2015 Release of Delay Adjustment Factors and Rates
Timely and accurate calculation of cancer incidence rates is hampered by reporting delay. Reporting delay is the time elapsed before a diagnosed cancer case is reported to the cancer registries. For national reporting, cases are first submitted about two years after the end of a diagnosis year (e.g., a complete 2012 year of diagnosis was first submitted in late 2014). In subsequent submissions, the data for that diagnosis year are updated to reflect: 1) new cases found to have been diagnosed within that diagnosis year; and 2) new information that has been received about previously submitted cases.Modeling reporting delay is used to adjust the current case count to account for anticipated future corrections to the data.These adjusted counts are needed to produce cancer incidence trends that are not impacted by late reporting.The adjustment for undercounts is largest in the most recent diagnosis year reported and diminishes each successive reporting year.But most recent data points are considered the most important, because the current data is used for cancer control and planning activities and is interpreted as a potential harbinger of future trends.
Previously, delay adjustment was available only for a subset of the SEER registries that had been established in the 1970’s through the early 1990’s (i.e. the SEER9 and SEER13 registry groups).However, as registries throughout North America continued to mature, a joint effort by NCI, CDC, and the North American Association of Central Cancer Registries (NAACCR) was mounted to develop a unified approach for estimating and reporting delay-adjusted rates across all of the U.S. and Canada.This new approach estimates delay factors for each registry, and then combines the factors from the selected registries to derive delay adjusted rates.For more details on this unified approach see http://surveillance.cancer.gov/delay/. These new delay adjustment factors and rates are available starting with the 2015 release of data (using the December 2014 NAACCR submission).
The NAACCR-based factors are stratified by Cancer Site, Registry, Age Group, Sex, Race, and Year of Diagnosis. The factors are linked to the appropriate cases (based on the stratifications above) in data submissions for each of the three partners in this joint effort (SEER, NAACCR, and NPCR).As of 2015, each of these groups is able to produce delay-adjusted rates using SEER*Stat.
While the eventual goal is to release delay adjustment factors and rates for individual registries, the goal in this first year of release is to produce delay adjusted rates for groups of registries (which are more stable than individual registry estimates). In 2015, NAACCR will release delay adjusted estimates for the US and Canada, while SEER will release delay adjusted rates for the SEER9, SEER13, and SEER 18 registry groups.
State-Level Delay Adjustment Factors
We encourage states to evaluate their state-specific delay adjustment factors and rates. To obtain access to the delay adjustment data for your state in SEER*Stat, please contact email@example.com.
NAACCR Delay Model Exclusions
The delay model was not able to be conducted on every Registry. States excluded from the model were notified in May 2015. Registries were excluded from the delay modeling for one of four reasons.
- Decline to participation: A registry chose not to participate in delay modeling;
- 75% Rule: Because of missing cells for combinations of submission and diagnosis years (either because they were not submitted or deemed not “fit for use”).For diagnosis years 1997-2012 with submissions from 1999-2014, a registry could have up to 16 1stsubmissions of data, up to 15 2ndsubmissions of data,… , and up to 4 13th submissions data. The proportion of cells with data was computed for the 1st through 13th submissions. If this calculated average was below 75%, the registry was excluded from consideration.
- Unexpected results: Some models for a specific registry, race group, or cancer site were difficult to model across submissions and produced results that did not accurately reflect the underlying data.This does not mean that the data were incorrect, but simply that the current models were not sufficiently flexible to fit some unusual pattern.
- Lacking 2012 diagnosis year data: Some registries were included in the delay modeling, but do not have 2012 data included in their Fall 2014 submission. Therefore, delay adjusted rates for 2012 cannot be computed.
Plenary presentation on this project was presented at the NAACCR Annual Meeting in Charlotte on Thursday, June 18, 2015 at 10:00 am.
In July 2015, each state included in the model received a state-level assessment of their delay adjustment estimates.
On July 23, 2015 NAACCR hosted a webinar that will outline how to produce rates and trends using delay adjusted data
Additional information about the development of the Delay Model and the Methodology, included reference articles, is available on the SEER Website: