Background The HIV incidence is the most valuable indicator of epidemiological surveillance as it establishes the degree to which HIV transmission is occurring and which groups are most at risk for HIV infection. However, the HIV incidence is a difficult indicator to estimate. This paper addresses a method to estimate the HIV incidence in Brazil in recent years.
Methods The information source is SISCEL, which is the national laboratory-based surveillance system created to monitor CD4+/CD8+ T lymphocytes and HIV viral load. The proposed method is based on the first CD4 count after HIV diagnosis among all treatment-naïve cases registered in SISCEL in the time period 2004–2011. A regression model that relates progress of CD4 to time of seroconversion was used to estimate time of HIV infection at the date of first CD4 count. The analysis was performed by sex and age group.
Results For all years, the proportion of HIV cases registered in SISCEL in the same year of infection was 31%, and approximately 6% for each year of the 5 subsequent years after infection, so that almost 40% cases are registered in SISCEL only 7 years or more after infection. The HIV incidence was stable in the period 2005–2011. After adjusting for HIV cases tested in private laboratory and undiagnosed cases not registered in SISCEL, the mean estimate was 41600 cases, corresponding to an HIV incidence rate of 0.27 per 1000 population. Analysis by sex and age showed a rising trend among males, particularly for young men, and a decreasing trend among females.
Conclusion In terms of late HIV diagnosis, the results indicate that coverage of HIV testing in the general population should be largely expanded. As to most at risk populations, the findings suggest that interventions should be focused on young adult men to reverse the increasing trend.