臧国全 贾瑞莹

摘 要：[目的/意义]计量医疗数据中的病种隐私值，并基于人口统计变量对病种隐私进行敏感性分析，为医疗隐私保护和医疗数据利用提供基础条件。[方法/过程]采用联合分析法，计算患者对医疗数据保护服务每个属性的边际支付意愿，求得医疗数据中病种的隐私值，通过独立样本T检验和单因素方差分析，研究不同个体特征之间病种隐私的差异是否具有统计学意义。[结果/结论]基于访谈法的预调查结果，艾滋病隐私值为最高;采用联合分析法和Logit模型，计量其他病种隐私值，结果显示，降序排列隐私值较高的病种为精神疾病、传染病（艾滋病除外）、恶性肿瘤、心脏与脑血管疾病;基于人口统计变量的隐私值分析显示，性别、年龄、受教育程度3个变量下医疗数据中病种隐私差异具有显著性，且男性高于女性，年龄段为41～65岁的人群高于其他年龄段人群，教育程度越高隐私越大。

关键词：医疗数据;病种隐私;隐私计量;联合分析

DOI：10.3969/j.issn.1008-0821.2020.05.019

〔中图分类号〕G250.252 〔文献标识码〕A 〔文章编号〕1008-0821（2020）05-0161-08

Measurement and Analysis of Disease Privacy in Medical Data

Zang Guoquan1，2 Jia Ruiying1

（1.School of Information Management，Zhengzhou University，Zhengzhou 450000，China;

2.Institute of Data Science，Zhengzhou University，Zhengzhou 450000，China）

Abstract：[Purpose/Significance]Measuring the privacy value of the disease in medical data and sensitivity analysis of disease privacy based on demographic variables provide basic conditions for medical privacy protection and medical data utilization.[Method/Process]Using the method of conjoint analysis，the privacy values of the diseases in medical data were obtained by calculating the patients marginal willingness to pay for each attribute of medical data protection service.And independent sample T-test and one-way ANOVA were used to investigate whether the differences in disease privacy between different individual characteristics were statistically significant.[Result/Conclusion]Pre-survey results based on interviews showed that AIDS had the highest privacy value;using the method of conjoint analysis and Logit model to measure the privacy value of other diseases，the results showed that the diseases with higher privacy value in descending order were mental diseases，infectious diseases（except AIDS），malignant tumors，heart and cerebrovascular diseases;Privacy value analysis based on demographic variables showed that there were significant differences in disease privacy among three variables：gender，age and education level.Males were higher than females，adults were higher than other age groups，and the higher education level，the greater privacy.

Key words：medical data;disease privacy;privacy measurement;conjoint analysis

1 問题提出

医疗数据是健康医疗大数据的重要部分。医疗数据是医疗机构通过其医疗信息系统采集的患者就医过程中产生的数据。中共中央、国务院印发的《“健康中国2030”规划纲要》和《关于促进和规范健康医疗大数据应用发展的指导意见》（2016）[1-2]，将健康医疗大数据分3类：健康数据（亦称院外数据）、医疗数据（亦称院内数据）和基因数据。据CHIMA（中华医院管理学会信息管理专业委员会）调查（2017），85%以上医院已实施医疗信息系统，产生的医疗数据在整个健康医疗大数据中规模占比最大，采集最完善[3]。……