Hosseintabar Marzoni Atefeh-Sadat, Moghimbeigi Abbas, Faradmal Javad
Golestan Research Center of Gastroenterology and Hepatology, Golestan University of Medical Sciences, Golestan, Iran.
Modeling of Noncommunicable Diseases Research Center, Department of Biostatistics and Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran.
Osong Public Health Res Perspect. 2015 Apr;6(2):100-5. doi: 10.1016/j.phrp.2015.01.004. Epub 2015 Feb 19.
Recent studies of esophageal cancer (EC) and gastric cancer (GC) have been reported to have high incidence rates of these cancers in Golestan Province of Iran. The present study describes the geographical patterns of EC and GC incidence based on cancer registry data and display statistically significant regions within this province.
In order to map the distribution of upper gastrointestinal cancer, relative risk (RR) were calculated. Therefore, to estimate a more reliable RR, Poisson regression models were used. The adjusted models (adjusted to urban-rural area, sex, and grouped age proportion) were utilized. We considered two-component random effects for each observation, an unstructured (noncorrelated) and a group of "neighbor" (correlated) heterogeneities. We estimated the model parameters using Gibbs sampling and empirical Bayes method. We used EC and GC data that were registered with Golestan Research Center of Gastroenterology and Hepatology in the years 2004-2008.
The EC and GC maps were drawn for 2004-2008 in the province. Kalaleh and Minoodasht counties have a high RR of EC and GC in the years of study. In almost all years, the areas with a high RR were steady.
The EC and GC maps showed significant spatial patterns of risk in Golestan province of Iran. Further study is needed to multivariate clustering and mapping of cancers RRs with considering diet and socioeconomic factors.
据报道，近期对食管癌（EC）和胃癌（GC）的研究表明，伊朗戈勒斯坦省这些癌症的发病率很高。本研究基于癌症登记数据描述了EC和GC发病率的地理模式，并展示了该省内具有统计学意义的区域。
为了绘制上消化道癌症的分布图，计算了相对风险（RR）。因此，为了估计更可靠的RR，使用了泊松回归模型。采用了调整模型（根据城乡地区、性别和分组年龄比例进行调整）。我们为每个观测值考虑了两部分随机效应，一个是无结构的（不相关的），另一个是一组“邻居”（相关的）异质性。我们使用吉布斯抽样和经验贝叶斯方法估计模型参数。我们使用了2004 - 2008年在戈勒斯坦胃肠病学和肝病研究中心登记的EC和GC数据。
绘制了该省2004 - 2008年的EC和GC地图。在研究年份中，卡拉莱赫县和米努达斯ht县的EC和GC相对风险较高。几乎在所有年份，相对风险较高的地区都是稳定的。
EC和GC地图显示了伊朗戈勒斯坦省显著的风险空间模式。需要进一步研究在考虑饮食和社会经济因素的情况下对癌症RR进行多变量聚类和绘图。
