Spatiotemporal Dataset and R-INLA Pipeline: Adolescent Suicide and Multidimensional Poverty in Colombia (2015–2024)

DESCRIPTION:
This dataset provides a high-resolution spatiotemporal characterization of adolescent suicide mortality (10–19 years) in Colombia for the period 2015–2024. The research utilizes a fully Bayesian disease mapping framework to overcome the statistical instability inherent in small-area health data, particularly in early adolescence.
The data is organized into a balanced department-year panel (1,320 observations) incorporating:
Mortality Counts: Derived from official medico-legal records (INMLCF), harmonized by age, sex, and mechanism of death.
Population at Risk: Intercensal projections from the National Administrative Department of Statistics (DANE).
Socioeconomic Covariates: The Multidimensional Poverty Index (MPI), pre-processed using Kalman smoothing to ensure temporal completeness.

NOTABLE FINDINGS:
The analysis identifies a steep developmental risk gradient (RR = 2.85 for late vs. early adolescence) and persistent geographic hotspots in departments such as Arauca, Putumayo, and Quindío. Notably, the data reveals a robust inverse association between multidimensional poverty and registered mortality (RR = 0.82), suggesting significant regional differences in death-investigation capacity and social fragmentation.

FORMAT/CONTENT
The documentation and code are organized within the Supplementary_Material directory, distributed into five specialized modules, each containing its respective raw inputs and derived outputs:
• 00_population: Contains official DANE intercensal projections (2015–2024).
• 01_MPI: Includes annual Multidimensional Poverty Index (MPI) data, including the backcast series for 2015–2017.
• 02_panel: Houses the individual-level mortality records from INMLCF and the final balanced analytic panel (1,320 cells).
• 03_Spatial: Contains official administrative boundaries and the topological adjacency graph used for spatial smoothing.
• 04_bayes: Includes the master modeling scripts and the automated table generation system for all supplementary results