Changes to the Sample Design and Frame of a National Survey of People Living with HIV

• Activity Number: 301 - Weighting Adjustments
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Christopher Johnson* and Ronaldo Iachan and Richard Lee Harding and Adam Lee and Tonja Kyle and Jason Craw and Linda Beer and Roy L. Shouse
• Companies: CDC and ICF International and ICF and ICF and ICF and CDC and CDC and CDC
• Keywords: population-based surveys ; health surveys ; weighting ; nonresponse ; post-stratification ; HIV

Abstract

Conducted since 2007 in 17 states with funding from the Centers for Disease Control and Prevention, the Medical Monitoring Project (MMP) provides national and local estimates of the experiences, behaviors, access to medical care and HIV medications, and needs of U.S. adults living with HIV. Initially a three-stage cluster sample of states, clinics, and patients, in 2015 MMP began to sample persons with a diagnosis of HIV reported to state surveillance registries, regardless of medical care status. We describe changes to the sampling frames and sample design, then outline two weight adjustment methods (propensity weighting vs. weighting class models) that were compared for two stages of non-response (sequential adjustments for non-contact and nonresponse vs. a single adjustment for nonresponse). We explain the use of updated surveillance registry data for non-coverage adjustments and for post-stratifying estimates to reduce bias. We describe how estimates available for this expanded study population differ from past MMP estimates and complement information from CDC'S National HIV Surveillance System.