中国麻风皮肤病杂志 ›› 2026, Vol. 42 ›› Issue (10): 696-704.doi: 10.12144/zgmfskin202610696

• 论著 • 上一篇    下一篇

银屑病药品不良事件主动监测触发器的构建与验证

卓裕骞1,彭上杰1,张晋昕2,李薇1,梁景耀1,钟金宝1,汪梦1,李曼1,张三泉1,朱慧兰1,张锡宝1,张玉艳1   

  1. 1广州医科大学皮肤病研究所，广州市皮肤病医院，广东广州，510095；2中山大学公共卫生学院，广东广州，510080
  • 出版日期:2026-10-15 发布日期:2026-10-09

Development and validation of active surveillance trigger tools for adverse drug events in psoriasis

ZHUO Yuqian1, PENG Shangjie1, ZHANG Jinxin2, LI Wei1, LIANG Jingyao1, ZHONG Jinbao1, WANG Meng1, LI Man1, ZHANG Sanquan1, ZHU Huilan1, ZHANG Xibao1, ZHANG Yuyan1   

  1. 1 Institute of Dermatology Guangzhou Medical University, Guangzhou Dermatology Hospital, Guangzhou 510095, China; 2 School of Public Health, Sun Yat-sen University, Guangzhou 510080, China
  • Online:2026-10-15 Published:2026-10-09

PDF (PC)

25

摘要： 目的：构建银屑病患者的药品不良事件（ADEs）主动监测触发器，并验证其监测效果，旨在保障患者用药安全。方法：通过系统检索银屑病相关指南、共识、药品说明书、全球触发工具（global trigger tool，GTT）白皮书及相关文献，初步形成46项触发器条目。通过两轮德尔菲法对条目进行评分和筛选，运用SPSS 25.0软件进行数据分析，依据条目评分的均分（Mj≥第15.87百分位数）、满分频率（Kj≥第15.87百分位数）及变异系数（CV≤第84.13百分位数）等条件确定最终触发器条目。以结构化字段自动识别联合人工复核的方式评价触发器监测效果，并将ADEs检出结果与医院在国家药品不良反应监测系统上报情况进行对比。结果：两轮德尔菲法回收率均为100%，专家权威程度（Cr）均＞0.7，CV均＜0.3，专家协调系数（W）=0.344，最终确定48项主动监测触发器条目，涵盖实验室指标（16项）、临床症状（22项）、合用药物（7项）及干预措施（3项）四个维度。结合2025年20 335条门诊和住院银屑病患者就诊记录，结果显示触发器共触发2038次，阳性预测值（PPV）为45.04%，阴性预测值（NPV）为99.72%，灵敏度为93.16%，特异度为95.70%，准确率为95.61%。主动监测识别ADEs 760例，ADEs检出率为3.74%。结论：本研究结合GTT与德尔菲法，构建了一套适用于银屑病患者的ADEs主动监测触发器，有助于实现用药风险早期预警。

关键词: 银屑病, 药品不良事件, 触发器, 德尔菲法, 主动监测

Abstract: Objective: To develop a trigger tools set for active surveillance of adverse drug events (ADEs) in patients with psoriasis and to evaluate its triggering performance, with the aim of strengthening medication safety in this population. Methods: The 46 initial trigger items were formed through a review of psoriasis-related guidelines, consensus, drug instructions, Global Trigger Tool white papers and literature. The items were rated and refined through two rounds of Delphi consultation. Software SPSS 25.0 was used to analyze the data. The final trigger items were determined based on the arithmetic mean (Mj ≥ the 15.87th percentile), the full-mark rate (Kj ≥ the 15.87th percentile), and the coefficient of variation (CV ≤ the 84.13th percentile) for each item. A single-center retrospective analysis was then conducted to validate the triggers. Trigger activation was assessed through automated identification of structured fields followed by manual review, and ADEs detected by the active surveillance approach were compared with those reported through the hospital's voluntary reporting system. Results: The recovery rates of questionnaires were both 100%, the authority coefficient(Cr) was＞0.7, CV was＜0.3, and the Kendall(W) coefficient was 0.344. Finally 48 trigger items were formulated, categorized into four dimensions: laboratory indices(16 items), clinical symptoms(22 items), concomitant medications(7 items) and intervention measures(3 items). Based on 20 335 outpatient and inpatient encounter records of patients with psoriasis in 2025, the triggers generated 2038 signals. The positive predictive value (PPV), negative predictive value (NPV), sensitivity, specificity, and overall accuracy were 45.04%, 99.72%, 93.16%, 95.70%, and 95.61%, respectively. Active surveillance identified 760 ADE cases, yielding a detection rate of 3.74%. Conclusion: By integrating the GTT and Delphi method, this study established an active surveillance trigger tools for ADEs in psoriasis patients, which is designed to facilitate precise surveillance and early warning of medication-related risks.

Key words: psoriasis, adverse drug event, trigger tools, Delphi method, active surveillance