The New Surveillance, Epidemiology, and End Results Prostate with Watchful Waiting Database: Opportunities and Limitations

Background: Active surveillance (AS)/watchful waiting (WW) strategy for localized prostate cancer (PCa) is increasingly and broadly endorsed as a preferred option for initial treatment of men with very low- and low-risk PCa, but outcomes can be difficult to analyze in traditional, population-based registries. The recently released Surveillance, Epidemiology, and End Results (SEER) Prostate with WW dataset provides an opportunity to understand national patterns and trends in AS/WW, but the data source itself has not been well described.

Objective: To provide a comprehensive description of this dataset and investigate possible biases due to missing data.

Design, setting, and participants: The SEER is a population-based epidemiologic registry in the USA. Newly diagnosed PCa patient data were collected from 18 SEER registries between 2010 and 2015, with inclusion of a new treatment variable for AS/WW. We identified 316 724 patients in the entire cohort and 257 060 men with clinically localized PCa (T1-2N0M0).

Intervention: Various primary treatments for PCa.

Outcome measurements and statistical analysis: The degree of missing data for each variable was measured. In order to investigate possible bias due to missing data for cancer characterization, we compared two versions of the data: one that excluded cases with missing data and one dataset generated applying multiple imputations.

Results and limitations: Only 46% of cases had complete data on basic cancer characteristics for risk stratification. The excluded dataset (N=118 821) differed significantly from the multiple imputation dataset (N=257 060) in the distribution of every reported variable (all p<0.001). The dataset does not distinguish WW from AS, which is a limitation.

Conclusions: While the SEER Prostate with WW dataset offers a new method to describe treatment trends for men with PCa, including the use of AS/WW, the amount of missing data should not be ignored.

Patient summary: While the Surveillance, Epidemiology, and End Results Prostate with Watchful Waiting dataset offers a new method to describe treatment trends for men with prostate cancer, including the use of active surveillance, it has a significant amount of missing data, which can be a source of potential bias if not addressed properly.