Grant Number:	1R01CA286051-01 Interpret this number
Primary Investigator:	Lash, Timothy
Organization:	Emory University
Project Title:	Improving Measurement of Cancer Registry Completeness
Fiscal Year:	2024

PROJECT SUMMARY Surveillance for cancer incidence and mortality is crucial to cancer control and cancer research. Cancer registries are the basis for allocating public health resources as well as designing and assessing interventions to control cancer morbidity and mortality through prevention, screening, and treatment. The most fundamental aspect of any cancer registry is “completeness,” which is a measure of the proportion of true cancer cases captured by a registry within a defined catchment area and period. In concept, completeness is a simple proportion, equal to the number of cancer cases recorded in a registry divided by the true number of cancer cases occurring in the catchment area over the period. Measuring cancer registry completeness is, however, extremely challenging. A registry is meant to include all cancer cases, so the number of recorded cases is also the best available measure of the true number. To address this fundamental problem, the most common method to estimate completeness calculates the expected number of cancer registry cases and compares that expected number to the reported number. These extant measures leave out the two most common cancers (prostate and breast) as well as less common cancers and universally include data only from Black people and White people, therefore routinely misestimating completeness for registries covering populations with high proportions of Asian people, Native American people, and people of Hispanic ethnicity. We propose a new comprehensive method to address these shortcomings. Our method will allow registries to determine which populations and geographic areas are contributing to a lack of completeness and help them to identify strategies most likely to improve completeness going forward. Ours aims are: (a) Develop methods for estimating registry-specific completeness specific to county, race and ethnicity, age, sex, cancer site, and calendar year—borrowing information across categories in a principled way—and adjusting for time trends in survival, reporting lags, and spatial correlation; (b) Develop methods to decompose completeness by demographic (race/ethnicity and sex) and geographic (rural/urban) characteristics, allowing individual registries to improve future completeness by targeting specific populations or regions for which data are less complete; and (3) Adapt the completeness and decomposition methods to a user-friendly tool, featuring intuitive control over model parameters, and guided, semi-automated code execution, and curated output to assure transparency and reproducibility by cancer registries. By achieving these aims, this research project will (1) improve the rigor and reproducibility of measurement of completeness for US cancer registries, (2) guide choices of where registries can invest resources to optimize overall and group-specific completeness, (3) demonstrate how differential lack of completeness has biased the current understanding of cancer disparities, and (4) provide transparent and replicable computational resources linked with easily updated data sources.





None