Quality of the notifications of interpersonal and self-inflicted violence in the state of Rio de Janeiro, Brazil, 2009-2016

Abstract Background Data from Health Information Systems are widely used for the production of knowledge. For their effective use, however, it is essential to ensure that they are valid and reliable. Objective Evaluate the quality of reports of violence in the state of Rio de Janeiro aiming to contribute to the improvement of information. Method Ecological study on the consistency of notifications of violence recorded to the Information System on Diseases of Compulsory Declaration in the state of Rio de Janeiro from 2009 to 2016. The temporal trend of the proportions of inconsistencies, for each group of related fields, was evaluated by simple linear regression and its statistical significance ( p≤0.05). Results During the study period, 103,841 cases of violence were reported, but in 31.7% of them it was not possible to identify whether they were interpersonal or self-inflicted. The main inconsistencies were incongruence between psychological violence and aggression, gender identity and sex, type of violence and age of the victim, self-inflicted violence and characteristics of the aggressor and type of violence; most of them presented stability in the period ( p>0.05). Conclusion The results show significant inconsistencies in the reports of violence that compromise the use of this information.

摘要 背景：健康信息系统（Health Information Systems）所产生的数据被广泛用于知识产出。然而，若要实现其有效应用，必须确保数据具备有效性与可靠性。研究目的：评估里约热内卢州暴力事件报告的质量，以期为信息优化工作提供助力。研究方法：本研究为生态学研究，针对2009年至2016年里约热内卢州强制申报疾病信息系统（Information System on Diseases of Compulsory Declaration）中收录的暴力事件申报一致性展开分析。针对每一组关联字段，采用简单线性回归（simple linear regression）分析不一致率的时间趋势，并检验其统计学显著性（p≤0.05）。研究结果：研究周期内，共计上报103841起暴力事件，但其中31.7%的案例无法明确其属于人际暴力还是自伤暴力。主要的数据不一致情形包括：心理暴力与攻击行为的界定不符、性别认同与生物学性别不匹配、暴力类型与受害者年龄不符、自伤暴力与施暴者特征及暴力类型不匹配；绝大多数不一致情形在研究周期内呈现稳定趋势（p>0.05）。研究结论：本研究结果显示，暴力事件报告存在显著的数据不一致问题，将对该类信息的实际应用造成不利影响。