Qualidade dos dados de notificação da violência autoprovocada em Manaus, Amazonas: perfil epidemiológico e tendência temporal

Abstract

Self-inflicted violence constitutes a relevant public health problem because of both the magnitude of notifications and deaths and its repercussions for victims, families, communities, and health services. Surveillance of this health event depends on the quality of the information recorded in the Notifiable Diseases Information System (SINAN), since incomplete, inconsistent, or untimely data may limit understanding of the phenomenon, identification of vulnerable groups, and planning of prevention and care actions. This study aimed to assess the quality of data recorded in self-inflicted violence notifications and to analyze the epidemiological profile and temporal trend of cases reported in Manaus, Amazonas, from 2015 to 2024. This was an observational, retrospective epidemiological study based on repeated cross-sectional data, with descriptive and analytical components and temporal trend analysis. Secondary data from the SINAN Interpersonal/Self-Inflicted Violence Notification Forms were used. Data quality was assessed according to the attributes of completeness, consistency, representativeness, timeliness, and acceptability. For the epidemiological analysis, annual notification rates and absolute and relative frequencies of the variables were calculated, and the chi-square test, standardized residual analysis, and Kendall’s tau coefficient were applied. Completeness of the variables related to the victims’ profile was classified as regular throughout the historical series. Among variables related to the circumstances of the events, completeness remained regular until 2022 and was classified as good in 2023 and 2024. The operational representativeness indicator met the established criterion in all years, while acceptability was classified as acceptable during most of the period and moderate in 2020. Consistency was classified as low in 2016, regular in five years, and excellent in four years. Timeliness showed unsatisfactory results throughout the study period. Regarding the epidemiological profile, notifications increased, with higher frequencies among women (69.1%), adolescents (46.1%), mixed race individuals (78.2%), single individuals (51.4%), and students (25.0%). Events occurred more frequently at home (83.7%) and at night (24.9%), were characterized by repeated violence (43.4%), and mainly involved poisoning (47.2%). Notifications were predominantly made by public health institutions (98.9%) and secondary care services (48.4%). The assessment revealed persistent weaknesses in completeness, timeliness, and some combinations used to verify record consistency. Gaps related to sexual orientation and gender identity limit the visibility of socially vulnerable groups, while the limited participation of private health services suggests weaknesses in notification coverage. It is concluded that interpretation of the epidemiological profile and temporal trend depends on the quality of the records. Strengthening surveillance requires improvements in data completion and information flow, continuing education for health professionals, and greater coordination among Primary Health Care, emergency services, and the Psychosocial Care Network, thereby supporting timely identification of risk situations, continuity of care, and prevention of recurrent events and severe outcomes.