Interpopulation variation in HIV testing promptness may introduce bias in HIV incidence estimates using the serologic...

Area of Science:

• Epidemiology
• Immunology
• Public Health

Background:

• The Serologic Testing Algorithm for Recent HIV Seroconversion (STARHS) estimates HIV incidence based on antibody levels.
• STARHS assumes independence between HIV infection dates and antibody testing dates.
• This assumption may be violated when individuals self-select testing, potentially due to risk behaviors or symptoms.

Purpose of the Study:

• To assess the consistency of HIV incidence estimates derived from STARHS compared to a more robust cohort-based method.
• To investigate potential biases in STARHS when applied to populations with self-selected testing.

Main Methods:

• Applied a cohort-based incidence estimator and two STARHS methods to a population of 3821 individuals tested for HIV.
• Compared overall and demographically stratified seroincidence estimates and incidence rate ratios across methods.
• Evaluated the proportion of HIV-infected individuals with low antibody levels against expected proportions under independence.

Main Results:

• STARHS generally produced higher incidence estimates than the cohort-based method.
• Incidence rate ratios between demographic groups were inconsistent between STARHS and the cohort method.
• A higher proportion of HIV-infected individuals exhibited low antibody levels than expected, indicating a violation of the independence assumption.

Conclusions:

• HIV incidence estimates derived from methods relying on changing antibody levels, like STARHS, may be biased in populations where individuals choose their own test timing.
• Self-motivated testing can compromise the validity of STARHS for accurate HIV incidence calculation.