全国医疗照射频度估算方法研究  

Study on estimation of medical exposure frequency in China

作　　者：牛亚婷 苏垠平[1] 梁婧[1] 侯长松[1] 孙全富[1] Niu Yating;Su Yinping;Liang Jing;Hou Changsong;Sun Quanfu(Key Laboratory of Radiological Protection and Nuclear Emergency,China CDC,National Institute for Radiological Protection,Chinese Center for Disease Control and Prevention,Beijing 100088,China)

机构地区：[1]中国疾病预防控制中心辐射防护与核安全医学所辐射防护与核应急中国疾病预防控制中心重点实验室,北京100088

出　　处：《中华放射医学与防护杂志》2019年第5期325-330,共6页Chinese Journal of Radiological Medicine and Protection

摘　　要：目的利用部分省份放射诊疗机构的调查数据,研究全国放射诊疗总频度的估算方法。方法全国部分省份医疗照射频度水平委托调查涉及25个省份的557家放射诊疗机构。对数据进行了变量间的相关性分析、多重线性回归分析,比较包含不同变量组合的模型拟合效果。结果医疗照射频度与门急诊人次、放射诊疗设备数量、放射工作人员数量之间存在大的相关性(|r|>0.5)。从调查数据中通过分层随机抽样,得到日均放射诊疗工作负荷的代表性样本。利用多重线性回归,建立放射诊疗总人次与门急诊人次、住院人次、放射诊疗设备数量、放射工作人员数量和医院级别之间的数学模型,并进行了模型诊断。估算2016年全国放射诊疗总人次为5.89亿。同时,采用稳健回归和中位数回归,对医疗照射频次进行了估算。结论对全国放射诊疗总频度的估算有多种方法,建议采用随机抽样,结合统计和监测数据进行。样本的代表性十分重要,模型的指定和优化也需要更深入的研究。Objective To study the estimation method of medical exposure frequency in China through the survey of diagnostic radiology and radiotherapy institutions in selected provinces.Methods A survey of medical exposure frequency was carried out involving 557 diagnostic radiology and radiotherapy institutions in 25 provinces.The correlation analysis and multiple linear regression analysis were conducted,with the fitting effects of the models with different variables being compared.Results The total medical exposure frequencies highly correlated with number of outpatient,number of equipment and number of radiation workers (|r|>0.5).Representative samples of the daily medical exposure workload were obtained by stratified random sampling from the survey data.Mathematical models were built using the multiple linear regression between total medical exposure frequency and hospital levels,number of outpatients,number of inpatients,number of equipment,and number of radiation workers.The total medical exposure frequency in 2016 was estimated to be 589 million examinations based on the models.In addition,the frequencies of medical procedures were derived using the robust regression and the median regression.Conclusions There are several methods for estimating the total medical exposure frequency.It is desirable to use the stratified random sampling combined with the published statistical and monitoring data.The representativeness of sample is critical.The specification and optimization of models also require further study.

关 键 词：放射诊断 放射治疗 医疗照射 频度 估算 

分 类 号：R81[医药卫生—放射医学]