Garcia Klauss Kleydmann Sabino, Xavier Danielly Batista, Soremekun Seyi, Abrahão Amanda Amaral, Drakeley Chris, Ramalho Walter Massa, Siqueira André M
Center of Tropical Medicine, University of Brasilia, Brasília 70904-970, Brazil.
Escola Superior de Agricultura Luís de Queiroz, Universidade de São Paulo, Piracicaba 13418-900, Brazil.
Trop Med Infect Dis. 2023 Dec 14;8(12):519. doi: 10.3390/tropicalmed8120519.
The objective is to describe the results and the methodological processes of record linkage for matching deaths and malaria cases.
A descriptive cross-sectional study was conducted with probabilistic record linkage of death and malaria cases data in Brazil from 2011 to 2020 using death records from the Mortality Information System (SIM) and epidemiological data from the Notifiable Diseases Information System (Sinan) and Epidemiological Surveillance Information Systems for malaria (Sivep-Malaria). Three matching keys were used: patient's name, date of birth, and mother's name, with an analysis of cosine and Levenshtein dissimilarity measures.
A total of 490 malaria deaths were recorded in Brazil between 2011 and 2020. The record linkage resulted in the pairing of 216 deaths (44.0%). Pairings where all three matching keys were identical accounted for 30.1% of the total matched deaths, 39.4% of the matched deaths had two identical variables, and 30.5% had only one of the three key variables identical. The distribution of the variables of the matched deaths (216) was similar to the distribution of all recorded deaths (490). Out of the 216 matched deaths, 80 (37.0%) had poorly specified causes of death in the SIM.
The record linkage allowed for the detailing of the data with additional information from other epidemiological systems. Record linkage enables data linkage between information systems that lack interoperability and is an extremely useful tool for refining health situation analyses and improving malaria death surveillance in Brazil.
描述用于匹配死亡病例和疟疾病例的记录链接的结果及方法过程。
采用描述性横断面研究，对2011年至2020年巴西死亡病例和疟疾病例数据进行概率性记录链接，使用了死亡率信息系统（SIM）的死亡记录以及法定传染病信息系统（Sinan）和疟疾流行病学监测信息系统（Sivep - Malaria）的流行病学数据。使用了三个匹配键：患者姓名、出生日期和母亲姓名，并分析了余弦和莱文斯坦差异度量。
2011年至2020年巴西共记录了490例疟疾死亡病例。记录链接使得216例死亡病例（44.0%）成功配对。三个匹配键均相同的配对占总匹配死亡病例的30.1%，有两个相同变量的匹配死亡病例占39.4%，只有三个关键变量中的一个相同的占30.5%。匹配死亡病例（216例）的变量分布与所有记录死亡病例（490例）的分布相似。在216例匹配死亡病例中，80例（37.0%）在SIM中的死亡原因说明不明确。
记录链接能够利用其他流行病学系统的额外信息详细说明数据。记录链接实现了缺乏互操作性的信息系统之间的数据链接，是完善巴西健康状况分析和改善疟疾死亡监测的极其有用的工具。
