Introduction: Health Information Systems are potential instruments to analyze health situation; however, the non-compulsory filling of a single common field makes it difficult to link systems’ data. This study aimed to describe and evaluate the adequacy of the strategies used to perform data linkage between databases from the Brazilian Public Health System (SUS) as to records for breast cancer control. Methods: The Breast Cancer Control Information Systems (SISMAMA), the Outpatient Information System (SIA, through Individualized Outpatient Service Production - BPA-I - and High-Complexity Outpatient Procedures Authorization Forms - APAC), the Hospital Information System (SIH), and the Mortality Information System (SIM) were linked probabilistically. The baseline was constructed by records with “suspected” and “highly suspected malignancy” from the second half of 2010. The linkage strategy included 15 steps. Registries with the national health service user card (CNS) or social security number (SSN) were used to estimate the sensitivity of the strategy, considering matches between records identified in the initial steps as gold standard, when these fields were used as key for blocking. Results: Using CNS and the SSN as a linkage strategy allowed to identify the high proportion of true matches across databases in which these variables were inputted: 47.3% in follow-up mammography records, 41.4% in SIH, and 45.5% in APAC. The sensitivity of the linkage strategy was 100%. Conclusion: The study showed that the strategies were satisfactory and the use of CNS and SSN allowed many matches, even without critical proceedings and with the possibility of linkage between databases based on information from only a few identification fields.
CITATION STYLE
Tomazelli, J. G., Girianelli, V. R., & Silva, G. A. E. (2018). Strategies used to link health information systems for the follow-up of women with abnormal mammograms in the Brazilian public health system. Revista Brasileira de Epidemiologia, 21. https://doi.org/10.1590/1980-549720180015
Mendeley helps you to discover research relevant for your work.