Giancarlo Sal y Rosas
The Partner Notification Study was conducted in Seattle between September 29th of 1998 to March 7th of 2003. Participants in the study were heterosexual men and women that tested positive for gonorrhea or Chlamydia. They were randomized to expedited partner treatment or standard partner referral. Patients in the intervention group were offered vouchers for medication to give to their sex partners, or if they preferred, staff members contacted their partners and provided them with medication without a clinical examination. Patients assigned to the standard partner referral were advised to refer their partners for treatment and were offered assistance-notifying partners. The primary outcome was persistent or recurrent gonorrhea or Chlamydia infection in the original participant at 3 and 19 weeks after treatment (participants had only one follow up visit).
A particular feature of this study is that the tests that define the outcome (infection of gonorrhea and Chlamydia) have perfect specificity, but their sensitivity was between 0.85-0.95. How this misclassification could affect the results of the study? Can we adjust our analysis to handle this outcome misclassification? I will show my work trying to answer these and other related questions.
