Detection of Patients with Self-Injuries and Suicide Attempts in the Electronic Registration System of Public Health Providers in the Autonomous City of Buenos Aires and Preliminary Evaluation of Timely Care

Abstract

Self-harm behavior is a deliberate, intentional, and direct act aimed at causing physical harm to one's own body, seeking to have an immediate impact without the intention of causing death. In contrast, a suicide attempt is self-inflicted harm with the intention of taking one's own life (Villarroel et al. 2013). The presentation of such episodes constitutes marker events that highlight the need for timely and appropriate therapeutic intervention. For this reason, it is necessary to generate epidemiological information on mental health that reflects the magnitude and frequency of these issues in the population. To this end, the Programmatic Area of Argerich Hospital, the Information and Statistics Management Office, and the Epidemiology Management Office initiated a collaborative work process for the detection in electronic medical records of problems related to self-injuries and suicide attempts. This article presents the initial advances of this work process with the intention of expanding this strategy in the future and contributing to public health in the quantification, characterization, and evaluation of access to treatments. Better understanding of the epidemiology of self-injuries, including the current reality of the care process, will allow us to quantify the problem and provide evidence for the consideration of public policies and programs, reassessing care pathways, organization of mental health services, and identifying current resource needs for care. The objective of this work was to build an automated process for both identifying patients and facilitating their subsequent access to a consultation with a mental health professional based on records in the electronic information system of public health providers in the City of Buenos Aires. To achieve this, various text processing tools were used. Initially, a survey of key terms associated with this issue was conducted through a bibliographic search of related concepts, reading medical histories, and interviewing professionals. Then, with a set of identified terms, a list of regular expressions was compiled, which were selected as terminology to include in the search process, along with a set of terms to exclude to remove false positives. Based on the results of this initial screening, validation of the problems of interest was carried out and the inclusion and exclusion criteria were adjusted to arrive at the final subset, resulting in a list of 298 unique expressions related to the entities of interest. Subsequently, a characterization of individuals was performed based on the reasons for consultation and their detection dates, considering the number of positive services following to monitor post-identification follow-up. In a second phase, the analysis will be extended by conducting searches in evolutions.