RASCO

"RASCO" is an R package to address the challenge of spatial confounding in cancer disease mapping (DM). Spatial confounding, where latent correlations between spatial random and fixed effects lead to misleading interpretations, is a significant concern in cancer epidemiology. RASCO introduces a solution by implementing restricted spatial regressions, which either project the latent effect onto the orthogonal space of covariates or displace spatial locations to mitigate the impact of spatial confounding.

Key Features and Functionalities:

- Alleviation of Spatial Confounding: RASCO is designed to tackle spatial confounding by employing restricted spatial regressions, ensuring more reliable interpretations of spatial patterns in cancer disease mapping.

- Implementation of Popular Parametric Count Data Models: The package supports parametric count data models such as Poisson, generalized Poisson, and negative binomial models for areal count responses, catering to the diverse needs of spatial DM analysis.

- Quantification of Spatial Association via CAR Model: Spatial association within the data is quantified using the conditional autoregressive (CAR) model, which is integral to understanding the spatial distribution of disease occurrences.

- Bayesian Inference with INLA: RASCO facilitates Bayesian inference, occasionally assisted by integrated nested Laplace approximation (INLA) for accelerated computing. This approach allows for comprehensive statistical analysis and model fitting.