Geospatial Research at the Surveillance Research Program of National Cancer Institute: Evaluation of Spatio-Temporal Projection Methods for Cancer Incidence
Background: A study is undertaken to evaluate the spatio-temporal projection models that is applied by the American Cancer Society to predict the number of new cancer cases.
Methods: Adaptations of a model that has been used since 2007 are evaluated. Modeling is conducted in three steps. In Step I, ecologic predictors of spatio-temporal variation are used to estimate age-specific incidence counts for every county in the country, providing an estimate even in areas with missing data in specific years. Step II adjusts the Step I estimates for reporting delays. In Step III, the delay-adjusted predictions are projected four years ahead to the current calendar year. Adaptations of the original model include updating covariates and evaluating alternative projection methods. Residual analysis and evaluation of five temporal projection methods are conducted.
Results: The differences between the spatio-temporal model-estimated case counts and the observed case counts for 2007 are below 1%. After delays in reporting of cases are considered, the difference is 2.5% for women and 3.3% for men. Residual analysis shows no significant pattern that suggests the need for additional covariates. The Vector Autoregressive model is identified as the best temporal projection method.
Conclusions: The current spatio-temporal prediction model is adequate to provide reasonable estimates of case counts. To project the estimated case counts ahead 4 years, the Vector Autoregressive model is recommended to be the best temporal projection method that produces the estimates closest to the observed case counts.
