Dr. Hughes' research program focuses on statistical methods related to the design and analysis of clinical trials and observational studies, particularly in the context of HIV research. He is well-known for his work on the development of methods for evaluating surrogate markers using meta-analysis of results from a number of clinical trials. An an important early application of this work was the assessment of the value of treatment-induced changes in CD4 cell counts and viral load as surrogate markers for the development of AIDS.
Dr. Hughes is the Principal Investigator of the Statistical and Data Management Center (SDMC) of the AIDS Clinical Trials Group (ACTG) funded by the National Institutes of Health. The ACTG undertakes numerous studies of interventions for the treatment of people infected with HIV, including co-infections such as tuberculosis and hepatitis B and C, both in the United States and internationally. He was previously the Principal Investigator for the SDMC of the Pediatric ACTG and has ongoing HIV research interests in the treatment of children infected with HIV, and the prevention of mother-to-child transmission of HIV. A particular focus of this work concerns studies evaluating the impact of prophylaxes for the prevention of transmission on restricted future treatment options for the mother due to viral resistance.
Dr. Hughes' statistical research is strongly motivated by problems encountered in his collaborative research. Ongoing research interests include the design and analysis of phase I/II clinical trials, and of bridging trials from one population to another population. One example concerns the design of studies that seek to identify the dose of a drug that maximizes the success of treatment (i.e. is both safe and effective) rather than just find the maximal tolerated dose. A second example involves the development of methodology for the design of dose-finding bridging studies in "special populations", such as infants or pregnant women, which are efficient in identifying a dose which minimizes the proportion of people who might be exposed to sub-therapeutic doses.
Another area of interest concerns statistical methods for the design and analysis of studies evaluating diagnostic and other biomarkers. For example, he is interested in the problem of comparing the performance of two biomarkers for determining when to start treatment in HIV-infected subjects, or when to change treatments. This work is motivated by the need to identify less expensive and more practically accessible technologies for addressing these questions in resource-limited countries.