Because Phase I does not involve a control or comparison group, Phase I studies were not relevant to the meta-analyses and thus were omitted from Attachment C, which presents a comparison of mental and functional status tests according to Phases II, III, and IV.

Healthy control subjects consisted of community volunteers or sometimes nonaffected spouses. Control subjects with comorbidity could include patients from primary care clinics, psychiatric clinic outpatients, memory or dementia clinic outpatients, or medical or surgical patients. Characteristics of case subjects were as follows: narrow spectrum of disease included moderate to severe Alzheimer's disease; expanded spectrum included mild or mixed Alzheimer's disease; and comorbidity among case subjects was most often depression, stroke, delirium, or other conditions that affected cognitive or functional performance.

Despite the voluminous literature on the subject, very few studies have been conducted on the most informative phases (i.e., IV or V). Thus, for Phase IV studies, only one instrument, the MMSE, had more than four contrasts; the remainder of the instruments had fewer than four (most had only two contrasts). Consequently, two studies, each with a single contrast at Phase IV, represent extremely important if limited information.

Instruments for assessment of functional impairment were also subjected to meta-analysis. The 18 key instruments evaluated in the meta-analyses are listed in Attachment B. The results of the meta-analyses for selecting specific assessment instruments for mental and functional status are described in Chapter 4 (see also Attachment C).

Table 1. Five-phase scheme for evaluation of mental status tests

Source: Nierenberg AA, Feinstein AR. How to evaluate diagnostic marker tests. Lessons from the rise and fall of dexamethasone suppression test. JAMA 1988;259(11):1699-1702. Copyright 1988, American Medical Association. Used with permission.

Panel Recommendations and Strength of Evidence

After extensive review and analysis of the available scientific literature, the panel developed guideline recommendations supported by the evidence. Strength of evidence was rated based on the following criteria:

A Strong Evidence. Evidence from studies that compare case patients who have dementia with control subjects who do not have dementia but do have comorbid or interfering conditions (most difficult level of discrimination).

B Suggestive Evidence. The same type of evidence as in category A, but involving a smaller number of studies or a less consistent pattern of findings, or both.

C Expert Opinion. Evidence from clinical experience described in the literature or derived from the consensus of panel members, or both.

The panel's recommendations, strength-of-evidence ratings, and discussion of supporting evidence appear in Chapter 4.

Public Comment and Peer Review

A public meeting was announced in the Federal Register and held April 12, 1993, in Washington, DC. The purpose of this meeting was to give interested organizations, individuals, and agencies an opportunity to present written or oral testimony for the panel's consideration. The meeting elicited oral and additional written testimony from eight interested parties.

Later in the process of guideline development, drafts of the Clinical Practice Guideline were distributed for peer review. A second peer review was conducted for a subsequent draft that incorporated major revisions based on earlier review comments. Reviewers were selected from the following groups to represent a broad range of disciplines and clinical practice areas: health care professional and consumer organizations, participants in the public meeting, government agencies, and interested health care providers. A total of 109 reviewers submitted comments, which were collated and reviewed by the panel cochairs.

Recommendations for Further Research

In reviewing and analyzing relevant literature, the panel identified several areas related to the initial recognition of dementia that have not been investigated or have not been studied adequately. Areas identified for future research include the following:

The relative costs of a simple approach to initial assessment of dementia versus much more expensive diagnostic procedures.

The nature of depressive symptoms observed in early dementia, including evaluation of the effects of other coexisting chronic diseases. Several studies have noted an apparent decrease in the prevalence of depressive symptoms with increasing dementia, but questions remain about whether this reflects problems with current assessment methods or a real interaction between affective pathology and the severity of dementia.

A prospective study to obtain valid estimates of the economic burden resulting from Alzheimer's disease and related dementias, including both direct and indirect costs.

The incremental validity and reliability of functional status questionnaires that assess higher social and instrumental activities in the initial assessment of persons with dementia.

Followup on suggestive new research, for example, on genetic aspects of dementia and written language as an early diagnostic clue to later development of Alzheimer's disease.

Clinical implications of advances in early diagnosis of Alzheimer's disease and other dementias.

Coding and standardization of database information to permit comparative studies.

Caregivers in the community and support for patients, family members, and caregivers in dealing with behavioral and other problems.

Health systems research, for example, health maintenance organizations and other systems, as more Americans become covered by alternatives to fee-for-service-based insurance systems.

Legal and ethical considerations and protections related to dementia, especially if definitive early diagnosis becomes possible years before symptoms manifest.

In addition, the panel noted that other related research questions and topics under investigation will advance diagnosis and management of persons with dementia. Genetic screening for persons who may be at high risk for Alzheimer's disease, such as those with the apolipoprotein E 4 gene locus on chromosome 19, is a highly debated issue, particularly because currently no demonstrated intervention can prevent or delay the dementing process, and no evidence exists about predicting age of onset for persons with this genotype. Efforts to develop biological markers for the presence of Alzheimer's disease, such as tests that could be performed on samples of blood or cerebrospinal fluid, are important research topics for confirming a suspected diagnosis of Alzheimer's disease.

Public and Professional Education

Both the public and health care providers need education about early recognition of cognitive problems among older persons. Although the National Institute on Aging and the Alzheimer's Association are conducting educational efforts for both these groups (see Attachment D), additional education is needed. For patients and their families and caregivers, such information is vital in the effort to distinguish reversible causes of dementia from normal aging and other cognitive problems and to obtain suitable care. Education is also relevant for a wide variety of health care professionals and other service providers who may be in a position to notice the onset of memory and cognitive problems in older persons they see frequently. In an urban setting, this might be a bus driver; in a rural community, it might be a grain dealer. Bank tellers, pharmacists, grocery store clerks, neighbors—all may become aware of cognitive changes in regular customers and friends.

In all cases, the purpose of educational activities is to increase the likelihood of early recognition and assessment of a potential dementing illness so that (a) concern can be eliminated if it is not warranted, (b) treatable conditions can be identified and addressed appropriately, and (c) nonreversible conditions can be diagnosed early enough to permit the family to plan for contingencies such as long-term care, and for the patient to participate as fully as possible in those decisions. For example, financial planning, supportive services, assistive devices, respite care, and use of appropriate service providers should be considered.

Educational programs related to dementia for the public and for health care providers should emphasize that symptoms of dementia are not a concomitant of normal aging and should not be ignored because of the mistaken belief that cognitive decline is an inevitable consequence of aging. Although changes in memory and cognition may accompany normal aging, significant impairment and disability are not a part of normal aging. In addition, the degree of concern felt by a person about memory problems and the severity of actual memory problems demonstrated during testing are not necessarily directly related (Cromwell, 1992).

Another important goal of educational programs related to dementia should be to clarify the difference between initial assessment and final diagnosis. A diagnosis or label such as Alzheimer's disease should not be applied prematurely.

Organization of the Guideline

This first chapter has described the purpose and scope of the guideline and the methods used by the expert panel to develop it. Chapter 2 includes background information on Alzheimer's disease and other dementias, such as major characteristics, prevalence and incidence, and costs associated with the condition. Chapter 3 addresses risk factors and causes of dementia. Chapter 4

presents specific guideline recommendations and summarizes supporting evidence. It includes a clinical algorithm for recognition and initial assessment and a table of symptoms that suggest possible dementia. Chapter 4 also contains information about delirium and depression, which can account for or mask symptoms of dementia. The attachments contain information about specific mental and functional status tests and list additional resources that will assist health care professionals and caregivers.

Two key features of this guideline are designed as clinical aids in the early recognition of dementia. The list of symptoms of possible dementia (Table 6) helps clinicians and others identify patient behaviors that indicate the need for an initial clinical assessment. The flow chart (Figure) provides an efficient, step-by-step guide to the initial recognition and assessment of patients, as well as interpretation of initial test results.