Regarding: ‘Incidence, mortality, and DALYs of global pharyngeal cancer: systematic analysis and projections based on global burden of disease study 2021’

Dear Editor

We read with great interest the article by Zhou et al. [1] on the global burden of pharyngeal cancer from 1990 to 2021. Drawing upon the latest Global Burden of Disease 2021 dataset, the authors comprehensively reported incidence, mortality, and disability-adjusted life years at global, regional, and national levels, highlighting striking differences across sex, age groups, and sociodemographic index strata. These findings provide important evidence for understanding the epidemiological patterns of pharyngeal cancer worldwide and lay a foundation for targeted public health interventions and policy-making.

For trend analysis, the authors estimated the estimated annual percentage change (EAPC) by fitting a single log-linear regression to age-standardized rates, which offers a simple and intuitive way to summarize long-term changes. However, EAPC assumes that the trend remains linear across the entire study period, which limits its ability to detect turning points or to reflect recent dynamics. We therefore suggest that, in addition to EAPC, the average annual percent change (AAPC) be reported, as it provides a more robust and comparable representation of recent trends, thereby enhancing the policy relevance of the results.

EAPC employs a single log-linear regression on age-standardized rates across the entire period, offering straightforward calculation and providing an overall average rate of change. However, it implicitly assumes a linear trend throughout the study period, making it susceptible to extreme values at the start and end years. It also fails to reveal potential turning points or phased changes occurring mid-period. This means that when disease burden experiences an initial rise followed by a decline, or when public health events cause phased turning points, EAPC often ‘averages out’ this information, obscuring short-term trends that are more valuable to decision-makers. In contrast, AAPC is derived from Joinpoint segmented regression: it first identifies significant inflection points, then calculates the time-weighted average of each segment’s APC within a preset time window (e.g. the last 5 or 10 years). The advantages of AAPC are threefold. First, it accurately reflects the overall direction and magnitude within a fixed interval, even when trends are nonlinear or phased. Second, it facilitates cross-country comparisons by standardizing the time window (e.g. ‘recent 5 years’), eliminating comparability issues caused by varying study durations or endpoint definitions. Third, it is more directly aligned with public health priorities, as policymakers often need to know whether the burden has been increasing or decreasing in the most recent decade. Furthermore, the computation of AAPC and its confidence intervals is supported by a mature methodological framework and widely used software, such as the National Cancer Institute Joinpoint Regression Program (https://surveillance.cancer.gov/joinpoint/), making the approach both transparent and reproducible [2–4].

In conclusion, we recommend that future analyses and reports supplement EAPC with AAPC, particularly for the most recent 5- or 10-year intervals. Doing so would yield clearer insights into recent changes in disease burden, enhance cross-country comparability, and better support evidence-based public health decision-making. We believe this methodological refinement would increase both the interpretability and practical utility of the findings.

Funding Statement

This study received no funding support.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

Data availability is not applicable to this article as no new data were created or analyzed in this study.