Statistics, Epidemiology and Data Management

The SED Core is a resource, on both classic and innovative research design and statistical analysis strategies. Methodological demands and design options for high-impact HIV social and behavioral research are becoming increasingly complex and diverse as the field evolves. Innovative approaches to design and statistical analysis are critical to address key research and implementation challenges of the HIV epidemic that as outlined in the US NHAS [1], apply to addressing disparities [2] and implementation gaps in resource-rich [3,4] and resource-limited settings [5-9]. Additionally, important new trials have demonstrated efficacy with the Test and Treat paradigm [10,11], pre-exposure prophylaxis (PrEP) [12], use of early antiretroviral therapy (ART) with infected people for prevention of further transmission [13], integration of TB and HIV treatment [14,15], and topical microbicides [16], In light of these recent developments as well as expected advances resulting from other major trials underway or planned, the need for rigorous studies of the real-world effectiveness of interventions as they are scaled-up (often in combination with other interventions) has never been greater. Likewise, well-considered application of comparative effectiveness research, cost-effectiveness analysis, and mathematical modeling inform policy formulation and potential for integration into practice. SED Core leadership and faculty are internationally recognized experts in the design and conduct of clinical [17-22] and behavioral [23,24] intervention trials - both in the US and in low- and mid-resource countries - and the sophisticated determinants studies that inform them. Examples of HIV Center projects developed with support from the Core include studies on the acceptability and use of new HIV prevention technologies, e.g., female condom, microbicides [23-27]; gender-specific prevention trials [23,24,28]; and determinants of risk, seroprevalence, and motivators for HIV testing among high risk men who have sex with men (MSM). The Core has supported the development and data management of innovative interventions for HIV prevention among severely mentally ill populations in Brazil and for adherence to ART in New York [29] and South Africa [30,31]. A range of studies on children, adolescents and families, including large longitudinal cohorts of perinatally infected children in New York City (NYC) [32,33] and interventions for high risk adolescents in NYC [34], New York State (NYS) [35], and South Africa [36], have relied on the analytical and data management foundations provided by the SED Core. SED Core faculty have themselves conducted studies on HIV incidence and prevalence estimations [37-40]; factors associated with HIV testing [40,41] and diagnosis [42]; linkage to [43-45] and retention in [46-48] HIV care following diagnosis; timely ART initiation [49,50]; and other HIV care and treatment outcomes [47,48,51-61]. They have been at the forefront of responding to the US NHAS and the implementation of effective HlV-prevention and treatment strategies, with studies on disparities among persons living with HIV [62], policy implications [63], mathematical modeling [41,64], comparative effectiveness design [17,18,20], and cost effectiveness modeling [65]. Thus, through the SED Core, HIV Center Investigators will have access to the expertise needed to further advance the field with cutting edge study design and statistical analysis, epidemiology, and data management techniques. Our established consultation and communication process (see Section 5) represents a successful integration of effort between Center investigators and SED Core members. Consultation sessions held by the Core in the previous cycle are provided in Appendix E.l. Problem resolution is a key indicator of SED Core performance. A few brief examples illustrate how the Core works in this capacity. ? The Core worked with both Claude Ann Mellins, Ph.D., on Adapting CHAMP+ for South Africa: Supporting HIV Infected Youth and Families [36] and with Robert Remien, Ph.D., on A Multimedia Social Support Intervention: Adherence to HIV Care In South Africa [30], to carefully identify the general goals of these NIH-supported pilot intervention studies. Pilot studies must determine auxiliary parameters such as prevalences of binary outcome measures and standard deviations for key continuous outcomes so that a future large study can be properly and realistically planned. Effect size estimation, however, is not a proper goal for a pilot study. The SED Core provided effective support to Center investigators to clearly state these key points in the submitted proposals that resulted in awarded studies. The SED Core's approach and recommendations were affirmed by the revised guidelines for RS4 pilot grant applications. 466 statistics. Epidemiology, and Data Management Core Remien, Robert H. ? Dr. Alex Carballo-Dieguez's study. Predictors of HIV Risk Behavior Among Men in Argentina [66] revealed unusually high estimates of HIV incidence using Respondent Driven Sampling and Serological Testing Assays for the detection of Recent HIV Seroconversion (STAPHS) in a sample of 500 MSM. SED Core members developed adjustments for specificity of recency determinations that were much more realistic.