The role of generalized linear models in handling cost and count data

Scientists from nursing and allied health disciplines frequently examine data with complex distributions. Key examples include data on cost that typically are skewed, and count data like the number of hospitalizations that regularly have greater variation than expected and a majority of observations at zero. Common approaches to handling complex data involve transformations that can interfere with interpretation, or force-fitting of data into linear or logistic regression. In this article, worked examples of generalized linear models, which allow for flexibility in the distribution of data, involving cost and count outcomes, are presented to help expose researchers to their nuances.