利用肿瘤登记数据评价职业和行业危险因素

An Approach to use Cancer Registration to Assess Cancer Risks by Occupation and Industry

本研究以肺癌为例探讨了应用肿瘤登记数据评价职业和行业肿瘤危险度的方法。以天津肿瘤登记处1981～1987年期间的登记数据为基础，利用病例一对照方法研究了肿瘤的职业和行业危险度，共分析了天津市区30个职业和27个行业。每一被研究的职业／行业为暴露组，其它是未暴露组。被研究的肿瘤为病例组，其它为对照组。统计分析方法是非条件logistic回归模型。调整因素包括性别、年龄和诊断年份。结果：以肺癌为例，研究结果与既往报告雷同。对天津市职业和行业的肿瘤危险度进行了系统的分析，提出天津市职业和行业的肿瘤危险因素。这种设计的主要优点是①可充分利用已经有的肿瘤登记的数据。②简便易行的、快速的和低成本地评价肿瘤危险因素或者潜在的危险因素。③所有肿瘤登记或任何疾病登记的数据库的因素均可以被评价，例如职业、行业、教育等。④减少分析中的混杂和偏倚。

This study explored how to use cancer registration data to assess cancer risk. Based onthe data of the Tianjin Cancer Registry, the occupational and industrial risks for lung cancer wereassessed as examples. All cases of cancer aged at least 20 years and reported during 1981 - 1987 to theTCR formed the basis of the study. A comparison was made using case - case methodology that cases ofone cancer site compared with cases of all other sitcs. Data were re - coded according to 30 occupationaland 27 industrial categories. The individuals employed in the interested category were considerd asexposed, and subjects employed in all other categories were considered as unexposed. Statistical analysiswas conducted by unconditional multiple logistic regression. Regression models included sex, age andyear of diagnosis. From our study, the results are constant with previous reports. The systematicanalysis of the association between occupation and industry of employment and cancer among men andwomen registered in the TCR provided confirmation in this population of a number of known orsuspected occupational cancer risk factors studied in industrialized countries. The main advantages of thisdesign are ①To use the data of a cancer registration that are already existed. ② It is easy, fast and lesscost to screening the cancer risk or potential cancer risk. ③ All factors in the database of cancerregistration could be evaluated, for example, occupation, industry, education,etc. ④The way to definecases and exposed groups could decrease the influence of confounding and bias.