The November 2022 SEER data submission includes new cancer cases diagnosed in 2020, the first year of the COVID-19 pandemic. In 2020, the highly publicized impact of COVID on health services and the consequential delays and reductions in cancer screening and diagnosis led to a decline in the 2020 incidence rates for most cancer sites.
Because 2020 was a temporary, anomalous year caused by the pandemic, it can bias estimates such as cancer incidence trends that are of substantive interest. To address the bias, SEER made some changes in the reporting of cancer statistics this year.
- The 2020 incidence data is included all SEER incidence databases and in most cancer statistics in SEER*Explorer, Stat Facts, and other web-based reports.
- The 2020 incidence data is excluded from two cancer statistics: estimation of Joinpoint trends and risks of developing cancer (DevCan).
The 2020 incidence cases are available in the SEER Research Data for analysis through SEER*Stat.
Exclusion of 2020 Incidence Data from Joinpoint Trends
The joinpoint regression model for the analysis of trends was not designed to accommodate a one-year anomaly in data nor to estimate a change in one single calendar year. Inclusion of the 2020 incidence can influence the location of joinpoints, the value of the trend measure (annual percent change), and provide a poor fit of the model and larger confidence intervals, challenging the interpretation of estimate’s association with risk factors and efforts in prevention and early detection.
The joinpoint models used by SEER to report trends were fitted to the November 2022 data submission using rates through 2019. The 2020 incidence rates are included in statistical reports and displayed in graphs but are not used in the joinpoint model. Refer to an example from SEER*Explorer for Recent Trends in SEER Incidence Rate for All Cancer Sites, 2000-2020.
Exclusion of 2020 Incidence Data from Risks of Developing Cancer
For the November 2022 data submission, SEER uses 2017-2019 incidence rates instead of 2018-2020 to compute the risks of being diagnosed with cancer. Including 2020 in the estimates would correspond to someone living their entire lifetime under the pandemic-induced 2018-2020 incidence rates. Estimates of the risk of being diagnosed with cancer are calculated by applying cross-sectional, age-specific incidence rates and death rates to a hypothetical cohort of 10,000,000 live births. The probabilities of developing the specified cancer and dying of other causes (before developing the specified cancer) are calculated using multiple decrement life tables implemented in DevCan. The methods use the most recent 3-years to gain stability. For this reason, we have chosen to use 2017-2019 rather than 2018 through 2020 to compute the risks of being diagnosed with cancer.
Decline of 2020 Cancer Incidence Rates Relative to 2019 Rates
To measure the impact of COVID on 2020 incidence, we provide the table below that reports the decline of 2020 cancer incidence rates relative to 2019 rates for 10 major cancer sites using the percent change (PC) of the rates in 2020 compared to 2019, i.e., the difference in the rates in 2020 and 2019 divided by the rate in 2019. The list will be expanded in the future.
|Cancer Site||2019 Delay-Adjusted Rate3,4||2020 Delay-Adjusted Rate3||Percent Change (PC)||95% Confidence Interval
|All sites||459.5||416.9||-9.26%||-9.56%, -8.97%|
|Breast (Females)||134.4||121.5||-9.60%||-10.35%, -8.85%|
|Kidney and Renal Pelvis||18.2||16.5||-9.52%||-10.97%, -8.04%|
|Melanoma of the Skin||22.5||19.1||-15.17%||-16.43%, -13.90%|
|Non-Hodgkin Lymphoma||19.3||17.8||-7.76%||-9.20%, -6.31%|
|Urinary Bladder||18.5||17.3||-6.23%||-7.68%, -4.75%|
For More Information
Learn more about the impact of COVID on cancer incidence trends in the following article: