Literature review process Search terms. Alzheimer’s disease, vascular or multi-infarct dementia, dementia with associated parkinsonian disorder (diffuse Lewy body disease, dementia with Lewy bodies, Parkinson’s disease with dementia), progressive supranuclear palsy, frontotemporal dementia (including Pick’s disease), and senile dementia. Additional search terms were question-specific:
- Question 1: Cholinesterase inhibitors, antioxidants, hormones, anti-inflammatory agents/drugs, cholinergic agents/drugs, nootropics (class of compounds structurally related to piracetam), metabolic enhancers, neurotrophic agents/drugs, (complementary) alternative medicines, treatment, and pharmacotherapy.
- Question 2: Hypnotics, antidepressants, anxiolytics, tranquilizers, sleep medications, selective serotonin reuptake inhibitors (SSRI), treatment and sleep, treatment and depression, treatment and anxiety, treatment and agitation, treatment and disinhibition, treatment and affective disorders, management, treatment and maintain (trunc), and treatment and discontinue (trunc).
- Questions 3 and 4: Counseling, education and caregiver, education and patient, environment, behavior (trunc) manag (trunc), behavior (trunc) and modif (trunc), advance directive, rehabilitation, and terminal care.
The key and index words for the special issue of cooperation between specialists were the following: treatment and multidisciplinary, treatment and team, treatment and community, neurologist (trunc), primary care, respite care, services, providers, transfer of care, long-term care, and family advocacy. Special issues related to economic models used the following key and index words: health care systems, health maintenance organization, preferred provider organization, medicaid, medicare, insurance, managed care, economic (trunc), and quality of care. Special issues related to determining capacity to consent used were the following: competency, decision-making, patient refusal of therapy/treatment, physician patient relations, patient rights, patient involvement, and informed consent.
Databases. For questions 1 and 2 the following databases were searched: MEDLINE, Embase, Current Contents, Psych Abstracts, and Cochrane databases. For questions 3 and 4: MEDLINE, Embase, Current Contents, Psychology Info, Cochrane, and CINAHL.
Inclusion/exclusion criteria and process. Studies selected included the following: randomized, controlled studies in all languages and other types of studies limited to English; human subjects with N greater than 20, regardless of outcome measured; and review articles published between January 1998 and November 1999. The nonpharmacologic intervention questions allowed smaller N for supporting studies to reflect the activity in the field. Members of the work group reviewed all search results and the bibliographies of the review articles to identify any articles that they thought were missing; these articles were submitted to the same inclusion/exclusion criteria. The initial search was conducted in October 1998; additional articles were added until July 2000.
Number and disposition of articles, data extraction, and classification of evidence. The search strategy identified 2,548 articles. A total of 380 met the predefined inclusion/exclusion criteria and were reviewed by at least two individuals. Selected items from each article were entered into a standardized data extraction form, and each article was assigned a class of evidence based upon a priori definitions, which determined whether or not study results were ultimately translated into Standards, Guidelines, or Practice Options.