Year: 2024
Type: article
Abstract: We reviewed the commentary provided by Semprini and Osazuwa-Peters.1 We appreciate the authors’ praise of our work and their efforts to further illuminate the impact of COVID-19 among people living with cancer. In this reply, we address their comments to clarify our research. The authors state that although the Surveillance, Epidemiology, and End Results (SEER) data represents the cancer burden of the United States, the select states in SEER may have different COVID-19 rates than the rest of the country. Regarding this concern about the representativeness of SEER data and potential regional differences in COVID-19 rates, we have acknowledged that SEER only covers selected states and regions in our previous work.2-4 However, the SEER population-based tumor registries are designed to include geographically distinct regions in the United Stats, chosen to represent the racial and ethnic heterogeneity of the country.2,3 The SEER database aims to be representative of the US cancer population, capturing approximately up to 48.0% of the US population.5 Thus, it remains a gold standard in population-based cancer epidemiological research.6 One limitation of SEER data, however, is its single-cause attribution for death, which may oversimplify the complexity of mortality among cancer patients. Many cancer patients have multiple comorbid conditions, and COVID-19 can exacerbate underlying illnesses, such as advanced metastatic disease, indirectly contributing to their death. Claims databases, as discussed in recent work by Gross et al.,7 may offer additional insights into these multimorbidity patterns, providing a more nuanced understanding of COVID-19’s impact on patients with cancer and other coexisting conditions. The authors state that although standardized mortality ratio calculations have been considered the gold standard for comparing mortality rates, using the SEER databases to calculate standardized mortality ratios may underestimate the death toll attributed to COVID-19 among people with cancer during 2020 compared with absolute differences in COVID-19 mortality. We acknowledge the limitations of standardized mortality ratios but maintain that they remain a gold standard in epidemiological studies for comparing mortality rates between populations. Standardized mortality ratios are preferred when comparing comparative studies or when demographic adjustments are necessary. Standardized mortality ratios adjust for differences in age and sex distributions, which is essential when comparing populations like people living with cancer, who have distinct risk profiles, and are older than the general population. After careful consideration with our statistical team and the National Cancer Institute’s SEER statistical team, we elected to use standardized mortality ratio calculations as the primary outcome for analyzing COVID-19 mortality. Standardized mortality ratios are calculated by dividing the observed number of deaths in the study population (numerator) by the expected number of deaths based on age-specific rates in a standard population (denominator). Although it’s true that small changes in the numerator or denominator can influence the standardized mortality ratio, this sensitivity highlights the measure’s ability to accurately adjust for population differences. Importantly, standardized mortality ratios are indirect measurements and are not intended for direct comparison between different studies or populations without considering underlying demographic and epidemiological contexts. Therefore, Semprini and Osazuwa-Peters1 report a 365% higher mortality rate in people with cancer compared with the general public based on absolute mortality rates, however, this figure cannot be directly compared with our reported 130% increased risk of mortality based on calculated standardized mortality ratios. Our objective was to understand the relative risk of COVID-19 mortality among cancer patients compared with the general population but not quantify the additional mortality burden due to COVID-19 in cancer patients. Absolute mortality rates do not adjust for differences in age, sex, or other critical risk factors between populations. As a result, they may overestimate or underestimate the true relative risk when used for direct comparisons. In contrast, standardized mortality ratios provide a standardized measure that facilitates a more accurate comparison of mortality risks between populations with differing demographic profiles. We still agree that absolute mortality rates are valuable for understanding the overall burden of disease and for public health planning. However, for comparative risk assessments, especially between populations with differing demographic profiles, standardized measures like standardized mortality ratios remain the gold standard. Ultimately, the work of Semprini and Osazuwa-Peters1 complements our results, and together they provide a more comprehensive understanding of the impact of COVID-19 on people living with cancer. Kyle Mani, BS (Conceptualization; Writing—original draft; Writing—review & editing), Xue Wu, MS (Writing—review & editing), Daniel E. Spratt, MD (Writing—review & editing), Ming Wang, PhD (Methodology; Writing—original draft; Writing—review & editing), and Nicholas G. Zaorsky, MD, MS (Conceptualization; Writing—original draft; Writing—review & editing). None declared. The authors declare no competing interests. The data are provided in the SEER database, which is freely accessible to the public. The relevant session information, that is, the user-submitted request, from the current work and abbreviated dataset (from SEER) are provided in the source data file. The source data underlying all tables and figures are provided as a source data file. We comply with all relevant ethical regulations. The datasets generated and analyzed during the current study are available in the SEER repository (https://seer.cancer.gov/seerstat/). These data are freely available via the National Cancer Institute SEER program, and thus the study was exempt from institutional review board review. There are no participants in the study, and thus there is no consent form.
DOI: 10.1093/jnci/djae290
Authors: Kyle A Mani, Xue Wu, Daniel E Spratt, Ming Wang, Nicholas G Zaorsky
Language: English
Open Access status: bronze
OpenAlex ID: W4404376913