The conceptualization of an international consensus conference for the purpose of beginning to establish an operational definition for “behavioral disturbances of dementia” was born at the 7th International Psychogeriatric Association (IPA) Congress in Sydney, Australia, in November 1995. At that time, a special symposium on “research methodological issues in evaluating behavioral disorders of dementia” took place. From the high level of interest and attendance, it was clear that this matter was a source of great attention and concern by clinicians and researchers working with demented elderly. Although we have been aware that patients with dementias, including Alzheimer's disease (AD), manifest psychotic, depressive, and behavioral symptoms over the course of the illness, such symptoms until recently have elicited little interest or research support. Further, research was hampered by the absence of viable measurements and scales developed for this specific population. In addition, rigorous clinical trials were rare and virtually nonexistent in those with moderate and advanced dementias. Prior to 1992, only seven randomly assigned, double-blind trials in this population had been published, only one in a nursing home setting.

As the population ages in the United States and other countries, we will be confronted with an increasing number of patients with neurodegenerative disorders, such as Alzheimer's disease and Parkinson's disease. The progressive cognitive impairment that characterizes Alzheimer's disease leaves patients increasingly unable to care for themselves. Patients with Alzheimer's disease also are prone to dementia-related behavioral disturbances, which can range from annoying (such as repetitive questioning) to disturbing (such as vocal outbursts) to dangerous (such as hitting and pushing).

Disturbances of mood, thought content, perception, and behavior are common concomitants of dementia that often produce suffering and excess disability. These disturbances also interfere with caregiving, precipitate institutionalization, and hasten death. The current lack of an effective means of preventing or controlling the pathophysiologic events that lead to dementia in late life has stimulated efforts to understand and treat these phenomena. In particular, clinically significant depression and psychosis often emerge in patients with dementia, regardless of its origins, and are an important focus of treatment.

Behavioral disturbances of dementia result from both biologic and environmental factors. This article focuses on the biologic causes of behavioral problems in patients with dementia.

An impressive body of literature has been published on the relationship between psychotic symptoms in Alzheimer's disease and pathology demonstrated primarily by neuroimaging and biochemistry studies. Jacoby and Levy in 1980 and Burns and colleagues in 1990, for example, reported less severe atrophy in delusional patients with Alzheimer's disease than in nondelusional patients with Alzheimer's disease. The author and colleagues have found that smaller ventricle-brain ratios are associated with delusions of theft in Alzheimer's disease. Zubenko (1991) and Doty (1989) have reported that delusions and hallucinations in patients with Alzheimer's disease are associated with decreased amounts of serotonin.

Neuroimaging studies have contributed greatly to our understanding of Alzheimer's disease and other dementias. Computed tomography and magnetic resonance imaging reveal brain structure and aid in the diagnostic evaluation of patients with cognitive impairment. Functional neuroimaging studies use positron emission tomography, single-photon emission computed tomography, and other methods to measure regional cerebral activity, including metabolic rate, blood flow, and neuroreceptor density. Functional neuroimaging results can be useful clinically and have also been used in a variety of research applications to examine physiologic variables in neuropsychiatric illnesses.

Age-related changes in circadian rhythm (e.g., fragmented sleep-wake patterns) occur in many older persons but are particularly pronounced in patients with Alzheimer's disease. In these patients, disruptions of circadian rhythms can be severe enough to increase mental decline, agitation during the day, and restlessness at night. Moreover, patients whose nocturnal restlessness disrupts the sleep of the caregiver are more likely to be institutionalized than those who have cognitive impairment alone.

Although the essence of one's identity—one's self-esteem—is eroded and devastated by Alzheimer's disease, little attention has been paid to the regression and dissolution of the self experienced by patients with this disease. Investigations into the psychology of the self by psychoanalyst Heinz Kohut and others have provided new ways of understanding a demented patient's attempts to maintain some semblance of self-esteem and identity in the wake of progressive cognitive decline.

Are the behavioral disturbances of dementia a primary manifestation of a devastating degenerative brain disease or a consequence of a demented patient's inability to cope with a hostile environment that does not adapt to his or her needs? Our current state of knowledge regarding the origin of noncognitive symptoms in dementia does not allow us to neatly dichotomize behavioral disturbances into either syndromes that can be attributed to altered neurobiologic substrates or behavioral reactions to a change in social routine or environment. The best fit model is one that incorporates biologic (neurochemical, neuropathologic, or genetic); psychological (premorbid personality and behavior, and response to stress); and social (environmental change and caregiver factors) aspects. For any particular symptom or group of symptoms, the relative input from each causal source can vary and may change in the future as our knowledge and understanding grow. The importance of an interactive causal model is that is has direct and immediate relevance to the development of treatment strategies.

Three issues concerning behavioral problems of dementia emerge from research on family caregivers. The first issue concerns the reliability of caregivers' reports of behavioral disturbances. The second issue is the realization that caregivers can provide valuable information not available from other sources. The final issue concerns how caregivers' reports can clarify the determinants behavioral problems.

The episodic and nonuniversal nature of behavioral disturbances of dementia challenges the notion that these symptoms result only from permanent changes in the brain. Few longitudinal data are available to correlate the occurrence of behavioral symptoms with the neurobiologic model of dementia, and most of these data are from clinical trials of institutionalized patients. Nevertheless, Dr. Cummings noted that even though behavioral symptoms are highly episodic, they tend to recur in the same subgroups of patients. The neurobiologic model, therefore, may help determine the vulnerability of a patient to behavioral disturbances but not the consistent expression of the behavioral symptom complex.

Establishing a medical diagnosis serves two utilitarian purposes: providing information necessary to initiate treatment and communicating information regarding prognosis. A nosology or diagnostic nomenclature (i.e., a classification of diagnoses) provides further utility by establishing a foundation for clinical research. In his book, Wulff outlined four types of diagnoses: (1) symptomatic or pseudoanatomic diagnoses (e.g., chronic headache, persistent diarrhea, or irritable bowel); (2) syndromes; (3) anatomic diagnoses; and (4) causal diagnoses. By definition, syndromes have no means of being validated by measures external to the constructs themselves. Often, specific syndromes reflect diverse origins, and conversely, specific etiologies may cause multiple syndromes (e.g., syphilis, human immunodeficiency virus, and diabetes).

Among the challenges clinicians and researchers face regarding behavioral disturbances of dementia are how to conceptualize them and how to measure them. The Diagnostic and Statistical Manual of Mental Disorders (4th ed.) recognizes that dementia can be associated with mood symptoms and delusions by formally coding for them but does not have codes for coexisting hallucinations or other specific behavioral symptoms associated with dementia. One reason for this lack of coding classification is the difficulty in conceptualizing the many behavioral disorders, symptoms, and problems seen in patients with dementia.

Science depends on the ability to measure natural phenomena. To increase our knowledge, accurate definitions and measurements of these phenomena are essential (Thompson, 1989). Over the past 20 years, many researchers have concentrated on learning how to define and measure psychiatric phenomena. The results of such research allow us to develop diagnostic criteria for various psychiatric disorders.

Quality-of-life issues affect both the patient and the caregiver. Although several measures of caregiver burden and stress, which reflect quality of life, are available, measuring the quality of life of the patients themselves is difficult.

Clinical scientists developing rating scales to assess the behavior of patients with dementia have adopted a variety of evaluation strategies. Scales differ according to the source of information (e.g., caregiver versus patient), type of behavior assessed (e.g., mood, agitation, or delusions), origin of the scale (i.e., imported from psychiatry, adapted from psychiatric scales, adapted from scales for neurologic conditions, or developed specifically for dementia), and anticipated application of the tool (e.g., behavioral characterization, longitudinal follow-up, or differential diagnosis). Investigators have rarely articulated the theoretic framework on which their scales are based, and in most cases, theories were eschewed in favor of empirically based assessments of observed behaviors. Theoretic assumptions, however, can be inferred from the structure of the scales.

Before the development of the Behavioral Pathology in Alzheimer's Disease (BEHAVE-AD) rating scale in 1987 by Reisberg and colleagues and its predecessor scale, the Symptoms of Psychosis in Alzheimer's Disease (SPAD) rating scale, in 1985 by Reisberg and Ferris, other scales were available for measuring behavioral disturbances and psychiatric disorders in patients with Alzheimer's disease. However, these scales generally mixed together cognitive disturbances with behavioral symptoms and sometimes included functional impairments as well. These predecessor scales also were not specifically designed to assess the types of behavioral problems seen in Alzheimer's disease. If a scale did address behavioral disturbances of dementia, it tended to be seriously underspecified in terms of the nature of behavioral disturbances.

Many rating scales are available to assess the behavior of patients with dementia. Some of these scales have a broad focus and allow clinicians to rate various domains of function, either directly through patient observation or indirectly through caregiver interviews. Other scales are more specific, such as the Cohen-Mansfield Agitation Inventory (CMAI; Cohen-Mansfield, Marx, & Rosenthal, 1989) and the Agitation Behavior Mapping Instrument (ABMI; Cohen-Mansfield, Werner, & Marx, 1989), which focus exclusively on agitated behaviors. As defined by Cohen-Mansfield and Billig (1986), agitation is inappropriate verbal, vocal, or motor activity that is not judged by an outside observer to result directly from the needs or confusion of the agitated person. Data gathered by using the CMAI and the ABMI have provided valuable insight into the subtypes of agitation, correlates of agitated behaviors, and the relationships between subtypes of agitation and cognitive functioning.

In the absence of a standardized technique for reliably and comprehensively describing changes in behavioral disturbances of dementia, the Behavioral Pathology Committee of the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) sought to develop a scale that could be used to evaluate a wide range of psychopathologic signs and symptoms in patients with differing severity of dementia. The goal of the committee was to develop a scale composed of well-anchored, homogeneously scaled items that could be administered by interviewers without extensive psychiatric training.

Although Alzheimer's original description of the dementing disorder that bears his name emphasized the prominence of troublesome and disruptive behaviors, a systematic investigation of behavioral disturbances of dementia did not begin in earnest until the 1980s. At that time, as the neuropathologic identity of presenile Alzheimer's disease and late-onset “senile dementia” was recognized, the redefinition of Alzheimer's disease abruptly increased the number of patients diagnosed with this condition. Physicians and other medical personnel working with Alzheimer's disease patients recognized both the importance of abnormal behaviors in this now large patient population and the need to describe, classify, and quantify these behaviors.