|Thursday, February 20|
|PS1 Poster Session I & Opening Mixer||
Thu, Feb 20, 5:15 PM - 6:45 PM
Methods of developing and validating a predictive model (302823)*Yu-Hui Chang, Mayo Clinic
Keywords: Model selection, model validation, predictive model
Developing a predictive model is usually seen in clinical research, and performing validation of a predictive model plays an essential role to ensure the combination of factors included in a model will have great predictive ability. This presentation will use the data from an emergency department (ED) to predict the probability of admission to hospital for patients with cancer. The ultimate goal is to use key elements that can capture patients clinical characteristics with the need for hospital admission so that the ED length of stay can be reduced. The potential factors will include the chief complaint, age, emergency severity index, vital signs taken from triage and the results from lab tests. We first investigate whether the chief complaint and other measurements taken from triage only will accurately predict if a patient will be admitted to the hospital, and evaluate if adding the results from lab test would improve prediction. The results from different model development techniques will be reported. Each of them will be internally validated by the bootstrapping method, and the performance of the prediction from the validation will also be presented and compared.