|
Activity Number:
|
281
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306232 |
|
Title:
|
A Probability Model for Recent Tb Transmissions in Genetic Clusters
|
|
Author(s):
|
Nong Shang*+
|
|
Companies:
|
Centers for Disease Control and Prevention
|
|
Address:
|
1600 Clifton Road, Atlanta, GA, 30333,
|
|
Keywords:
|
genetic cluster ; recent transmission ; dependent Bernoulli trial ; family history model
|
|
Abstract:
|
DNA fingerprinting of Mycobacterium tuberculosis is used to link tuberculosis into genetic clusters. Researchers have used such clusters as surrogates of recent TB transmissions. Most previous researchers have focused on patient-level characteristics to predict recent transmission. However, it is also important to consider the predictive power of cluster-level characteristics. Most of practical questions at the cluster level can be answered through modeling the probability that a new case is associated with recent transmission. Family history models were applied. Each of the parameters in the family history model is further modeled by the features of the clusters, resulting in two simultaneously estimated regression equations with observations as the sums of dependent Bernoulli random variables. Numerical methods were developed to estimate the parameters and to find the "best" models.
|
