Association for Public Policy Analysis & Management

Although New York State (NYS) is a longstanding HIV epicenter, new diagnoses have steadily declined from 6,700 in 2001 to 3,500 in 2014 due to its strong policies and programs to reduce transmission and improve HIV care outcomes. Building on this momentum, Governor Cuomo recently pledged to end AIDS as an epidemic in NYS by the end of 2020 (Ending the Epidemic (ETE)), defined as fewer new HIV infections than HIV-related deaths. Specific ETE policy goals are to reduce annual new infections to 750 and the rate at which newly diagnosed individuals progress to AIDS by 50%.

With finite resources, the NYS Department of Health must identify the most efficient strategies to achieve these objectives and communicate progress to stakeholders; however there are several challenges for data-driven decision-making. HIV testing and care is a complex system, with new infections driving future trends, interventions influence epidemic outcomes differently depending on how they target transmission-related behavioral or biological factors, and HIV-infected individuals have different risk behaviors and healthcare needs as they age. Some ETE strategies, such as “data-to-care” programs that use surveillance data to identify out-of-care individuals or pre-exposure prophylaxis antiretroviral medications for high-risk uninfected individuals, have limited long-term data from real-world settings because they are novel or not yet scaled up beyond clinical trials and pilot interventions. Other ETE strategies, such as supportive housing, have limited quantitative evaluation data. Finally, policies are often implemented simultaneously, making it difficult to isolate their effects or assess trade-offs.

We used system modeling to project the impact of ETE strategies on HIV prevalence, percent unaware cases, new infections, new diagnoses, and deaths. This computational method is commonly used in engineering but only recently valued as a “policy informatics” tool to assist with decision-making for complex health policy issues. We developed a simulation model of New York’s HIV testing and care system, with additional structure to represent three critical ETE programs: supportive housing; pre-exposure prophylaxis; and linkage to and retention in care. The model also stratifies the population by risk group. Data inputs come from CDC estimates of new infections, HIV surveillance data, early program evaluation data, survey data, literature, and expert consultation.

Preliminary results suggest that: 1) if current trends continue, significant strides towards achieving the ETE objectives can be made, but new and existing policies must be ramped up significantly to approach the 2020 milestones; 2) linking all newly diagnosed individuals to care and expanded housing improve outcomes but will likely not meet ETE goals without other policies in place; 3) wide-scale adoption and adherence to pre-exposure prophylaxis appears necessary to significantly reduce the rate of new infections, but covering individuals at highest risk and maintaining their adherence are important implementation challenges; and 4) it is critical to focus efforts on younger men who have sex with men, who have the highest transmission rates. More broadly, this analysis demonstrates how systems modeling can organize and synthesize diverse data sources meaningfully for real-time decision-making and stakeholder communication.