Group testing in presence of classification errors

Diwakar Gupta, Regina Malina

Research output: Contribution to journalArticlepeer-review

12 Scopus citations

Abstract

We modify Dorfman's and Sterrett's group testing protocols to make them suitable for testing blood samples for the presence of HIV antibodies, as well as for many industrial applications, when false negatives cannot be tolerated. We first propose that test kit sensitivity be increased to nearly 100 per cent by altering the reactive versus non-reactive threshold. Subsequently, group and repeat testing are used with a careful selection of group size and the number of times a test is repeated, in order to maximize efficiency while keeping the false positive predictive value (FPPV) within a specified limit. Numerical calculations show that our testing protocol is efficient, has low procedural complexity and keeps both types of classification errors below specified tolerance limits.

Original languageEnglish (US)
Pages (from-to)1049-1068
Number of pages20
JournalStatistics in Medicine
Volume18
Issue number9
DOIs
StatePublished - May 15 1999

Fingerprint

Dive into the research topics of 'Group testing in presence of classification errors'. Together they form a unique fingerprint.

Cite this