Characteristics of HIV Time-Space Alerts in the United States, 2018-2021

Abstract

Background: HIV time-space cluster detection is routinely conducted by CDC and health departments to identify counties with elevated diagnoses compared with historical levels. These analyses, originally applied to all diagnoses and persons who inject drugs (PWID), generate "alerts" for review that may indicate clusters or outbreaks. The reoccurrence of alerts has not been previously described nationally.

Setting: 50 U.S. states and the District of Columbia.

Methods: Quarterly, during 2018-2021, the number of HIV infections diagnosed during the preceding 12 months among each group of interest (all people ["overall alerts"], PWID, men who have sex with men [MSM], or MSM who inject drugs) was compared with the annual mean for that group from the preceding 36 months. An alert was generated if the number of diagnoses was >2 standard deviations and >2 diagnoses above the baseline mean. For counties with initial alerts during 2018-2019, reoccurrences of alerts during any of the following 8 quarters were assessed.

Results: During 2018-2021, compared to counties with alerts among PWID (n=154), 5 times as many counties had overall alerts (n=824) and nearly 3 times as many counties had alerts among MSM (n=445). Reoccurrences of alerts in ≥2 subsequent quarters were lowest among MSM and MSM who inject drugs (10% and 19% of counties respectively).

Conclusions: Applying time-space cluster detection criteria to populations beyond PWID results in numerous alerts nationally, many of which represent transient increases. Additional work to refine time-space cluster detection criteria, or to prioritize additional follow-up or investigation, is necessary.

Keywords: HIV; cluster detection and response; outbreak detection; spatiotemporal; surveillance; time-space.