PEC0511
Share
 
Title
Development and application of a data-driven approach to identify suspected HIV-1 recent infection clusters for public health response in Vietnam, 2019
Presenter
Alexandra Ernst
Authors
A. Ernst * (1), T.M.T. Tran (2), T.T.H. Bui (2), P.D. Manh (3), K. Curran (4), C. Said (1), S. Welty (1), V.N. Nguyen (3), T.T.H. Nguyen (3), G. Rutherford (1), V.H. Son (3)
Institutions
(1) University of California, Institute for Global Health, Global Strategic Information, San Francisco, United States, (2) Centers for Disease Control and Prevention, Division of Global HIV and Tuberculosis, Center for Global Health, Hanoi, Vietnam, (3) Ministry of Health, Vietnam Administration for HIV/AIDS Control, Hanoi, Vietnam, (4) Centers for Disease Control and Prevention, Division of Global HIV and Tuberculosis, Center for Global Health, Atlanta, United States

BACKGROUND: The greatest HIV burden in Vietnam is among key populations in specific geographic regions. To reach epidemic control, Vietnam has implemented rapid test for recent infection (RTRI) within the past twelve months into routine HIV testing services in 10 provinces. Identifying suspected clusters with possible high rates of transmission prompts public health investigation and response.
DESCRIPTION: We identified multiple types of clusters: transmission clusters (via contact tracing, molecular analysis), time-space clusters (geographic grouping of cases), or a single recent infection among priority populations. Stakeholders developed and applied several time-space cluster definitions to determine a threshold for response.
LESSONS LEARNED: From January-August 2019, 402 RTRI-recent HIV cases were identified in 10 provinces. We analyzed data by month, geography, and multiple cluster definitions. With most new diagnoses in two urban provinces, stakeholders agreed that cluster definitions be geographic-specific and field-tested with data.
Using a definition of '¥2 recent cases monthly, provincial-level analysis yielded 48 clusters (average size: 9.27 RTRI-recent cases). District-level analysis, using the same threshold of '¥2 recent cases monthly, identified 40 clusters (average: 4.42 RTRI-recent), with the majority of clusters in Hanoi (n=8, 7 districts) and Ho Chi Minh City (HCMC) (n=30, 12 districts). However, district-level analysis significantly reduced the number of clusters in non-urban provinces, suggesting the need for distinct definitions in high-burden provinces like Hanoi and HCMC. In HCMC, using a definition of '¥6 recent cases monthly, district-level analysis yielded 10 high-volume clusters (average: 9.30 RTRI-recent, 6 districts).
Considering limited resources for response, we developed and applied the following cluster definitions:

  1. Hanoi: '¥2 RTRI-recent cases per district monthly;
  2. HCMC: '¥6 RTRI-recent cases per district monthly; and
  3. National: '¥2 RTRI-recent cases per province monthly.

CONCLUSIONS: Using RTRI data, stakeholders developed and field-tested geographic-specific time-space cluster definitions. As health authorities investigate suspected clusters, it is critical to review data routinely in real-time to monitor cluster numbers and trends. Future evaluations may consider time-space clusters by facility. We recommend that initial cluster definitions be validated using statistical or spatial analysis software and revised accordingly, with input from stakeholders on their capacity to identify and respond to clusters.