|
Abstract:
|
Low birthweight (LBW) is a leading cause of newborns' mortality and morbidity. Many risk factors of LBW can be linked to the maternal socioeconomic status, which in turn contribute to large racial disparities in LBW incidence. Here, we employ Bayesian statistical models to analyze the LBW incidence rate in Pennsylvania (PA) counties by race/ethnicity. While leveraging spatial structure can help improve the precision of our estimates, we must be cautious to avoid letting the model overwhelm the information in the data and produce spurious conclusions. As such, we first develop a framework by which we can measure and control the informativeness of our spatial model. After demonstrating the properties of our framework via simulation, we analyze the LBW data from PA and examine the extent to which the commonly used conditional autoregressive model can lead to over-smoothing. We then reanalyze the data using our proposed framework and highlight its ability to detect (or not detect) evidence of racial disparities in LBW incidence.
|
