ASSESSMENT OF STRATEGIES FOR PROSTATE CANCER SCREENING

点击翻译按钮获取中文摘要

The objective of the work proposed here is to evaluate alternative strategies for screening for carcinoma of the prostate. Due to the increasing fife expectancy of males, this disease has become a major health problem in the U.S., and promises to become more important in the future. Despite some medical progress, however, the disease-specific mortality rate associated with prostate cancer has not declined over the last 30 years. One explanation is that a large percentage of cancers is incurable at diagnosis. Thus, a potential approach to reducing mortality is to improve early detection by routine screening, a strategy that has proven effective for breast cancer though not for lung cancer. The costs of routine screening, however, are enormous and the benefits uncertain. Moreover, although properly controlled, randomized screening trials may be initiated soon, meaningful results will not be available for another 10-15 years. It is therefore imperative that the most appropriate screening policy be identified for the current management of prostate cancer. To develop appropriate policies, we propose to use the methodology of decision analysis to evaluate the costs and benefits of different alternatives. Our overall goal is to develop a general health care policy in the form of guidelines for primary care physicians and individual patients with different co-morbid conditions. A specific focus of our work will be to determine the effectiveness of screening using limited resources as especially applied to minority groups who have high mortality rates from prostate cancer and reduced access to preventive health care. An important issue raised by our analysis is the need for better information concerning the natural history of untreated prostate cancer. Thus, a large part of our work consists of using data from many different sources to estimate critical parameters of our model. These include statistics on incidence, mortality, and treatment, as well as data from studies that have specifically sought to track the outcome of untreated patients. We also make extensive use of sensitivity analysis to determine possible effects of changes in model assumptions and parameter values.