Page 109 - 《中国药房》2024年15期
P. 109

·药师与药学服务·


          基于莱文斯坦距离的易混淆药品目录自动生成算法及软件实现
                                                                                                           Δ

                                                       #
          陈 杨 ,淡重辉,何 瑶,阮 一,陈 肖,郑晓媛(重庆市急救医疗中心药剂科,重庆 400014)
                *
          中图分类号  R952      文献标志码  A      文章编号  1001-0408(2024)15-1899-06
          DOI  10.6039/j.issn.1001-0408.2024.15.17

          摘  要  目的  构建一种高效的易混淆药品目录自动生成算法,并在此基础上开发一种易混淆药品目录管理系统,以此提升对易
          混淆药品目录的管理效率。方法  本研究以莱文斯坦距离算法为理论基础,深入研究易混淆药品组的自动识别机制以及相似性阈
          值筛选方法,进而构建易混淆药品目录自动生成算法。在系统开发层面,本研究采用Visual Basic.NET作为编程语言,结合SQL
          Server 2008 R2 Express数据库管理平台,设计开发易混淆药品目录管理系统。结果  相似性阈值δ是易混淆药品目录自动生成算
          法的关键参数,随着δ的逐渐增大,易混淆药品的总数逐渐减少,而易混淆药品组数则呈现先上升后下降的变化规律。在实际应
          用中,可根据药品的通用名或品种名构建易混淆药品目录,对应的相似性阈值可取0.75和0.83。此外,本课题组开发的易混淆药
          品目录管理系统将原本耗时约1周的目录建立时间缩短至不到1 h,极大地提升了工作效率。结论  本研究构建的易混淆药品目
          录自动生成算法高效快速,为易混淆药品管理提供了强有力的技术支撑。开发的易混淆药品目录管理系统极大地减轻了目录建
          立和维护的时间成本,提高了对易混淆药品目录的管理效率。
          关键词  易混淆药品;莱文斯坦距离;药品目录;相似性阈值


          Software  implementation  of  Levenshtein  distance-based  algorithm  for  automatically  generating  easily
          confused drug catalogs
          CHEN Yang,DAN Chonghui,HE Yao,RUAN Yi,CHEN Xiao,ZHENG Xiaoyuan(Dept.  of  Pharmacy,
          Chongqing Emergency Medical Center, Chongqing 400014, China)

          ABSTRACT   OBJECTIVE  To  create  a  highly  effective  algorithm  for  automatically  generating  easily  confused  drug  catalogs
         (ECDC),  as  well  as  to  develop  a  management  system  for  ECDC  based  on  this  algorithm,  in  order  to  improve  the  management
          efficiency  of  ECDC.  METHODS  This  study,  based  on  Levenshtein  distance  algorithm,  delved  deeply  into  the  automatic
          identification mechanism of easily confused drugs and the screening method for determining similarity thresholds, ultimately leading
          to the development of an algorithm for automatically generating ECDC. Besides a management system was designed and developed,
          using  SQL  Server  2008  R2  Express  as  the  data  storage  platform  and  Visual  Basic.NET  as  the  programming  language.  RESULTS
          The  similarity  threshold  δ  played  a  crucial  role  in  the  algorithm  for  automatically  generating  ECDC. As  the  value  of  δ  gradually
          increased,  the  total  count  of  easily  confused  drugs  decreased  gradually,  while  the  count  of  drug  groups  exhibited  a  pattern  of
          initially  increasing  and  then  decreasing.  Practically,  ECDC  could  be  created  using  either  the  generic  or  varietal  names  of  drugs,
          with  corresponding  similarity  thresholds  of  0.75  and  0.83.  Furthermore,  ECDC  management  system  had  significantly  reduced  the
          time  required  to  establish  a  catalog  from  about  one  week  to  less  than  one  hour,  resulting  in  a  substantial  enhancement  in  work
          efficiency.  CONCLUSIONS  The  algorithm  used  to  automatically  generate  ECDC  is  highly  efficient  and  rapid,  offering  robust
          technical assistance for the management of easily confused drugs. Implementing the ECDC system can greatly reduce the time cost
          related to building and maintaining the catalogs, thus significantly improving the efficiency of managing ECDC.
          KEYWORDS    easily confused drugs; Levenshtein distance; drug catalogs; similarity threshold


             Δ 基金项目 重庆市卫生健康委医学科研项目(No.2024WSJK030);
                                                                 易混淆药品指的是在外观(看似)、读音(听似)上相
          重庆市临床药学重点专科建设项目(No. 渝卫办发〔2023〕2 号,No. 渝
          卫办发〔2023〕69号)                                      似,或存在一品多规、多剂型、多厂家等情形,容易引发
             *第一作者 主管药师,硕士。研究方向:医院药学、化学计量学。                  混淆的药品      [1―2] 。这类药品是引起调配差错的重要原
          E-mail:yangchen8786@sina.com
                                                             因 [3―4] 。据估计,美国每年与用药错误相关的费用高达
             # 通信作者 主任药师,硕士。研究方向:药理学、医院药学。
          E-mail:thymolblue@163.com                          数百万美元,其中听似和看似药品导致的用药错误分别


          中国药房  2024年第35卷第15期                                              China Pharmacy  2024 Vol. 35  No. 15    · 1899 ·
   104   105   106   107   108   109   110   111   112   113   114