Harnessing Big Data to end HIV

The introduction of highly successful HIV prevention and treatment options, including treatment as prevention (TasP) and pre-exposure prophylaxis (PrEP), led to many national and global plans to eliminate HIV infection (USDHHS, 2019). However, the current progress is too slow in pace to achieve these goals. For example, the number of new infections in the U.S. has not decreased substantially since 2014 (remaining at ∼38,000/year), with a rising trend among several subpopulations (e.g., individuals aged 25–34 years, Hispanic or Latino MSM) (Sullivan et al., 2021). Nearly 15% of people with HIV (PWH) do not know their serostatus. In addition, about 20% of the new diagnoses nationwide remain late diagnosis (e.g., being diagnosed with AIDS at the time or within 12 months of initial diagnosis) (Beyrer et al., 2021; Sullivan et al., 2021). The unequal risks of HIV infection among key populations, combined with increasing rise of sexually transmitted diseases like Syphilis makes impacts HIV incidence and makes ending the HIV epidemic unattainable (CDC; Lyons et al., 2020).

Despite substantial strides in HIV prevention and treatment at global, national and regional levels, some significant knowledge gaps remain in our understanding of the various individual, contextual, and structural factors that may influence the HIV treatment cascade outcomes. These gaps include the lack of robust evidence on the roles of these factors and their interplays, lack of integration of diverse sources of real-world data (e.g., electronic health records [EHR] data, social media data, geospatial data, metadata, U.S. census data, and other publicly available data), and lack of progress in translating the research findings to clinical and public health practices. Besides the existing barriers to HIV prevention, treatment, and care, the global COVID-19 pandemic has also imposed unprecedented challenges to HIV management and care (Qiao et al., 2021) and at the same time presents a natural opportunity to explore, document, and quantify expected or unexpected consequences of a major world-wide disruption in HIV treatment and care.

The increasing availability of large and complex data sets for HIV research has presented wonderful opportunities to address these gaps and challenges by data science approaches. We hope that this special issue will offer a timely and unique avenue to review and synthesize exciting research that apply Big Data (e.g., EHR, social media data, genomics data, geospatial data) and innovative data science methodologies (texting mining, natural language processing, deep learning, machine learning, artificial intelligence) to identify gaps in rare, unseen, and otherwise undiscovered biomedical, behavioral, social patterns/determinants that shed light on HIV acquisition, transmission, the development of comorbidities, and long-term viral load control across the HIV treatment continuum.