NAACCR Program Manager of Data Use and Research, Recinda Sherman, provides updates for CiNA data products, Call for Data, and more.
I am looking forward to seeing many of you at our Annual Conference in New Orleans in June. I am also eager to eat a big ole muffuletta from Central Grocery & Deli. But in between the eating and the catching up, there will be some excellent presentations.
For instance, RDU has two workshops on Wednesday, June 21st:
NAACCR Data Confidentiality and Data Release Recommendations
This workshop will present an early look at the draft NAACCR Data Release Guidelines, under development by the Data Security & Confidentiality Workgroup. This workshop will be a targeted discussion forum on three topics related to data release: issues related to patient contact studies, release of restricted data or cases, and suppression guidelines.
Numerators, Denominators, and ABSMs – Critical Issues With Calculating Cancer Statistics
This panel discussion will provide an overview of some of the work of the Population and Area-Based Social Measures (ABSM) Workgroup and describe ongoing issues, including population resources available, potential impact of differential privacy on the 2020 Census data, non-bridged race rate calculations, and changes to the definition of rurality by the census.
Overall, CiNA data showed about a 10% drop in expected cases for diagnosis year 2020, a number that is in alignment with NPCR and SEER data, including pathology reports. As mentioned in our last RDU Update, we will not use 2020 data in trend analysis.
Completeness Adjustment for 2020
Historically, the NAACCR uses a five-year Incidence to Mortality Ratio (I/M) as the comparator to estimate expected rates to evaluate the current submission year by completeness. However, as expected, this method did not appropriately capture completeness of case ascertainment for the unusual 2020 diagnosis year.
To address this issue, a NAACCR Workgroup was formed that included NAACCR, NPCR, and SEER representatives. The workgroup evaluated potential adjustments with two goals in mind:
- Do not penalize registries for decreased numbers of cancer cases diagnosed due to changes in medical access; and
- Do not adjust away operational issues (due to COVID-19 or other factors) that resulted in decreased number of cancer cases collected.
To this end, the workgroup reviewed surveys of the central cancer registries on case ascertainment, evaluated multiple years of available pathology data, and assessed multiple adjustments to the standard method using submitted data. Based on these evaluations, the workgroup unanimously identified a standard method of adjustment to the 2020 submission that will be used by both NAACCR and NPCR.
NAACCR, along with NPCR, applied a “self-correcting” adjustment to the Completeness Estimate for Certification of the 2020 data. The goal was to adjust the expected case count to account for the fewer cases identified due to changes in medical care due to COVID-19, but not to adjust for registry operational issues that impacted ascertainment of cases. The adjustment was made to the underlying completeness calculation by using a single year, 2020, instead of five years for both the incidence and the mortality components of the comparison I/M Ratio.
This approach was applied to all US and Canadian registries and resulted in completeness measures that aligned with historical performance as well as our understanding of current challenges within the registries.
It is important to clarify that this adjustment is for the 2020 diagnosis year only. The 2021 data will require separate evaluation for any potential adjustment. This year, NAACCR is leading the production of the Annual Report to the Nation on the Status of Cancer with a focus on the 2020 rates as the special topic.
NAACCR Call for Data
This was a difficult year for many registries. This was reflected in the December 2022 submissions, which had a number of unexpected data quality issues that required additional attention to resolve both during and post-submission.
After applying an adjustment to the Completeness Calculation, 3 US registries and 2 Canadian registries submitted data but did not achieve any level of NAACCR Certification for the 2020 diagnosis year data. For Canada, 10 registries achieved Gold, no registries were Silver, and 2 registries did not submit data. For the US, 49 registries achieved Gold, 6 registries achieved Silver, and all US registries submitted data.
Certification results and associated data assessments, including the Tableau visualizations, were provided to registries in April through the Call for Data Portal. Registries will receive their certificates and public recognition doing our annual meeting—in person!
Tumor-level Deduplication With Match*Pro for December 2023 CFD
NAACCR provided training on the new tumor-level deduplication protocol using Match*Pro in March. We solicited feedback from registries on the protocol and used their input to finalize a graduated approach for incorporating tumor level deduplication into the NAACCR CFD submission requirements. Registry feedback was used to inform our approach to duplication for Certification. The NAACCR Board voted and approved the following roll-out approach for deduplication for NAACCR Certification using Match*Pro.
CFD December 2023
- 100% Patient-level deduplication resolution on full submission (1995-2022)
- Gold = 5 years of tumor-level resolution (2017-2021); <.1% tumor-level duplicates (match run on all cases diagnosed 2007-2021)
- Silver = 1 year of tumor-level resolution (2021); <.1% duplicates (match run on all cases diagnosed 2007-2021)
CFD December 2024
- 100% Patient-level deduplication resolution on full submission (1995-2023)
- Gold = 2007+ tumor-level resolution; <.1% tumor-level duplicates (match run on all cases diagnosed 2007-2022)
- Silver = 5 years of tumor-level resolution (2018-2022); <.1% tumor-level duplicates (match run on all cases diagnosed 2007-2022)
CFD December 2025
- 100% Patient-level deduplication resolution on full submission (1995-2024)
- Gold = 2007+ tumor-level resolution; <.1% duplicates (match run on all cases diagnosed 2007-2023)
- Silver = 2007+ tumor-level resolution; <.2% duplicates (match run on all cases diagnosed 2007-2023)
At this time there is no intention of resolving tumor-level duplicates on cases prior to 2007. Patient-level deduplication will be assessed based on the honor system, and we will collect information in the CFD Portal similar to last year. However, tumor-level deduplication will be assessed on your submission file. We are currently working on the procedure for registries to identify cases that were reviewed and determined to not be duplicate tumors.
NAACCR Special Edition of JRM
This edition of the Narrative has an article describing the contents of our 3rd Annual NAACCR JRM collaboration with NCRA.
I am pleased to announce that articles from our edition won both “Best Paper” and “Honorable Mention” Awards. This year, article submissions are due October 3rd, 2023. We encourage registries to submit short reports or editorials, in addition to research articles.
As always, if you have any questions, concerns, revolutionary ideas, or new projects that might help us get an Ig Noble Prize in the NAACCR Research and Data Use area, we’d love to hear it.
What to Read Next
CAP PERT Update & Cancer Data Summit A CAP Pathology Electronic Reporting (PERT) meeting was held in Chicago on Thursday,…
S&RD Steering Committee Explores Minimum Data Set for Early Incidence Reporting In 2022, the NAACCR Board of Directors identified a…