Scholarlink

Purpose <p>Prostate cancer constitutes a major public health challenge in China; however, its spatiotemporal dynamics remain unclear. Clarifying these patterns is critical for guiding targeted prevention and control efforts to alleviate the disease burden.</p> Materials and methods <p>A descriptive spatiotemporal study was conducted utilizing city-level prostate cancer registry data from 2013 to 2016. The data were retrieved from the Annual Reports on Cancer Registration in China 2016 to 2019 published by the Chinese National Cancer Center. The analytical framework integrated spatial autocorrelation analysis (global and local clustering) and Bayesian spatiotemporal modeling. Disease dynamics were comprehensively assessed using Bayesian spatiotemporal models, which incorporated structured and unstructured spatial effects, temporal trends, and spatiotemporal interactions.</p> Results <p>Significant spatial clustering and geographic variation in prostate cancer incidence was identified. Rates were higher in southeast coastal regions and lower in northwest and northern China. Global Moran’s I values increased from 0.141 in 2013 to 0.173 in 2016 (all <i>P</i> &lt; 0.001), indicating strengthened spatial dependence over time. The global spatiotemporal Moran’s I for 2013–2016 was 0.633 (Z = 31.933, <i>P</i> &lt; 0.001), further confirming the aggregated pattern across space and time. Prostate cancer incidence showed a general increasing trend. Bayesian models revealed that spatial variation was dominated by unstructured spatial effects, with structured effects accounting for only 7.9% of the total spatial variance. This suggests that the observed variation was primarily driven by local, city-specific characteristics rather than broad regional clustering. Negative spatiotemporal interactions indicated that the evolution of incidence risk varied substantially across cities after adjusting for main spatial and temporal effects.</p> Conclusions <p>Spatiotemporal interaction was identified as a major driver of incidence variation, which generally reduced relative risk in most areas but amplified it in a specific subset of cities. Therefore, policymakers should prioritize targeted interventions in these high-risk hotspots, rather than relying solely on generalized temporal trends for resource allocation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Spatiotemporal patterns and clustering of prostate cancer incidence in China: a Bayesian modeling study of cancer registry data

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月14日

查看全文
  • Xu Zhu,
  • Zhan Chen,
  • Meng-Wei Ge,
  • Attiq-Ur Rehman,
  • Hong-Lin Chen,
  • Hua Zhu,
  • Bing Zheng

摘要

Purpose

Prostate cancer constitutes a major public health challenge in China; however, its spatiotemporal dynamics remain unclear. Clarifying these patterns is critical for guiding targeted prevention and control efforts to alleviate the disease burden.

Materials and methods

A descriptive spatiotemporal study was conducted utilizing city-level prostate cancer registry data from 2013 to 2016. The data were retrieved from the Annual Reports on Cancer Registration in China 2016 to 2019 published by the Chinese National Cancer Center. The analytical framework integrated spatial autocorrelation analysis (global and local clustering) and Bayesian spatiotemporal modeling. Disease dynamics were comprehensively assessed using Bayesian spatiotemporal models, which incorporated structured and unstructured spatial effects, temporal trends, and spatiotemporal interactions.

Results

Significant spatial clustering and geographic variation in prostate cancer incidence was identified. Rates were higher in southeast coastal regions and lower in northwest and northern China. Global Moran’s I values increased from 0.141 in 2013 to 0.173 in 2016 (all P < 0.001), indicating strengthened spatial dependence over time. The global spatiotemporal Moran’s I for 2013–2016 was 0.633 (Z = 31.933, P < 0.001), further confirming the aggregated pattern across space and time. Prostate cancer incidence showed a general increasing trend. Bayesian models revealed that spatial variation was dominated by unstructured spatial effects, with structured effects accounting for only 7.9% of the total spatial variance. This suggests that the observed variation was primarily driven by local, city-specific characteristics rather than broad regional clustering. Negative spatiotemporal interactions indicated that the evolution of incidence risk varied substantially across cities after adjusting for main spatial and temporal effects.

Conclusions

Spatiotemporal interaction was identified as a major driver of incidence variation, which generally reduced relative risk in most areas but amplified it in a specific subset of cities. Therefore, policymakers should prioritize targeted interventions in these high-risk hotspots, rather than relying solely on generalized temporal trends for resource allocation.