Online Program

Sample Surveys of HIV-infected Patients: Applications of Real Time Sampling Methods

*Ronaldo Iachan, ICF Macro 
Tonja M Kyle, ICF Macro 
Beer Linda, CDC 
Christopher Johnson,  
Eduardo Valverde, CDC 
Kathleen Brady, Philadelphia Department of Public Health 
Jacek Skarbinski, CDC 
Kim Brinker, CDC 

Keywords: HIV, Real time sampling,

Any survey of the HIV-infected patient population can encounter many of the challenges typical of studies of hard-to-reach populations. HIV-infected patients often include hard-to-reach subpopulations including racial, ethnic, sexual, linguistic, and cultural minorities; illicit drug users; as well as immigrant, indigenous, mobile, migrant and homeless populations. In addition, the stigma often associated with HIV can make it difficult to achieve cooperation with data collection. For the past five years, the Centers for Disease Control and Prevention (CDC) has funded the Medical Monitoring Project (MMP), which is designed to obtain locally and nationally representative data on behaviors and clinical outcomes from a national probability sample of adult patients in care for HIV infection. MMP’s patient sampling method involves sampling from patient lists obtained from participating HIV medical care facilities. Patient recruitment often presents enormous challenges that have been reflected in low response rates and resource-intensive efforts to locate these patients. In an effort to improve data timeliness and response rates, particularly among the subgroups that may be hardest to reach, CDC tested real-time sampling (RTS) as a supplemental sampling methodology in selected facilities in one project area. This presentation describes the methods and results of an RTS pilot conducted in the 2011 data collection period. We discuss a wide range of issues in the following areas. Sampling design: frame construction and stratification, sampling stages, sampling units, measures of size, sample allocation and selection, and sampling patients on site Data collection: engaging facilities and patients, improving cooperation Weighting: sampling weights, non-response and multiplicity adjustments The presentation summarizes the promise as well as the challenges of implementing RTS methods to survey HIV-infected patients and other hard-to-reach populations.