The matching of healthcare cost models to the analytic objectives and characteristics of the data available to a study requires caution. The study results and interpretation can be heavily dependent on the choice of model with a real risk of spurious results and conclusions.
EVA PAGANO 4OBJECTIVE -Hyperglycemia is a common condition in hospitalized patients. The aim of this study was to investigate the relationships between glycemia upon admission and mortality in a heterogeneous group of adult patients.RESEARCH DESIGN AND METHODS -The 3-year records released from a general hospital were associated with a plasma glucose dataset of its general laboratory. A matched case-control study was implemented (3,338 case-control subject pairs). All-patient refined diagnosis-related groups and the relative risk of death were the matching criteria. A multivariate conditional logistic regression model was used to evaluate the associations between death and glycemia.RESULTS -Higher in-hospital mortality was associated with hyperglycemia or hypoglycemia, whereas lower risk was observed for values between 78 and 101 mg/dl. CONCLUSIONS -Our data confirm the relation between glycemia upon admission and mortality and suggest that slightly increased or decreased plasma glucose can be linked with increased mortality risk.
This study shows, with reference to a real multicenter trial, that center information cannot be neglected and should be collected and inserted in the analysis, better in combination with one or more random effect, taking into account in this way also the heterogeneity among centers because of unobserved centers characteristics.
We aim at evaluating how data-mining statistical techniques can be applied on medical records and administrative data of diabetes and how they differ in terms of capabilities of predicting outcomes (e.g. death). Data on 3,892 outpatient patients with a diagnosis of type 2 diabetes from the San Giovanni Battista Hospital in Torino. Six statistical classifiers were applied: Logistic regression (LR), Generalized Additive Model (GAM), Projection pursuit Regression (PPR), Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA), Artificial Neural Networks (ANN). All models selected the same subset of covariates. ANN is the model performing worse, whereas simpler models, like LR, GAM and LDA seem to perform better. GAM is associated with a very small misclassification rate. The agreement in predicting individual outcomes among models is 0.23 (SE 0.06, Kappa). Monitoring on the basis of patients' characteristics is highly dependent from the statistical properties of the chosen statistical model.
The implementation of locally adapted GL on VTE prophylaxis may lead to a benefit in terms of both costs and effects, especially for the highest-risk patients.
The main advantage of the ER model is to unify these approaches in a unique framework, where the estimation of the cut-offs and the production of the prediction rules are performed simultaneously for a continuous response variable. The final model can thus be analysed at any desiderate quantile of the cost distribution, avoiding the need of pre-specifying any cut-off.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.