JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 294
Type: Contributed
Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:20 AM
Sponsor: Biometrics Section
Abstract - #305457
Title: Application of Zeros Hurdle or Zero-Inflated Models in Populations with Excess at-Risk Zeros: A Case of Nodal Involvement in Breast Cancer
Author(s): Alok Dwivedi*+ and Rakesh Shukla and MB Rao and Sada Nand Dwivedi and SVS Deo
Companies: University of Cincinnati and University of Cincinnati and University of Cincinnati and All India Institute of Medical Sciences and All India Institute of Medical Sciences
Address: Department of Environmental Health, Cincinnati, OH, 45267, United States
Keywords: Zero inflated models ; Hurdle models ; Count data
Abstract:

Count data often have excess zeros in many clinical studies. Zero hurdle (ZH) or Zero inflated (ZI) models have been proposed to handle excess zeros in count outcome. ZH models are more appropriate when population involves at-risk zeros only while ZI models should be preferred when population involves mixture of zeros (not at-risk and at-risk zeros). In clinical studies, we often deal with disease problems making each individual at risk of disease/outcome of interest. Thus, all zeros are at risk zeros indicating use of zero hurdle models. Alternatively, if there are excess zeros then one can define them into two classes (low/no risk or high risk) and thus ZI models can be used. We investigate that ZI models can be exploited appropriately in case of only "at risk" zeros without any loss in the results. We use data on the number of involved nodes in breast cancer patients. Simulations are carried out to compare the ZH and ZI models using different regression coefficients in case of at-risk zero population. Our proposed approach indicates that ZI models have an added merit of estimating and identifying high risk zeros over hurdle models without any loss of information.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program




2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.