Modeling Reporting Delay for Cancer Registries

Central cancer registries work hard to ensure every new cancer diagnosis is counted so that incidence data can be used for cancer control and planning and research. NAACCR Gold Certification requires that registries report 95% of cases within 23 months; and the vast majority of central registries in North America meet that benchmark

Cancer registry data are continually updated as new cases are reported and existing records are updated, reflecting the complexity of healthcare reporting and data collection. Reporting delays happen when new cancers are reported to the central registry years after diagnosis, rather than within the national submission period for the year of diagnosis..  As a result, recent data submissions are more incomplete than older data, and failing to adjust for reporting delays can bias incidence estimates and lead to misleading conclusions about cancer trends. To address this, composite delay factors are modeled for each registry to adjust recent data for expected under-counts, ensuring more accurate trends for cancer control and planning.

Late last year, Drs. Huann-Sheng Chen, Eric J. (Rocky) Feuer, and colleagues¹ published a comprehensive description of the delay methodology used to produce cancer incidence trends that are not impacted by late reporting. This peer reviewed article provides a comprehensive look at the development of the delay model, initially for SEER only registries and now applied to all NAACCR member registries.

The paper outlines the evolution of cancer registration in the U.S. and Canada and explains how delay modeling has adapted to meet changing data needs. It also highlights the critical role of NAACCR data in supporting these models, and the authors express strong appreciation for NAACCR and central cancer registries for their role in enabling and promoting delay-adjusted cancer incidence analyses. The paper’s Discussion section offers a thoughtful examination of the strengths and limitations of this methodology, helping readers better understand and appropriately interpret cancer incidence trends. While delay-adjusted rates are critical for accurate trend analysis, the model relies on past patterns and cannot correct for one‑time operational disruptions. This makes it important to use additional indicators to distinguish real changes in cancer incidence from temporary reporting delays.

Registries can access their delay factors by submitting a registry-specific data request via DaRT.

More information about delay adjustment can be found here.

For additional questions about delay adjustment methodology and use contact Dr. Chen, and for questions about release via DaRT contact Dr. Recinda Sherman

Link to free pdf of the article here.

[1] Chen, Huann-Sheng, Douglas Midthune, Zhaohui Zou, Daniel Miller, and Eric J. Feuer. “Modeling Reporting Delay in Cancer Incidence Counts in the Evolving US and Canadian Population-Based Cancer Registry Environment.” Cancer Epidemiology, Biomarkers & Prevention 34, no. 10 (2025): 1710-1721