Hospitalizations and adverse drug events in the Brazilian unified health system: a ten-year retrospective analysis of routine data

• Autor(a) principal: Silva, Lunara Teles
• Data de Publicação: 2022
• Outros Autores: Modesto, Ana Carolina Figueiredo, Oliveira, Rodrigo Alves de, Amaral, Rita Goreti, Lopes, Flavio Marques
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Revista de Saúde Pública
• Texto Completo: https://www.revistas.usp.br/rsp/article/view/203522
• topic: Drug-related side effects and adverse reactions; Pharmacoepidemiology; Hospitalization; Databases, factual

OBJECTIVE To describe the frequency and characteristics of hospitalizations for/with adverse drug events in the Brazilian unified health system routine data. METHODS Nationwide retrospective study using data obtained from a period of ten years from the Brazil Hospital Information System (SIH-SUS), an administrative database that registers hospitalizations in the unified health system. We selected hospitalizations with primary and/or secondary diagnosis related to adverse drug events according to a list of validated International Classification Disease 10th edition (ICD-10) codes. These events were described according to year, age group, sex, length of hospital stay, mortality, hospital costs, Brazilian geographical region, and category of ICD-10 codes. Crude hospitalization rates of adverse drug events per 100,000 inhabitants were obtained and Joinpoint Regression was used to analyze temporal changes in these rates along the years. The most frequent ICD-10 codes were also identified. RESULTS Over ten years, 603,663 hospitalizations in Brazil were found in the database, out of which 2.5% of the patients died. Though 2009 had the highest prevalence of hospitalization per 100,000 inhabitants (32.57), no significant annual change in rates was found for the entire period. All age groups and sexes presented a jointpoint in temporal series; however, only women had a significative increase trend. The most frequent codes were from the chapter of mental and behavioral disorders (F19.2, F19.0, and F19.5 codes). CONCLUSIONS The database methodology can be useful to estimate frequencies of adverse drug events and perform characterization nationwide and to help monitor morbidity along the years.