CONTEXT

More than 31,000 men were expected to die from prostate cancer in 2001. Researchers and the public have given most attention for controlling prostate cancer to screening. No well-conducted randomized controlled trial (RCT) of screening has been completed. We are thus left with examining indirect evidence to determine the efficacy of screening for reducing prostate cancer mortality.

OBJECTIVE

To examine the evidence of the benefits and harms of screening and earlier treatment in reducing prostate cancer mortality and to assist the US Preventive Services Task Force in making recommendations on this topic.

DATA SOURCES

We first developed an analytic framework and 9 key questions that represent the logical chain between screening and reduced mortality. We then systematically searched MEDLINE from January 1994 to September 15, 2002, using the Medical Subject Heading prostate neoplasms and combining this term with predefined strategies to identify English language studies concerning the 9 key questions. We also searched the Cochrane Library, contacted experts, and scanned review bibliographies.

STUDY SELECTION

We examined abstracts and full articles of all identified studies to determine whether they met preset inclusion and exclusion criteria for each key question. We selected studies that met the following inclusion criteria: (1) randomized controlled trials (RCTs), case-control studies, and ecologic studies that examined links between screening and reduced mortality, (2) studies that addressed the accuracy, reliability, and yield of screening tests by applying the test and a reference standard uniformly to a defined population; (3) RCTs with clinical outcomes for treatment questions; (4) studies of patient reports about their experience with screening or various treatments; and (5) studies that examined or modeled the costs and benefits of screening. For key questions about treatment, we required RCTs with clinical outcomes. We graded the quality of all included articles according to criteria established by the USPSTF.

DATA EXTRACTION

Members of the study team abstracted relevant information from included studies and entered it into established abstraction forms. The first author checked all abstractions against the original papers.

DATA SYNTHESIS

No conclusive direct evidence shows that screening reduces mortality from prostate cancer. Although we could not precisely determine the sensitivity and specificity of screening with prostate-specific antigen (PSA) and digital rectal examination (DRE), research is clear that these tests can detect prostate cancer at an earlier stage than clinical detection. Because of the heterogeneity in the natural history of prostate cancer, we could not determine how many screen-detected cancers would eventually become clinically important. The efficacy of treatment for clinically localized prostate cancer detected by screening with any of the currently used approaches is unknown. Each treatment is associated with several well-documented potential harms. The cost of a national screening program is potentially large. Modeling studies show that men ages 50 to 69 years could receive benefit at reasonable cost from screening under favorable assumptions about the efficacy of earlier treatment. These studies do not adjust for the potential harms of screening. Given the current strategy for screening, men with a life expectancy of less than 10 years are unlikely to benefit even under favorable assumptions.

CONCLUSIONS

We are unable to determine the net benefit of screening because we cannot establish the presence and, if present, the magnitude of benefit from screening. We can establish the presence of potential harms.