Format

Send to

Choose Destination
Sex Transm Dis. 2004 Oct;31(10):588-95.

The impact of inaccurate reporting of condom use and imperfect diagnosis of sexually transmitted disease infection in studies of condom effectiveness: a simulation-based assessment.

Author information

1
National Center for Birth Defects and Developmental Disabilities, Centers for Disease and Control and Prevention, Office of the Director, Mail-Stop E-87, 1600 Clifton Road, Atlanta, GA 30333, USA. ODevine@cdc.gov

Abstract

BACKGROUND AND OBJECTIVES:

The effectiveness of condoms in reducing sexually transmitted disease (STD) infection risk has been debated in the face of equivocal epidemiologic evidence. We assessed the potential magnitude of bias in condom effectiveness studies resulting from inaccurate reporting of condom use behavior and misdiagnosis of STD infection status.

GOAL:

The goal of this study was to illustrate the magnitude of bias in condom effectiveness studies in the presence of inaccurate condom use reporting and diagnostic misclassification.

STUDY:

We used probabilistic simulations to mimic plausible outcomes for hypothetical prospective and retrospective condom effectiveness studies subject to both inaccurate reporting of the participants' true condom use and diagnostic error. The simulations were conducted by generating a series of binomial (yes, no) random variables corresponding with STD infection status and accurate diagnosis of infection.

RESULTS:

The simulation results illustrate that failure to address reporting and diagnostic errors can lead to a substantial bias in studies of condom effectiveness. This bias resulted in a roughly 25% to 30% reduction in the probability of detecting a true 2-fold reduction of infection risk resulting from using condoms.

CONCLUSION:

Inaccurate reporting of condom use and reliance on imperfect diagnostic tests can substantially bias observable measures of condom effectiveness. The potential for these biases should be addressed in the design and analysis of effectiveness studies.

PMID:
15388995
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Wolters Kluwer
Loading ...
Support Center