Abstract
Objectives:
We examined the survival rates after diagnosis of HIV infection stage 3 (AIDS) in the United States by population density area of residence at diagnosis.
Methods:
We used data from the National HIV Surveillance System to calculate survival rates among people aged ≥13 with HIV infection stage 3 (AIDS) diagnosed from 2005 through 2010. We determined survival rates for more than 12, 24, and 36 months after diagnosis; overall and by demographic characteristics; and across 3 population density area categories (large metropolitan statistical areas [MSAs, ≥500 000 people], small-to-medium MSAs [50 000 to 499 999 people], and nonmetropolitan areas [<50 000 people]).
Results:
The survival rates for more than 12, 24, and 36 months after diagnosis were highest among people residing in large MSAs (90.2%, 87.2%, and 84.9%, respectively) and lowest among people residing in nonmetropolitan areas (87.3%, 84.1%, and 81.4%, respectively). With a few exceptions, survival rates were lower in those residing in nonmetropolitan areas than those residing in large MSAs and small-to-medium MSAs across most subgroups by age at diagnosis, race/ethnicity, sex, transmission category, region of residence, and year of diagnosis. Between 2005 and 2010, significant year-to-year increases occurred in the proportion of people surviving more than 36 months after diagnosis across all 3 population density area categories (estimated annual percentage change: large MSAs [0.88; 95% confidence interval (CI), 0.56-1.20]; small-to-medium MSAs [0.94; 95% CI, 0.06-1.83]; and nonmetropolitan areas [1.26; 95% CI, 0.07-2.46]).
Conclusions:
Although survival rates for those with HIV infection stage 3 (AIDS) improved in all 3 population density area categories, efforts to remove barriers to care and promote treatment adherence in nonmetropolitan areas will be necessary to eliminate survival disparities.
The goals of the National HIV/AIDS Strategy include improving the health outcomes for people living with human immunodeficiency virus (HIV) infection by increasing access to care and reducing health-related disparities. 1 Survival is a useful outcome measure after the diagnosis of HIV infection (regardless of infection stage) and particularly after the diagnosis of HIV infection stage 3 (AIDS) (hereinafter, stage 3 [AIDS]).
In the United States, survival rates after HIV infection diagnosis and stage 3 (AIDS) diagnosis vary by demographic subgroups. 2 –4 Among people with stage 3 (AIDS) diagnosed from 2005 through 2010, survival was greater among males aged ≥13 than among females aged ≥13, varied by race/ethnicity and HIV transmission category, generally increased with more recent years of diagnosis, and decreased with older age at the time of diagnosis. 2
Most people aged ≥13 with newly diagnosed HIV infection in 2015 (82.3%) and living with diagnosed HIV infection at the end of 2014 (83.9%) resided in large metropolitan statistical areas (MSAs), defined as areas with ≥500 000 people. 5 Similarly, the rate of new diagnoses of HIV infection and of stage 3 (AIDS) and the prevalence of people aged ≥13 living with diagnosed HIV infection and stage 3 (AIDS) were higher in larger population density areas than in nonmetropolitan areas (defined as areas with <50 000 people).
Despite this higher incidence and prevalence of HIV infection in larger population density areas, the proportion of people aged ≥13 with a late diagnosis of HIV infection (defined as a stage 3 [AIDS] diagnosis within 3 months of HIV infection diagnosis) in 2012 was actually lower in larger population density areas than in smaller population density areas. 6 Indeed, more barriers to accessing care and treatment adherence are reported in rural (mostly nonmetropolitan) areas than in urban (mostly metropolitan) areas, a finding that may negatively affect health outcomes after diagnosis of HIV infection or stage 3 (AIDS). 7 –9 In addition, poor retention in care is associated with poorer survival after HIV infection diagnoses. 10 At least 1 study examined the differences in survival after stage 3 (AIDS) diagnosis by population density areas; however, this study focused only on 9 southern US states. 11 To our knowledge, no studies have measured and compared stage 3 (AIDS) diagnosis survival outcomes by population density areas for the entire United States.
Our objectives were to assess the survival rates of people aged ≥13 throughout the United States with newly diagnosed stage 3 (AIDS) and to determine whether or not those rates differed by population density areas of residence at the time of diagnosis (hereinafter referred to as population density area categories), which were potentially related to breakdowns in care. Identifying differences in survival after diagnosis of stage 3 (AIDS) among population density area categories may help to identify populations with disparities in care and target interventions to improve survival in those areas.
Methods
Data Source and Inclusion Criteria
Our study included all people aged ≥13 with stage 3 (AIDS) diagnosed from 2005 through 2010 in the 50 US states and the District of Columbia. We used surveillance data collected by state and local health departments and reported to the Centers for Disease Control and Prevention National HIV Surveillance System through July 2015. 2 We analyzed the surveillance data to follow those with diagnosed stage 3 (AIDS) for a 36-month period after diagnosis and then for an additional 19 months after that, to account for any delays in reporting of deaths. Therefore, techniques to account for censoring of people due to an incomplete follow-up period were not required, and we did not need to adjust the data for reporting delays in the diagnosis of stage 3 (AIDS) or death. We used data collected as part of routine public health surveillance activities classified as nonresearch. As such, institutional review board review was not required.
Demographic Data
We collected data on demographic characteristics. We categorized race/ethnicity into the following groups: American Indian/Alaska Native, Asian/Native Hawaiian/other Pacific Islander, black/African American, Hispanic/Latino, white, and multiple races. We classified sex according to sex assigned at birth. We classified transmission category by the most likely route of HIV transmission, based on a hierarchy of reported risk information (male-to-male sexual contact, injection drug use [IDU], male-to-male sexual contact and IDU, heterosexual contact, and other [ie, hemophilia, blood transfusion, perinatal exposure, or unknown]). 2 We did not adjust data to account for those with unknown HIV transmission. We assigned region of residence according to where people were living at the time of their stage 3 (AIDS) diagnosis, using US Census Bureau regions. 12
Other Measures
We grouped data into 3 population density area categories: MSAs with populations of ≥500 000 (ie, large MSAs), MSAs with populations of 50 000 to 499 999 (ie, small-to-medium MSAs), and nonmetropolitan areas with populations <50 000. 13 MSAs consisted of a core urban area with ≥50 000 people combined with adjacent counties that are closely tied socially and economically to the core urban area.
Analysis
We calculated the proportions of people who survived (ie, survival rate) more than 12, 24, and 36 months after diagnosis of stage 3 (AIDS) by population density area category, and we further presented these results by age at the time of diagnosis, race/ethnicity, sex, transmission category, region of residence, and year of diagnosis. However, because the surveillance data that we used represented a complete census of people with diagnosed stage 3 (AIDS) rather than population samples, we did not apply statistical testing to compare these results.
We analyzed 2005 to 2010 survival trends by using Poisson regression analysis to calculate estimated annual percentage changes in survival rates for more than 12, 24, and 36 months after stage 3 (AIDS) diagnosis during that period, and we presented the results by the 3 population density area categories. We considered the estimated annual percentage change to be significant when the 95% confidence interval (CI) did not contain an estimated annual percentage change of 0. We performed all analyses using SAS version 9.3. 14
Results
During the period 2005-2010, 190 771 people aged ≥13 were diagnosed with stage 3 (AIDS) in the United States. Of these, at the time of diagnosis, 157 047 (82.3%) resided in large MSAs, 21 084 (11.1%) in small-to-medium MSAs, and 11 737 (6.2%) in nonmetropolitan areas (Table 1). The survival rates for >12 months, >24 months, and >36 months after diagnosis were highest among people residing in large MSAs (90.2%, 87.2%, and 84.9%, respectively) and lowest among people residing in nonmetropolitan areas (87.3%, 84.1%, and 81.4%, respectively) (Table 2).
Number and proportion of HIV infection stage 3 (AIDS) diagnoses among people aged ≥13, by population density area categorya of residence and other demographic characteristics, United States and District of Columbia, 2005-2010b
Abbreviations: HIV, human immunodeficiency virus; MSA, metropolitan statistical area.
aLarge defined as ≥500 000 population; small-to-medium as 50 000-499 999 population; nonmetropolitan as <50 000 population.
bBased on data from the Centers for Disease Control and Prevention National HIV Surveillance System collected through July 2015.
cRow percentages do not add to 100 because some people (fewer than 0.5%) with diagnosed stage 3 (AIDS) could not be classified into a population density area category.
dHispanic/Latino includes all Hispanic and Latino people of any race.
eOther includes people whose likely route of transmission was hemophilia, blood transfusion, perinatal exposure, or unknown; data were not adjusted to account for those with unknown transmission category.
Survival rates among people aged ≥13 at >12 months, >24 months, and >36 months after diagnosis of HIV infection stage 3 (AIDS), by population density area categorya of residence and other demographic characteristics, United States and the District of Columbia, 2005-2010b
Abbreviations: HIV, human immunodeficiency virus; HRH, heterosexual contact; IDU, injection drug use; MSA, metropolitan statistical area; MSM, men who have sex with men (ie, HIV infection attributed to male-to-male sexual contact).
aLarge defined as ≥500 000 population; small-to-medium as 50 000-499 999 population; nonmetropolitan as <50 000 population.
bBased on data from the Centers for Disease Control and Prevention National HIV Surveillance System collected through July 2015.
cTotals include people with unknown population density area category.
dHispanic/Latino includes all Hispanic and Latino people of any race.
eOther includes people whose likely route of transmission was hemophilia, blood transfusion, perinatal exposure, or unknown; data were not adjusted to account for those with unknown transmission category.
Survival Rates by Age Group
Survival rates for more than 12, 24, and 36 months after diagnosis were highest among people residing in large MSAs and lowest among people residing in nonmetropolitan areas for all age groups, with a few exceptions (Table 2). Among adults aged 20-24, those residing in large MSAs and small-to-medium MSAs had nearly equal survival rates for >12 months (96.2% and 96.4%, respectively) and for >24 months (94.6% and 94.7%, respectively) after diagnosis (Table 2). In addition, among adults aged 55-59, compared with those residing in small-to-medium MSAs, those residing in large MSAs had lower survival rates for >12 months (81.6% and 82.5%, respectively), >24 months (77.0% and 78.3%, respectively), and >36 months (73.3% and 74.8%, respectively) after diagnosis. Also, the survival rate for >12 months after diagnosis was lower among people aged ≥65 residing in large MSAs than among those residing in nonmetropolitan areas (69.0% and 69.9%, respectively), and survival rates for >12 months (65.9% and 69.9%, respectively), >24 months (58.1% and 62.4%, respectively), and >36 months (53.0% and 57.7%, respectively) after diagnosis were lower among those residing in small-to-medium MSAs than among those residing in nonmetropolitan areas.
The difference in survival rates among the 3 population density area categories was ≥5.0 percentage points in only a few age groups (Table 2). The survival rate for >36 months after diagnosis was 5.0 percentage points higher among people aged 55-59 residing in small-to-medium MSAs (74.8%) than among those of that age residing in nonmetropolitan areas (69.8%). Also, for adults aged 60-64, the survival rate for >12 months after diagnosis was 5.0 percentage points higher among those residing in large MSAs (77.6%) than among those residing in nonmetropolitan areas (72.6%); the survival rate for >36 months after diagnosis was 6.4 percentage points higher for those residing in large MSAs (68.2%) and 5.4 percentage points higher for those residing in small-to-medium MSAs (67.2%) than for those residing in nonmetropolitan areas (61.8%).
Survival Rates by Race/Ethnicity
Survival rates for more than 12, 24, and 36 months after diagnosis were highest among people residing in large MSAs and lowest among people residing in nonmetropolitan areas for all racial/ethnic groups, with a few exceptions (Table 2). Survival rates for more than 12, 24, and 36 months after diagnosis were lower among American Indians/Alaska Natives residing in small-to-medium MSAs (82.9%, 80.6%, and 75.2%, respectively) than among those residing in nonmetropolitan areas (84.7%, 81.8%, and 76.4%, respectively).
Among Asians/Native Hawaiians/other Pacific Islanders, survival rates for more than 12, 24, and 36 months after diagnosis were lower for those residing in large MSAs (91.9%, 90.6%, and 90.0%, respectively) than among those residing in nonmetropolitan areas (93.6%, 93.6%, and 91.5%, respectively), and survival rates for >24 months and >36 months after diagnosis were lower in Asians/Native Hawaiians/other Pacific Islanders residing in small-to-medium MSAs (92.9% and 91.2%, respectively) than among those residing in nonmetropolitan areas (93.6% and 91.5%). However, the number of Asians/Native Hawaiians/other Pacific Islanders residing in nonmetropolitan areas was small (n = 47).
In addition, among Hispanic/Latino people residing in small-to-medium MSAs compared with those residing in nonmetropolitan areas, survival rates were lower for >12 months (89.6% and 90.1%, respectively), >24 months (87.4% and 88.4%, respectively), and >36 months (85.9% and 87.0%, respectively) after diagnosis (Table 2). Among people of multiple races, survival rates for more than 12, 24, and 36 months after diagnosis were lower for those residing in large MSAs (91.4%, 88.1%, and 85.7%, respectively) and small-to-medium MSAs (91.4%, 88.2%, and 84.7%, respectively) than among those residing in nonmetropolitan areas (92.2%, 89.5%, and 87.3%, respectively).
The difference in survival between population categories was ≥5.0 percentage points in only 1 racial/ethnic group (Table 2). The survival rate for >36 months after diagnosis was 6.6 percentage points higher among American Indians/Alaska Natives residing in large MSAs (83.0%) than among those residing in nonmetropolitan areas (76.4%).
Survival Rates by Sex and Transmission Category
Among both sexes, survival rates for more than 12, 24, and 36 months after diagnosis were highest in all transmission categories for those residing in large MSAs and lowest for those residing in nonmetropolitan areas, with a few exceptions (Table 2).
Among males, survival rates for more than 12, 24, and 36 months after diagnosis were lower in those with stage 3 (AIDS) attributed to IDU residing in large MSAs (85.3%, 80.2%, and 76.2%, respectively) and in small-to-medium MSAs (86.9%, 83.0%, and 79.2%, respectively) than among those residing in nonmetropolitan areas (88.3%, 85.2%, and 81.6%, respectively) (Table 2). Similarly, survival rates for >24 months and >36 months after diagnosis were lower among males with stage 3 (AIDS) attributed to both male-to-male sexual contact and IDU residing in large MSAs (88.7% and 86.2%, respectively) and in small-to-medium MSAs (88.7% and 85.7%, respectively) than among those residing in nonmetropolitan areas (88.8% and 86.8%, respectively).
Among females, survival rates for more than 12, 24, and 36 months after diagnosis were lower among those with stage 3 (AIDS) attributed to IDU residing in large MSAs (88.4%, 83.1%, and 78.6%, respectively) than among those residing in small-to-medium MSAs (90.2%, 85.0%, and 81.2%, respectively) (Table 2). The survival rate for >24 months after diagnosis was lower among females with stage 3 (AIDS) attributed to IDU residing in large MSAs (83.1%) than among those residing in nonmetropolitan areas (83.8%).
Survival by Region of Residence
Survival rates were highest among people residing in large MSAs and lowest among people residing in nonmetropolitan areas for all regions, with a few exceptions (Table 2). In the northeast region, survival rates for more than 12, 24, and 36 months after diagnosis were lower among those residing in large MSAs (91.1%, 88.3%, and 86.0%, respectively) and small-to-medium MSAs (90.9%, 87.6%, and 85.4%, respectively) than among those residing in nonmetropolitan areas (92.7%, 91.0%, and 89.7%, respectively). In the west region, the survival rate for >24 months after diagnosis was lower among those residing in small-to-medium MSAs (87.1%) than among those residing in nonmetropolitan areas (87.6%).
Survival by Year of Diagnosis
For all years of stage 3 (AIDS) diagnosis, survival rates for more than 12, 24, and 36 months after diagnosis were highest among people residing in large MSAs and lowest among those residing in nonmetropolitan areas, with one exception (Table 2). The 2010 survival rate for >24 months after diagnosis was lower among those residing in small-to-medium MSAs (86.2%) than among those residing in nonmetropolitan areas (86.4%).
Trends in Survival by Population Density Area Category
For those residing in large MSAs at the time of their stage 3 (AIDS) diagnosis, survival rates for more than 12, 24, and 36 months after diagnosis all increased significantly between 2005 and 2010 (estimated annual percentage change = 0.42, 95% CI, 0.11-0.73; estimated annual percentage change = 0.68, 95% CI, 0.37-0.99; and estimated annual percentage change = 0.88, 95% CI, 0.56-1.20, respectively). In addition, for those residing in small-to-medium MSAs and nonmetropolitan areas, survival rates for >36 months after diagnosis also increased significantly during that period (estimated annual percentage change = 0.94, 95% CI, 0.06-1.83; and estimated annual percentage change = 1.26, 95% CI, 0.07-2.46, respectively) (Table 3).
Survival rates and estimated annual percentage changes among people aged ≥13 at >12 months, >24 months, and >36 months after diagnosis of HIV infection stage 3 (AIDS), by population density area categorya of residence, United States and the District of Columbia, 2005-2010b
Abbreviations: HIV, human immunodeficiency virus; MSA, metropolitan statistical area.
aLarge defined as ≥500 000 population; small-to-medium as 50 000-499 999 population; nonmetropolitan as <50 000 population.
bBased on data from the Centers for Disease Control and Prevention National HIV Surveillance System collected through July 2015.
Discussion
Overall, we found that survival rates for more than 12, 24, and 36 months after diagnosis of stage 3 (AIDS) were highest among people residing in large MSAs and lowest among people residing in nonmetropolitan areas. The higher survival rates for those residing in large MSAs may have been because of better access to HIV care and tighter adherence to treatment once in care. Rural (mostly nonmetropolitan) areas have more barriers to accessing HIV care than urban areas, 7,8 including transportation issues, lack of qualified providers, confidentiality concerns, lack of affordable health insurance, and HIV-related stigma. 7,8,15 HIV-related stigma has in turn been associated with poor medication adherence. 9 Thus, the higher levels of HIV-related stigma experienced by people in rural areas were likely to affect their willingness to access care and their adherence to recommended medical treatment.
We also found larger differences in survival rates between large MSAs and nonmetropolitan areas than between small-to-medium MSAs and nonmetropolitan areas. This finding may also reflect that people residing in large MSAs had better access to HIV care than those residing in small-to-medium MSAs. An analysis of survival for more than 60 months after a diagnosis of stage 3 (AIDS) in 2003 and 2004 in 9 southern states also found that survival rates were highest in large MSAs. 11 One important exception that we observed was in the northeast region of the United States, where survival rates for more than 12, 24, and 36 months after diagnosis were highest among those residing in nonmetropolitan areas. The smaller geographic distances between MSAs and nonmetropolitan areas in the northeast region compared with other US regions may have resulted in fewer barriers to accessing HIV care in that region.
For most demographic subgroups, including age at diagnosis, race/ethnicity, sex, transmission category, and year of diagnosis, survival rates for more than 12, 24, and 36 months after diagnosis were higher for those residing in large MSAs and small-to-medium MSAs than for those residing in nonmetropolitan areas. When comparing those living in either type of MSA with those living in nonmetropolitan areas, the differences in survival rates for >12 and >36 months after diagnosis among people aged 60-64 residing in large MSAs and the survival rates for >36 months after diagnosis among people aged 55-59 and 60-64 in small-to-medium MSAs were all ≥5.0 percentage points higher. In contrast, survival rates among people aged ≥65 were similar in large MSAs and nonmetropolitan areas, perhaps because access to medical care among people enrolled in Medicare is similar in rural and urban areas. 16
Survival rates for >36 months after diagnosis among American Indians/Alaska Natives residing in large MSAs were more than 5.0 percentage points higher than among those residing in nonmetropolitan areas. Greater concerns with confidentiality, stronger HIV-related stigma, and mistrust of Indian Health Services on reservations may have limited the accessing of medical care by American Indians/Alaska Natives who reside in these nonmetropolitan areas. 17
Despite the general trend that survival rates were higher for those with diagnosed stage 3 (AIDS) residing in large MSAs and small-to-medium MSAs than for those residing in nonmetropolitan areas, there were some important exceptions. For example, among males with HIV diagnoses attributed to IDU, survival rates for more than 12, 24, and 36 months after diagnosis were lower among those residing in large MSAs and small-to-medium MSAs than among those residing in nonmetropolitan areas. These findings suggest that in certain demographic subpopulations, there may be factors other than population density area of residence that influence access to care and adherence to HIV treatment and that subsequently affect survival.
Finally, survival rates for >36 months after stage 3 (AIDS) diagnosis increased in all 3 population density area categories from 2005 to 2010. These year-to-year increases may have resulted from introduction of new HIV therapies, including 2 new classes of medications (entry inhibitors and integrase inhibitors), 18 improvements in retention in care for people with HIV, 19 improvements in access to care, or improvement in treatment adherence across all population density area categories between 2005 and 2010.
Limitations
This study had several limitations. First, the designation of the population density area category was based on people’s residence at the time of stage 3 (AIDS) diagnosis, and this residence may have differed from where people resided during the years after diagnosis. However, migration after diagnosis likely had minimal effect on the results, given that a previous study reported that migration after diagnosis of HIV infection was similar between urban and rural areas and that the demographic characteristics of people who migrated were similar to those of people who did not migrate. 20
Second, because MSAs are defined as counties with a core urban area with populations of ≥50 000 and include adjacent counties with close social and economic ties to the core urban area, the composition of these geographic areas may have varied. For example, certain locations in MSAs may be geographically distant from the core urban area. As a result, access to medical care may have varied somewhat substantially within each population density area category, which may have affected our results.
Third, although we used surveillance data to follow those with stage 3 (AIDS) for a 36-month period after diagnosis and then for an additional 19 months subsequently, additional delays in reporting of deaths may have occurred. State and local health departments are responsible for linking their HIV surveillance registries with state and national data sources to identify deaths, but the timeliness for completing the linkages varies by state. Thus, deaths may have been underreported in our results.
Finally, we calculated survival rates based on deaths among people with diagnosed stage 3 (AIDS) regardless of cause of death. Although survival rate differences were likely mostly influenced by variations in access to HIV-related medical care by geographic density areas, survival rates may also have been influenced by disparities in medical care in general or other environmental differences (eg, violence, drug overdoses).
Conclusions
Although survival rates after stage 3 (AIDS) diagnosis improved in all 3 population density area categories from 2005 through 2010, survival rates for more than 12, 24, and 36 months after diagnosis were highest among people residing in large MSAs and lowest among people residing in nonmetropolitan areas overall and in most demographic subgroups. Continued efforts from community-based organizations, health care providers, health departments, and government agencies are needed to remove barriers to HIV care and promote treatment adherence in nonmetropolitan areas to eliminate disparities in survival.
Footnotes
Acknowledgments
The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported as public health surveillance by the Centers for Disease Control and Prevention.