This technical review is based on a systematic review of the available scientific literature, beginning with a comprehensive examination of literature published from 1966 through 1994. Literature review was accomplished by (1) establishing a prioricriteria for relevant studies, (2) developing and completing abstract review forms for each study identified by computer searches and bibliography scans, (3) compiling and reviewing full articles, (4) developing and completing a data collection form for each article selected for inclusion in evidence tables, and (5) compiling evidence tables summarizing these articles.

Computer literature searches of Medline from 1966 to 1994 and Cancerlit from 1980 to 1994 were performed. Supplemental Medline searches were also performed. Approximately 3,500 citations were reviewed to identify the literature appropriate for evaluation. An abstract review form (Exhibit 1) was completed for each article; 350 articles were selected for further review. A 16-page data collection form (Exhibit 2) was used to summarize these articles and included data such as number of study participants, research question(s) addressed, study design, outcomes studied, and study results.

Medical subject heading (MeSH) terms used to search Medline and Cancerlit included the following:

• colonic neoplasm
• rectal neoplasm
• colorectal neoplasm
• sigmoidoscopy
• colonoscopy
• barium
• genetic screening
• hereditary disease
• occult blood

Data from the core articles were used to develop a set of evidence tables on colorectal cancer natural history and epidemiology, risk factors, screening test performance characteristics, outcomes associated with screening and surveillance, test cost and cost-effectiveness, and patient and provider participation.

An appropriate outcome for studies of colorectal cancer screening is a reduction in colorectal cancer mortality, and an appropriate intermediate outcome is a reduction in the incidence of colorectal cancer. Other potential outcomes include effect on patient quality of life and colorectal cancer-related morbidity, such as effects associated with surgery and adjuvant chemotherapy. Data were sought that addressed these specific outcomes.