Background: Human immunodeficiency virus (HIV) remains a global challenge and novel measures for transmission disruption are needed. Contact tracing is limited by reluctance or inability of newly diagnosed individuals to name at-risk contacts. Molecular cluster analysis is mostly used for outbreak investigations, and its role in routine public health activities remains uncertain.
Methods: We conducted a 2-year prospective statewide study in Rhode Island to evaluate integration of HIV cluster analyses into routine contact tracing, by attempting to reinterview all new diagnoses who clustered, notifying them of clustering, and evaluating benefits of this strategy. Clustering was compared between a phylogenetic ensemble versus distance-based HIV-TRACE.
Results: Of 100 new diagnoses during 2021-2022, 52 individuals clustered, of whom only 31% were reinterviewed. Reinterviewing did not improve contact tracing beyond initial interviews, and the study was stopped early for futility. Clustering concordance within the phylogenetic ensemble was high (88%-89%), but lower (74%) for HIV-TRACE. Despite hypothesis rejection, we established a public health-academic partnership, developed a bioinformatics pipeline enabling near real-time cluster analysis, and identified gaps and unique opportunities for intervention.
Conclusions: Attempting to reinterview all statewide new HIV diagnoses in molecular clusters showed no evidence of improving contact tracing. However, a strong academic-public health partnership enabled near real-time, longitudinal integration of molecular cluster analysis into routine public health activities, and identified barriers and opportunities tailoring data-driven approaches to unique individual and community characteristics, guiding future work on optimal use of molecular epidemiology to disrupt HIV transmission.