Our team recently published an article using registry cohort to examine whether county-level prevalence of prostate-specific antigen (PSA) screening was associated with lower odds of advanced stage prostate cancer, and lower mortality among those diagnosed. We found that among those diagnosed with prostate cancer, those residing in counties with higher PSA screening experienced lower odds of advanced stage disease and lower mortality. These patterns held even after adjusting for multiple demographic and area-level socioeconomic, environmental, and health systems factors.
This was a colossal effort, involving support from research staff and administrators at several academic institutions and close collaborations with registry officials in eight states. An important strength of this project was the use of geomasked residential addresses to link environmental confounding variables, although because this involved use of sensitive personal information, we had to develop approaches to balance privacy with precision of exposure assessment. We also built on an earlier ecological study led by colleagues at the BWH Center for Surgery and Public Health, which compared county-level rates of metastatic prostate cancer and prostate cancer mortality among counties with low vs high PSA screening.
Here, because we had individual level case data and geomasked addresses, we could control for multiple sociodemographic, clinical, and environmental measures. Our exposure (county-level PSA screening) remained ecological. On one hand, this is a limitation – we don’t know if living in a county with high screening means you actually were screened, and vice versa. On the other, it may suggest that systems-level factors that influence population behaviors and use of screening may also impact downstream cancer outcomes. We are actively working in other cohorts with individual-level data, and across different health systems (see our earlier post on collaborations with the Veterans Health Administration) to confirm these findings.