A critique of the decision analysis for clinically localized prostate cancer.

The question of when to choose surgery over watchful waiting for treatment of clinically localized prostate cancer is difficult. Recently, the Prostate Patient Outcomes Research Team (PORT) published a decision analysis that promoted watchful waiting as a reasonable alternative to invasive treatment for many men with localized prostatic carcinoma. Criticisms were leveled at the analysis itself, its structure, and especially the probabilities and utilities used in the model. We reexamine the PORT decision analysis. Structural sensitivity of the model was conducted, as were multi-way analyses incorporating data more recent than those used in the PORT analysis. The model structure is concluded to be sound and reasonable given the available literature on prostate cancer progression. However, recent data suggest that some probabilities used in the PORT analysis may be understated, causing life expectancy for surgical treatment to be understated. Using these recent data, the decision to operate on all grades of prostate cancer is strongly supported. This finding contradicts that of the original PORT analysis, which suggested that surgery would decrease quality-adjusted life expectancy for patients with well differentiated lesions. Ethical considerations aside, a clinical trial, such as the Prostate Cancer Intervention Versus Observation Trial, would help to establish reliable progression rates. Also, additional quality of life studies with actual prostatic cancer patients are needed for accurate decision making. Decision scientists and academic urologists should collaborate on model refinement and subsequent analyses.