Studies focused on the implementation of HIV prevention are sparse and have a wide range of objectives. Some studies describe the barriers and facilitators to implement evidence-based interventions (EBIs), while others use advanced implementation trial designs and assess multiple implementation research (IR) outcomes. There is a pressing need to identify current and ongoing research gaps and new directions in IR in the field of HIV prevention. In this symposium, we present the use of machine learning and text mining methodologies for the purpose of searching and extracting the results of systematic and scoping reviews. An example of this method is presented from a scoping review of NIH-funded HIV studies between 2013 and 2018. The results describe the proportion and characteristics of HIV-related IR studies funded by NIH and include the IR outcomes and trial designs, types of HIV research, and the stage of IR. Our study points to the need to increase the amount and rigor of IR on effective HIV interventions, with the goal to speed translation and increase their reach to prevent and treat HIV. Automating the review process with machine learning methods can facilitate rapid re-evaluation of the characteristics of HIV IR studies as they are funded to track trends and inform research foci.
The first paper, “Using a Combined Text Mining and Human Coding Approach to Accelerate Scoping Reviews of NIH-funded HIV Implementation Studies,” describes the use of machine learning, text mining, and rule-based heuristics to screen the titles and abstracts of NIH-funded grants to identify HIV-related interventions focused on addressing the HIV continuum and determine whether they are IR studies.
The second paper, “Implementation Research Methods in Recently-Funded HIV Studies by the NIH: A Deep Dive into their Aims, Study Design, Primary Outcomes, and Stage of Investigation” examines and characterizes the typology of IR of recently funded HIV-related studies by NIH with a focus on stage of IR and IR outcomes.
The third paper, “Landscape of NIH Funding for Implementation Research in HIV/AIDS: Establishing a Baseline and Informing New Opportunities” describes the spectrum of HIV-related interventions that used IR, including information on the target population, stage of the HIV care continuum, types of interventions, delivery methods, and stage of IR.