The development of the study is mainly based on the combination and analysis of data from different public registries that are representative of the Catalan population and that include variables of a very diverse nature.
A retrospective nested case-control study defined from the dynamic cohort of population with access to public health services in Catalonia in the period 2014-2019 will be carried out. The study population or cases are those persons attended for suicide attempts in Catalan public health services and who are registered in CRS during the study period.
We address an important problem of underreporting of suicide attempt; not all cases of suicide attempt are detected or reported as such, so they are not correctly coded and registered in CRS. Therefore, hospital, emergency and primary care CMBD databases will also be reviewed and patients with a diagnosis coded for self-injury will be included as cases, since the ICD codes (ICD-9-CM and ICD-10-CM), currently used to code suicide attempts, do not distinguish between self-injury with or without suicidal intent. Overall, a total sample of about 15,000 cases is estimated.
All available data are of public origin and are managed through the Agència de Qualitat i Avaluació Sanitària de Catalunya (AQuAS).
As results, measures of incidence of attempted and reattempted suicide will be estimated and the relationship with the main risk factors at individual and population level will be calculated. The great challenge lies in the large number of risk factors that may be explanatory of suicidal behavior and their diverse origins: sociodemographic (e.g., age, sex, or area of residence), socioeconomic (e.g., income level), clinical (e.g., medical or psychiatric diagnoses), and health care use (e.g., treatment visits and medication) factors.
To address this challenge, we will develop advanced classification algorithms (e.g., regularized logistic regression, random forests or gradient boosting) to account for the various additive and interaction effects among the large number of risk factors under study. This will result in the identification of several sets of relevant risk factors and provide new information to clinicians on how to assess suicide risk in daily practice. In addition, our study will result in useful tools to help clinicians assess actual suicide risk, i.e., estimate the likelihood of suicide attempt following health care visits in different health care settings.
Finally, the effectiveness of the program will be estimated taking into account mortality and hospitalizations due to suicide, as well as the costs derived from the implementation and use of the program, comparing such results with the usual treatment of these patients. Likewise, the implementation of the program in Catalonia will be evaluated using coverage and quality indicators (i.e., indicators that take into account the proportion of the target population that receives the intervention and the degree to which the program complies with the planned protocol of actions).
