Novel threat detection methodology to detect HIV outbreaks in Washington

Abstract/summary Despite recent significant advances in biomedical HIV prevention, HIV diagnosis rates in Washington State have been stable over the past several years, with a rate of 5.3 per 100,000 in 2014 and 5.4 per 100,000 in 2018. Identifying and responding to outbreaks is a cornerstone of public health practice; rapid public health action in such cases may decrease the transmission of HIV and other infectious diseases. While the Centers for Disease Control and Prevention (CDC) require that local health departments conduct analyses to identify HIV molecular clusters and/or time-space clusters of new HIV diagnoses, no recent HIV outbreaks have been identified by methods other than local health department staff or providers noticing increases in cases in specific areas or populations. Working in collaboration with the Washington State Department of Health (DOH), we propose the following aims to evaluate a new methodology to identify HIV outbreaks: (1) To adapt rapid threat detection methodology that was developed by our team to identify outbreaks in gonorrhea to instead identify spatial- temporal clusters of HIV cases, or changes in previously identified clusters in Washington State, based on retrospective data from multiple counties, and (2) To compare the performance of this methodology to the performance of current cluster identification methods being used by the Washington DOH, including CDC recommended strategies for molecular and time-space cluster identification. We will determine whether this method could have identified outbreaks sooner than the routine, CDC-recommended approaches, including an HIV outbreak that occurred among injection drug users in King County, WA in 2018. We propose an innovative approach; the methods are adapted and expanded from statistical change detection and spatiotemporal approaches to identify the time and location of the clusters rapidly. If our approach is found to provide added value to ongoing work at Washington DOH, we will help implement the approach so that it can be used in conjunction with HIV surveillance in Washington State. While our approach does not rely on HIV genetic sequence information, if successful, it can also be used to assist in identifying more appropriate local criteria to identify HIV molecular clusters. This approach is flexible and can be applied in numerous scenarios, including for other infections such as syphilis or SARS-CoV-2.