Sponsoring Section/Society: ASA-SPES
Session Slot: 8:30-10:20 Thursday
Estimated Audience Size: 60-80
AudioVisual Request: Two Overheads
Session Title: Reliability Models with Incompletely Identified
Causes of Failure
Theme Session: No
Applied Session: Yes
Session Organizer: Yashchin, Emmanuel IBM
Address: IBM, Thomas J. Watson Research Center Dept. of Math. Sciences, Box 218 Yorktown Heights, NY 10598
Phone: 914-945-1828
Fax: 914-945-3434
Email: YASHCHI@WATSON.IBM.COM
Session Timing: 110 minutes total (Sorry about format):
110 total (missing 10 below) First Speaker - 25 minutes (or 25) Second Speaker - 25 minutes Third Speaker - 25 minutes Discussant - 20 minutes Floor Discussion - 5 minutes
Session Chair: Yashchin, Emmanuel IBM
Address: IBM, Thomas J. Watson Research Center Dept. of Math. Sciences, Box 218 Yorktown Heights, NY 10598
Phone: 914-945-1828
Fax: 914-945-3434
Email: YASHCHI@WATSON.IBM.COM
1. Survival Analysis With Partially Resolved Failures
Reiser, Benjamin, Univ of Haifa, Israel
Address: Dept. of Statistics, University of Haifa, Haifa 31905,Israel
Phone: 972-4-8240177
Fax: 972-4-8253849
Email: RSST305@UVM.HAIFA.AC.IL
Flehinger, Betty J., IBM Research
Yashchin, Emmanuel, IBM Research
Abstract: We consider a life testing situation in which systems are subject to failure from independent competing causes. Following a failure, immediate (stage-1) procedures are used in attempt to reach a definitive diagnosis. If these procedures fail to result in a diagnosis, this phenomenon is called masking. Stage-2 procedures, such as failure analysis or autopsy, provide definitive diagnosis for a sample of the masked cases. We show how stage-1 and stage-2 information can be combined to provide statistical inference about (a) survival functions of the individual risks, (b) the proportions of failures associated with individual risks and (c) probability, for a specified masked case, that each of the masked competing risks is responsible for the failure.
2. Product Design With Incomplete Failure Data
Doganaksoy, Necip, General Electric Corporate R&D
Address: P.O. Box 8, K1-4C43, Schenectady, NY 12301
Phone: 518-387-5319
Fax: 518-387-5714
Email: doganaksoy@crd.ge.com
Abstract: Proper identification of failure modes (and mechanisms) plays a central role in design of high reliability products. In practical situations, such knowledge and understanding of failure modes evolves throughout the product life cycle. It then becomes essential that this information be used properly to guide design changes as well as design of new products. This talk will present some practical issues encountered in utilizing incompletely identified failure data and provide recommendations for research in this area.
3. Cause-Specific Survival Analysis With Covariate-Dependent Failure Type Missingness
Ryan, Louise M., Department of Biostatistics, Harvard School of Public Health, and Dana Farber Cancer Institute
Address: 44 Binney Str., Boston, MA 02115-6084
Phone: 617-632-3602
Fax: 617-632-2444
Email: ryan@jimmy.harvard.edu
Andersen, Janet, Department of Biostatistics, Harvard School of Public Health
Abstract: In studies of cancer and many other chronic disease settings, there is often interest in conducting cause-specific survival analyses which focus on time to death from a particular type of disease, rather than overall survival. Problems arise when failure type information is missing for some individuals in the study. We will discuss methods to adjust for such missingness. We will show that particular care is needed when the missingness probability depends on covariates, and will discuss a model that accommodates this situation. Findings will be illustrated with data from a study in colon cancer where older patients were more likely to have missing failure type information.
Discussant: Guess, Frank M. Dept. of Statistics, University of Tennessee
Address: Knoxville, TN 37996-0532
Phone: 423-974-1637
Fax: 423-974-2490
Email: fguess@utk.edu
List of speakers who are nonmembers: None