Development and application of a rapid identification algorithm for cutaneous lupus erythematosus and its subtypes based on medical insurance databases
10.12173/j.issn.1005-0698.202504078
- VernacularTitle:基于医疗保险数据库的皮肤型红斑狼疮及其亚型快速识别规则构建与应用
- Author:
Yutong WANG
1
;
Xianglong MENG
;
Yu PAN
;
Chen WEI
;
Hui JIN
;
Shengfeng WANG
Author Information
1. 北京大学公共卫生学院流行病与卫生统计学系(北京 100191);教育部重大疾病流行病学重点实验室(北京 100191)
- Publication Type:Journal Article
- Keywords:
Cutaneous lupus erythematosus;
Medical insurance;
Rapid identification algorithm
- From:
Chinese Journal of Pharmacoepidemiology
2025;34(7):743-752
- CountryChina
- Language:Chinese
-
Abstract:
Objective To develop and validate data extraction and patient identification algorithms for cutaneous lupus erythematosus(CLE)and its two subtypes,discoid lupus erythematosus(DLE)and subacute cutaneous lupus erythematosus(SCLE),and to enable high-efficiency patient identification in large-scale electronic health databases.Methods This study utilized data from the 2013-2017 National Insurance Claims for Epidemiological Research(NICER)to construct data extraction and rapid patient identification algorithms.The manual verification results were used as gold standard to assess the sensitivity and specificity of the algorithms.Additionally,the basic characteristics of the identified patients were analyzed.Results Initially,standardized expressions were developed based on medical terminology and diagnostic codes.These were further refined with input from clinicians to include potential synonyms and common misspellings,improving the preliminary screening expressions.Through iterative verification by clinicians and data management engineers,a final disease-specific screening algorithm was established.The developed extraction and identification algorithms for all 3 targeted disease demonstrated strong performance,with sensitivity values of 0.985,1.000,and 0.991,and specificity values of 0.997,0.999,and 0.998 for CLE,DLE,and SCLE,respectively.A total of 34,554 CLE cases,including 2,879 DLE cases,and 623 SCLE cases were identified between 2013 and 2017,with a higher prevalence among females than males.Conclusion This study developed and validated an identification algorithm for CLE patients based on medical insurance databases,demonstrating high performance.The proposed algorithm provides a methodological framework and empirical evidence for designing and optimizing big data-driven rapid patient identification algorithms in dermatology research.
