Questions and Answers: The 15% Increase in HIV Diagnoses from 2004-2007 in 34 States and General Surveillance Report Questions

• CDC’s new HIV/AIDS Surveillance Report shows that from 2004-2007, HIV diagnoses increased 15% in the 34 states that have long-term, name based HIV reporting. Why?

  There are four main reasons that could explain the increase in diagnoses.

  • The increase in new diagnoses may be due to changes in state reporting regulations.
    • Several states have revised their HIV reporting laws to include laboratory reporting of all viral loads and CD4s. This may have lead to more complete reporting of new diagnoses as well as identification of prevalent cases of HIV that were previously unreported. As states continue to implement revised surveillance practices, the annual numbers of reported and diagnosed cases of HIV infection are likely to fluctuate.
      • Electronic lab reporting of viral loads and CD4 counts enables states to capture HIV cases among persons diagnosed with HIV infection who may not have been reported when they were initially diagnosed, or that were diagnosed prior to implementation of confidential name-based reporting in that area.
      • We cannot determine what proportion of the increase is due to changes in regulations.
  • There may be more people getting tested for HIV, which could result in more new diagnoses.
    • There has been increasing emphasis on the benefits of increased HIV testing among persons at high risk which may have lead to an increase in HIV testing, and therefore an increase in diagnoses reported to surveillance systems, among various populations and in certain states. Increases in HIV testing may have also resulted from broader use of rapid HIV tests and changes in testing practices following publication of CDC’s guidelines for routine HIV screening in health care settings in 2006. More HIV testing may have lead to more diagnoses (i.e., a decrease in undiagnosed infections, which we saw with the release of the recent HIV prevalence data).
    • CDC has emphasized HIV testing since the release of its 2006 testing guidelines and has a number of efforts to increase testing, e.g., increasing testing efforts among African Americans and working to implement routine HIV testing in healthcare settings. Therefore, it is not unrealistic to think that more people are getting tested, which could lead to more HIV diagnoses.
      • An MMWR published in October 2008 showed that testing rates are higher for pregnant women, persons who acknowledge HIV risks, young persons (ages 18-34), and African Americans than other groups. This suggests that testing efforts have reached many of the people disproportionately affected by HIV.
  • The increase in new diagnoses may be linked to instability in the data.
    • The increase appeared mostly in the last year of reporting (2007).
    • Due to some uncertainty inherent in statistical estimates, the estimated data for the most recent year is always the least stable.
  • There may actually be an increase in new HIV infections.
    • The surveillance report notes, “…it is not possible to rule out a rise in HIV infections since the estimation models include a degree of uncertainty. Moving forward, data from CDC’s new HIV incidence surveillance system will provide the best indication of changes in new HIV infections. Data on trends in new HIV infections will be available after at least three years of data have been reported from the new systems.”
      • CDC will be better able to tell if this increase in diagnoses is due to an increase in incidence once we have at least three years of data from the HIV incidence surveillance system (STARHS).
• Are there increases in certain races/ethnicities, age groups, or transmission categories that stand out, or is there an across-the-board increase?

  Increases in annual estimates of HIV/AIDS diagnoses were seen among subpopulations.

  • From 2004 to 2007, there was a 26% increase in estimated annual HIV/AIDS diagnoses among MSM. This increase may represent an increase in HIV incidence in MSM, which was found in the recent incidence data. It may also be affected by increases in testing.
    • HIV diagnoses among MSM increased all four years of the analysis, with the largest increase occurring in the last year.
    • Although it is difficult to determine whether the increase in diagnoses represents a true increase in incidence, a recent CDC analysis indicated that HIV incidence among gay and bisexual men has been increasing since the early 1990s.
      • Additional years of data from CDC’s incidence surveillance system will allow us to determine whether these increases have continued.
  • During this same period a 9% increase in annual diagnoses was seen among male high-risk heterosexuals, and a 14% increase among females.
    • Additional incidence surveillance data will allow us to determine whether these increases represent increases in incidence or are due to the other factors outlined in question #1.
  • There were increases in annual diagnoses among all racial/ethnic groups.
    • Again, future years of incidence surveillance data will allow us to determine if the increases represent increases in incidence or are due to the other factors outlined in question #1.
• The report says that HIV diagnoses have increased among people 50 and older. Does this mean HIV is increasing among older Americans?

  It should not be concluded that the annual number of new HIV infections among older Americans is increasing based on HIV diagnosis data alone. HIV diagnosis data should be interpreted with caution as some persons may have been infected recently while others were infected many years ago. However, an increase in routine HIV screening could account for an increase in HIV diagnoses. Data from the HIV incidence surveillance system (after at least 3 more years of reporting) will provide answers to this question.

• What do HIV diagnoses data tell us that cannot be discerned from incidence data?

  Estimates from CDC’s new HIV incidence surveillance system provide the clearest picture to date of new infections (HIV incidence), both overall and among the most affected populations. However, HIV diagnosis data are needed to provide a measure of the burden of disease, testing, prevalence, and treatment/care needs.

  • HIV diagnosis data, in conjunction with HIV incidence estimates, will provide information to help evaluate efforts to increase the proportion of HIV-positive individuals who are tested and aware of their status.
  • HIV diagnosis data also allow us to evaluate efforts to improve early HIV diagnosis, as data on concurrent HIV and AIDS diagnoses provide an indication of the proportion of individuals being diagnosed late in the course of their disease.
  • HIV diagnosis data also provide a picture of the current burden of disease (HIV prevalence), as they indicate all individuals who have been diagnosed with HIV infection and/or AIDS.

  Also, despite the potential limitations of using HIV diagnosis data as a proxy measurement for HIV incidence, data on HIV diagnoses continue to provide the best information on the distribution of HIV infection in

  • areas that do not collect data for HIV incidence surveillance;
  • areas without sufficient incidence data;
  • smaller populations (such as smaller racial/ethnic groups), in which it may not be possible to generate reliable HIV incidence estimates;
  • young people, ages 13-24, due to their recent infections.
• From 2004-2007, how many states were added to the name-based HIV reporting system? How does this addition affect the data? Would this/these additions account for the 15% increase?

  Georgia was added to the tables for estimated HIV/AIDS, increasing the number of states included in these analyses from 33 to 34. Cases from Georgia did contribute to the increase seen from 2004 to 2007; however, the addition of Georgia alone did not result in the increase. Estimates of annual diagnoses in Georgia increased 10% from 2004 to 2007; however, numerous other states, including a few with higher HIV morbidity (e.g., Florida, New York) showed increases in annual estimates.

  States are added to tables for estimates of HIV/AIDS after sufficient time (at least 4 years of confidential name-based reporting) has passed in order to allow for stabilization of reporting, calculation of reporting delay weights, and more accurate interpretation of trends.

• Are there state or regional increases in HIV diagnoses, or is the increase evenly distributed across the country?

  Increases were seen in 28 out of the 34 states.

  1. What do we know about trends in the South (if not based on HIV data – then what do we know based on AIDS data)?

     Increases were seen in most (12/13) of the southern states for which estimated HIV diagnosis data is available (13/17). We do not provide annual estimates of HIV/AIDS, by region, in the annual report.

     AIDS data should not be used to evaluate trends in HIV infection as AIDS diagnoses are a result of missed opportunities and/or failed treatment. However, looking at the data by region from 2003 to 2007, the estimated number of AIDS cases decreased 5% in the South.

• The 15% increase covers years 2004-2007. Was there an increase in diagnoses in each year?

  The annual number of HIV/AIDS diagnoses remained stable from 2004 through 2006. A 15% increase in annual HIV/AIDS diagnoses from 2006 to 2007 accounts for the majority of the increase seen from 2004 to 2007. As we know, the most recent year of data is the most subject to change as the estimation models include a degree of uncertainty

• The surveillance report says that AIDS cases have decreased. Why? It also says that AIDS deaths have decreased. Why?

  AIDS cases have decreased due to widespread use of highly active antiretroviral therapy which prolongs the time from diagnosis with HIV infection to progression to AIDS.

  Estimated deaths among persons with AIDS have most likely decreased as a result of HAART. As HAART increases the time from diagnosis with HIV infection to progression to AIDS it also extends the lifespans of persons with AIDS.

  PLEASE NOTE FOR ALL DEATH- RELATED QUESTIONS: It is believed that the decrease seen in annual deaths from 2004-2007 among persons with AIDS is real; however, the extent of the decrease may be overemphasized due to delays in reporting of deaths in the most recent year(s) to CDC. CDC is currently re-evaluating its protocols for reporting deaths in order to provide a more accurate picture.