To permit the display of HIV prevalence data by ZIP code, AIDSVu obtained a data release agreement with health departments overseeing HIV surveillance in the following jurisdictions: Atlanta, Chicago, Dallas, Detroit, Houston, Los Angeles, Miami, New Orleans, New York City, Philadelphia, San Francisco, San Juan, and Washington, DC. Data were obtained directly from the public health departments rather than from the U.S. Centers for Disease Control and Prevention (CDC) because CDC does not collect street addresses for reported cases, and street addresses are required for mapping at the ZIP code level. Because data were not obtained directly from CDC, ZIP-code-level data are not directly comparable to the state- and county-level HIV prevalence data also displayed on AIDSVu.

Each public health department defined the geographic area (e.g., ZIP codes) in their jurisdiction for which they desired to display data on AIDSVu. ZIP-code-level HIV surveillance data reflect persons living with an HIV or AIDS diagnosis in the defined geographic area as of December 31, 2009. For details regarding each jurisdiction’s assignment of ZIP codes to cases missing an address at diagnosis, or to cases where the ZIP code at diagnosis has changed because of the redrawing of ZIP code boundaries, see the Data Caveats table below. Denominators used to calculate rates for ZIP codes were obtained from the U.S. Census Bureau’s 2010 census of ZIP Code Tabulation Areas (ZCTAs). Additional information on ZCTAs can be found here.

AIDSVu allows viewers to look at ZIP-code-level HIV surveillance data at the overall geographic level, and by race/ethnicity, sex and age groups. The black, white and Asian/Pacific Islander race groups are non-Hispanic, and the Hispanic/Latino ethnicity is inclusive of all races. Sex was defined as “sex at birth.” Cases were assigned to age groups based on age at the end of 2009. Caution should be exercised when viewing and interpreting the ZIP code maps because the scales change across the different breakdowns; the correct scale is always displayed to the left of the map, regardless of the breakdown.

Rates of persons living with an HIV/AIDS diagnosis were calculated per 100,000 population to permit data standardization and comparison. For ZIP codes that house correctional institutions, caution should be exercised when interpreting the data as these statistics are inclusive of institutional populations and their inclusion may artificially inflate the actual rate and case count of a ZIP code when an institution is housed in it. As is standard in the display of health statistics, rates generated from a numerator less than 12 are considered unstable and should be interpreted with caution. To determine if a specific ZIP code meets this criterion, place the mouse cursor over a ZIP code and, in the balloon that appears, there will be a message noting that a rate meets this condition, if applicable.

To protect the confidentiality of persons living with an HIV/AIDS diagnosis, AIDSVu does not display rates and case counts when the numerator (number of persons living with an HIV/AIDS diagnosis) is less than 5 and/or the denominator (number of people in the ZIP code in that population group) is less than 1,000. ZIP codes appear in a shade of gray when one or both of these conditions are met (see footnote below map scale). Because rates are not displayed when the numerator is less than 5, the “unstable rate” indicator mentioned above will only be displayed when the numerator is 5 or greater, but less than 12.

Range intervals were developed using deciles in SAS analytic software (SAS Institute, Cary, NC) and are based on the combined data of all jurisdictions providing ZIP-code-level data so that comparisons can be made across cities. In order to illustrate the variation in rates and case counts, ranges were developed specific to each grouping (i.e., overall, by race/ethnicity, by sex and by age group). Thus, a total of eight sets of map scales exist (rate and count, both by overall/race/sex/age). For each scale calculation, we determined deciles by combining the individual rates or counts for each ZIP code and grouping.

Jurisdiction-Specific Data Caveats

Data Sources

To permit the display of census-tract-level HIV prevalence data, AIDSVu obtained a data release agreement with the public health departments overseeing HIV surveillance in Washington, DC and Philadelphia. Data were obtained directly from the public health departments and not from CDC because CDC does not collect street addresses for reported cases, and street addresses are required for mapping at the census tract level. Because of this, census-tract-level data are not directly comparable to the state- and county-level HIV prevalence data also displayed on AIDSVu.

Census-tract-level HIV surveillance data reflect persons living with an HIV infection or AIDS diagnosis as of December 31, 2009. Residence at earliest HIV or AIDS diagnosis was used in the assignment of cases to 2010 census tracts. For more information on how these two jurisdictions handle the assignment of a census tract to cases missing an address at diagnosis, see the Data Caveats table below. Denominators used to calculate rates for census tracts were obtained from the U.S. Census Bureau’s 2010 census of census tracts.

AIDSVu allows viewers to look at census-tract-level HIV surveillance data at the overall geographic level, and by race/ethnicity, sex and age groups. The black and white race groups are non-Hispanic, and the Hispanic/Latino ethnicity is inclusive of all races. Sex was defined as “sex at birth.” Cases were assigned to age groups based on age at the end of 2009. Caution should be exercised when viewing and interpreting the census tract maps because the scales change across the different breakdowns; the correct scale is always displayed to the left of the map, regardless of the breakdown.

Rates of persons living with an HIV/AIDS diagnosis were calculated per 100,000 population to permit data standardization and comparison. For census tracts that house correctional institutions, caution should be exercised when interpreting the data as these statistics are inclusive of institutional populations and their inclusion may artificially inflate the actual rate and case count of a census tract when an institution is housed in it. As is standard in the display of health statistics, rates generated from a numerator less than 12 are considered unstable and should be interpreted with caution. To determine if a specific census tract meets this criterion, place the mouse cursor over a census tract and, in the balloon that appears, there will be a message noting that a rate meets this condition, if applicable.

To protect the confidentiality of persons living with an HIV/AIDS diagnosis, AIDSVu does not display rates and case counts for census tracts when the numerator (number of persons living with an HIV infection diagnosis) is less than 5 and/or the denominator (number of people in the census tract in that population group) is less than 500. Census tracts appear in a shade of gray when one or both of these conditions are met (see footnote below map scale). Because rates are not displayed when the numerator is less than 5, the “unstable rate” indicator mentioned above will only be displayed when the numerator is 5 or greater, but less than 12.

Scale intervals were developed using deciles in SAS analytic software and are based on the combined data of the two jurisdictions providing census-tract-level data so that comparisons can be made between cities. In order to illustrate the variation in rates and case counts, ranges were developed specific to each grouping (i.e., overall, by race/ethnicity, by sex and by age group). Thus, a total of eight sets of map scales exist (rate and count, both by overall/race/sex/age). For each scale calculation, we determined deciles by combining the individual rates or counts for each census tract, including all levels of race/ethnicity, sex or age, when applicable.

Jurisdiction-Specific Data Caveats

Data Sources