood et al., 1998). Although no studies have estimated the annual costs of alcohol and substance abuse among older adults, there is evidence that the presence of drug abuse and dependence greatly increases health care expenditures among individuals with comorbid medical disorders. For example, in a study of over 3 million Medicare patients who were hospitalized and discharged with a diagnosis of cardiovascular disease, average annual hospital charges were $17,979 for older patients with a concomitant diagnosis of drug dependence and $14,253 for those with a concomitant diagnosis of drug abuse, compared with only $11,387 for older patients with no concomitant drug disorder (Ingster & Cartwright, 1995). In addition, increased expenditures due to the presence of a drug disorder were greatest among older patients who also had a mental disorder.

A long-standing assumption holds that substance abuse declines as people age. Winick (1962) proposed one of the most popular theories to explain apparent decreases in substance abuse, particularly narcotics, with aging. His “maturing out” theory posits that factors associated with aging processes and length of abuse contribute to a decline in the number of older narcotic addicts. These factors include changes in developmental stages and morbidity and mortality associated with use of substances. Consistent with these hypotheses, substance abusers have higher mortality rates compared with age-matched nonabusers (Finney & Moos, 1991; Moos et al., 1994). However, some research contradicts the “maturing out” theory. For example, some studies show that persons who have been addicted for more than 5 years do not become abstinent as they age (Haastrup & Jepsen, 1988; Hser et al., 1993). Also, addicts approaching 50 years of age who were followed for more than 20 years remained involved in criminal activities (Hser et al., 1993). These findings emphasize the need to focus more attention on substance abuse in late life, especially in light of demographic trends.

Treatment of Substance Abuse and Dependence
The treatment of substance abuse and dependence in older adults is similar to that for other adults. Treatment involves a combination of pharmacological and psychosocial interventions, supplemented by family support and participation in self-help groups (Blazer, 1996a).

Pharmacotherapy for substance abuse and dependence in older adults has been targeted mostly at the acute management of withdrawal. When there is significant physical dependence, withdrawal from alcohol can become a life-threatening medical emergency in older adults. The detoxification of older adult patients ideally should be done in the inpatient setting because of the potential medical complications and because withdrawal symptoms in older adults can be prolonged. Benzodiazepines are often used for treatment of withdrawal symptoms. In older adults, the doses required to treat the signs and symptoms of withdrawal are usually one-half to one-third of those required for a younger adult. Short- or intermediate-acting forms usually are preferred.

Pharmacological agents for treatment of substance dependence rarely have been studied in older adults. Disulfiram use in older adults to promote abstinence is not recommended because of the potential for serious cardiovascular complications. Compounds recently proposed for use in treatment of addiction, such as flagyl, deserve further study. A rare controlled clinical trial of substance abuse treatment in older patients recently revealed naltrexone to be effective at preventing relapse with alcohol dependence (Oslin et al., 1997).

5 Alogia refers to poverty of speech, and avolition refers to lack of goal-directed behavior.

Back to Top

Home | Contents | Previous | Next