| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 346.18 KB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
In this study we consider logistic regression models to predict
mutation carriers in family members affected by Brugada Syndrome.
This Syndrome is an inherited cardiopathy that predisposes individuals
without structural heart disease to sudden cardiac death. We focused
on five electrocardiographic markers, which have been explored as good
discriminators between carriers and non-carriers of the genetic mutation
responsible for this disease. Logistic regression models which combine
some of the five markers were investigated. Our objective was to assess
the predictive ability of these models through internal validation procedures.
We also applied shrinkage methods to improve calibration of the
models and future predictive accuracy. Validation of these models, using
bootstrapping, point to some superiority of two models, for which fairly
good measures of predictive accuracy were obtained. This study provides
confidence in these models, which offer greater sensitivity than the usual
screening by detecting a characteristic pattern in an electrocardiogram.
Description
Keywords
Bootstrapping Logistic Regression Ridge Regression Calibration Discrimination
Citation
Publisher
B. Murgante et al.