Spatio-temporal modelling using B-spline for disease mapping: analysis of childhood cancer trends

• Author(s) / Creator(s)

  • Torabi, Mahmoud
  • Rosychuk, Rhonda J.

• To analyze childhood cancer diagnoses in the province of Alberta, Canada during 1983-2004, we construct a generalized linear mixed model for the analysis of geographic and temporal variability of cancer rates. In this model, spatially correlated random e®ects and temporal components are adopted. The interaction between space and time is also accommodated. Spatio-temporal models that use conditional autoregressive smoothing across the spatial di- mension and B-spline over the temporal dimension are considered. We study the patterns of incidence rates over time and identify areas with consistently high rate estimates as areas for further investigation. We apply the method of penalized quasi-likelihood to estimate the model parameters. We illustrate this approach using a yearly data set of childhood cancer diagnoses in the province of Alberta, Canada during 1983-2004.

• Date created

  2011

• Subjects / Keywords

  • Geographic epidemiology
  • Conditional autoregressive
  • Disease mapping
  • Generalized linear mixed model
  • Penalized quasi-likelihood

• Type of Item

  Article (Published)

• DOI

  https://doi.org/10.7939/R3HD7NS4D

• License

  Attribution-NonCommercial-NoDerivatives 3.0 International