
Estimating the Number of Persons Who Inject Drugs in the United States by MetaAnalysis to Calculate National Rates of HIV and Hepatitis C Virus Infections



Download the PDF here
Analysis to Calculate National Rates of HIV and Hepatitis C Virus Infections
Published: May 19, 2014 plos one
Amy Lansky1*, Teresa Finlayson1, Christopher Johnson1, Deborah Holtzman2, Cyprian Wejnert1, Andrew Mitsch1, Deborah Gust1, Robert Chen1, Yuko Mizuno1, Nicole Crepaz1
1 Division of HIV/AIDS Prevention, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America, 2 Division of Viral Hepatitis, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia, United States of America
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0097596
Using data from national populationbased U.S. surveys, we estimated that persons who ever injected drugs comprised 2.6% (CI: 1.8%3.3%) of the U.S. population. This represents approximately 6,612,488 million PWID (range: 4,583,1888,641,788) aged 13 years or older in 2011. Although PWID comprise 3% or less of the U.S. population, they account for 22% of all persons living with HIV infection [2]. Our estimates also quantified the recognized disparity of HIV disease rates among black and Hispanic/Latino male and female PWID when compared with white male and female PWID.
The disparity in disease burden among PWID compared to their population size has been difficult to quantify......We conducted a metaanalysis of national surveys to estimate the proportion of persons in the U.S. who have injected drugs, used these estimates to calculate disease metrics for PWID using national surveillance data for HIV infection, and calculated rate ratios by gender, race/ethnicity, and age. We used other methods to estimate rates of HIV infection among PWID in Puerto Rico and rates of HCV infection among adult PWID in the U.S. The estimates of the number of PWID and rates of HIV and HCV infection among PWID are needed to effectively plan, implement at an appropriate scale, and evaluate programs that serve PWID with or at risk for bloodborne infections such as HIV, HBV, or HCV.
Proportion of the Population and Number Estimated to be PWID in the U.S
we estimate that approximately 774,434 adults and adolescents (range: 494,6051,054,263) injected drugs in the past year in the United States.........Table 2 shows the estimated population proportion of lifetime PWID overall and for males and females for each populationbased survey and the combined estimates from the metaanalysis. The overall combined estimate for the ever recall period was 2.6% (confidence interval [CI]: 1.8%3.3%). As noted, Q statistics and I2 indicated heterogeneity of results across the surveys (Q2 = 45.1, P <.0001, I2 = 95.6). The combined estimate for males was 3.6% (CI: 2.4%4.8%) and for females was 1.6% (95% CI: 1.1%2.0%). Applying these proportions to the U.S. population age 13 years or older for 2011, we estimate that approximately 6,612,488 adults and adolescents ever injected drugs, with a range from 4,583,188 to 8,641,788 persons; using the sexspecific proportions an estimated 4,532,348 males (range: 3,040,4476,024,250) and 2,059,709 females (range: 1,513,9692,605,450) ever injected drugs.
We calculated lifetime population proportion estimates for male and female PWID by race/ethnicity and by age group (Table 4). The population proportion of PWID was highest among white males (3.8% [CI: 2.7%4.9%]) and lowest among Hispanic/Latino females (0.7%, [CI: 0.5%1.0%]). The population proportion of PWID increased with age among those aged 1849 years.
HCV Infection Rate among PWID in the U.S: The prevalence rate of HCV infection among PWID aged 4065 years was 43,126 per 100,000 PWID (CI: 34,02458,875).
Discussion
Using data from national populationbased U.S. surveys, we estimated that persons who ever injected drugs comprised 2.6% (CI: 1.8%3.3%) of the U.S. population. This represents approximately 6,612,488 million PWID (range: 4,583,1888,641,788) aged 13 years or older in 2011. Although PWID comprise 3% or less of the U.S. population, they account for 22% of all persons living with HIV infection [2]. Our estimates also quantified the recognized disparity of HIV disease rates among black and Hispanic/Latino male and female PWID when compared with white male and female PWID.
The rates we calculated for living with a diagnosis of HIV infection in 2010 represent approximately 2% among male PWID and 3% among female PWID. In recent years, national HIV seroprevalence data among PWID have originated primarily from NHANES [23], which was a data source in our metaanalysis. More recent data on HIV seroprevalence among pastyear PWID in 20 cities with high AIDS prevalence was 9%, with similar patterns of higher prevalence among black and Hispanic/Latino compared to white PWID [24]. However, those data on pastyear PWID are not directly comparable to our rates which were based on lifetime PWID.
Our past year estimate represents about 774,434 PWID in 2011 (range: 494,6051,054,263). Tempalski and colleagues found a 2007 population estimate for PWID of approximately 1,500,000 with minimum and maximum estimates of approximately 1,300,000 and 1,700,000 million [10]. In view of the differences between our study and that by Tempalski and colleagues in terms of methods (metaanalysis vs. multiplier methods), datasets (national surveys vs. drug and HIV testing data), and time periods (20002008 vs. 19922007), and the lack of an accepted gold standard method for PWID population size estimates [6], it is unclear whether our estimate represents underestimation or the Tempalski method represents overestimation. The use of any PWID population size estimate should be accompanied by acknowledgement of the limitations of the methods and data sources.
Our rate of adults and adolescents living with a diagnosis of HIV infection among PWID in Puerto Rico (14%) is higher than that published by Perez and colleagues (2.8% [CI 0.6%12.4%]) [22]. The estimate of the population proportion of PWID in Puerto Rico (1.5%), based on a single household survey, could be an underestimate. In addition, the investigators cautioned that the small number of HIVinfected persons limited their ability to make reliable prevalence estimates stratified by injection drug use [22].
Our HCV infection prevalence rate among PWID aged 4065 years was 43,126 per 100,000 population, reflecting the substantial impact of injection drug use on acquiring HCV infection. The HCV infection prevalence we found (43.1%) was similar to a previous NHANES estimate of 48% [4]; the higher NHANES prevalence may be due to the inclusion of PWID in a broader age range than our estimate. The prevalence estimate points to the importance of national efforts to raise awareness of HCV testing among persons who have injected drugs [25]. In an era of improved treatment, it is also important that those who are infected are linked to appropriate care [11]. CDC recommends integrated prevention services for PWID, which address risk for HIV and HCV infections and are expected to result in increased access to services, improved timeliness of service delivery, and increased effectiveness of prevention efforts [26].
Our results are subject to several limitations. While the study designs are robust in the 4 national surveys, they are hampered by small proportion of participants reporting injection drug use. Because PWID are a small proportion of the general population, obtaining adequate numbers to produce stable estimates is difficult without very large sample sizes. This difficulty is exacerbated when estimates are stratified by sex, age, or race/ethnicity. In addition, the illicit nature of and stigma associated with injection drug use may have resulted in underreporting of this behavior; however this bias should be mitigated in part by use of ACASI for most surveys included in our analysis. A second limitation is coverage bias. The surveys in the metaanalysis exclude individuals without stable housing. Given that a high proportion of PWID are unstably housed [27], [28], they are likely underrepresented in our analysis.
This coverage bias would result in an underestimate of the population proportion of PWID and an overestimate of disease rates. A third limitation is the degree of heterogeneity among surveys. Although all surveys are populationbased, the sampling methods, age range, and question wording vary across surveys. We used randomeffects models to account for variance beyond sampling errors. As noted, the rates among those aged 1324 years and 50 years or older may be underestimates given that the metaanalysis was limited to those aged 1864 years. Other limitations are inherent from the surveillance data used in the rate calculations [5]. The HCV infection data are subject to the limitations of NHANES [4].
Given the potential factors affecting the data in the 4 surveys and the surveillance data, the population estimate and disease rates should be presented with acknowledgement of their limitations and interpreted in the context of the confidence intervals presented; wider confidence intervals for some groups indicate less precision in the estimates. The estimation method presented here (metaanalysis results of ongoing, national survey data) represents one method, as our expert consultants recognized, for estimating the size of the PWID population in the United States. As more research is conducted to estimate population size of groups at risk for HIV and HCV infection, we will consider using different methods in the future, should they prove more accurate or more tractable than metaanalysis of national survey data. These methods include capturerecapture, using data collected from the population at risk, and network scaleup method based on data collected from the general population [6]. For purposes of determining whether there are better methods or data sources for use in the future, our expert consultants recommended smallscale feasibility studies of multiple estimation methods in one geographic area to be able to compare estimates generated by these different methods (e.g., national survey, network scaleup, and capturerecapture). However, for the short term, synthesizing national surveys was recommended by the expert consultants, the approach we took in this paper.
Estimating the population proportion of PWID allowed calculation of rates of HIV and HCV infection, which quantifies the disproportionate impact among PWID nationally. Trends from populationbased surveys will be monitored as part of CDC's behavioral surveillance analyses and the metaanalysis can be updated as new data emerge. For HIV infection, rates can be calculated on an annual basis with the most recent surveillance data. Other disease metrics can be used to calculate rates, such as HIV incidence [1] or national HIV prevalence estimates [2], which include persons with undiagnosed HIV infection. Because we calculated past year as well as lifetime estimates, others can use either, depending on which best fits their needs. However, our estimates may not be well suited for calculating disease rates at the state or local level as the population sizes of PWID vary across the U.S. [10].
The best available data must be used to guide decisionmaking for disease prevention. The estimate of the number of PWID (lifetime and past year) in the U.S. and quantifying the burden of disease and disparities among PWID can be particularly important for planning and evaluating programs serving disproportionately affected populations and addressing health inequities at the national level. The estimate of the number of PWID in the U.S. and resulting rates are important additions to cost effectiveness and other data used to make critical decisions about resources for prevention of HIV and HCV infections.

Estimating the Number of Persons Who Inject Drugs in the United States by MetaAnalysis to Calculate National Rates of HIV and Hepatitis C Virus Infections
Published: May 19, 2014
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0097596
Abstract
Background
Injection drug use provides an efficient mechanism for transmitting bloodborne viruses, including human immunodeficiency virus (HIV) and hepatitis C virus (HCV). Effective targeting of resources for prevention of HIV and HCV infection among persons who inject drugs (PWID) is based on knowledge of the population size and disparity in disease burden among PWID. This study estimated the number of PWID in the United States to calculate rates of HIV and HCV infection.
Methods
We conducted metaanalysis using data from 4 national probability surveys that measured lifetime (3 surveys) or pastyear (3 surveys) injection drug use to estimate the proportion of the United States population that has injected drugs. We then applied these proportions to census data to produce population size estimates. To estimate the disease burden among PWID by calculating rates of disease we used lifetime population size estimates of PWID as denominators and estimates of HIV and HCV infection from national HIV surveillance and survey data, respectively, as numerators. We calculated rates of HIV among PWID by gender, age, and race/ethnicity.
Results
Lifetime PWID comprised 2.6% (95% confidence interval: 1.8%3.3%) of the U.S. population aged 13 years or older, representing approximately 6,612,488 PWID (range: 4,583,1888,641,788) in 2011. The population estimate of pastyear PWID was 0.30% (95% confidence interval: 0.19 %0.41%) or 774,434 PWID (range: 494,6051,054,263). Among lifetime PWID, the 2011 HIV diagnosis rate was 55 per 100,000 PWID; the rate of persons living with a diagnosis of HIV infection in 2010 was 2,147 per 100,000 PWID; and the 2011 HCV infection rate was 43,126 per 100,000 PWID.
Conclusion
Estimates of the number of PWID and disease rates among PWID are important for program planning and addressing health inequities.
Introduction
Injection drug use provides an efficient mechanism for transmitting bloodborne viruses, including human immunodeficiency virus (HIV), hepatitis C virus (HCV), and hepatitis B virus (HBV). In the United States (U.S.), 8% of all new HIV infections in 2010 were among persons who inject drugs (PWID) and 3% were among PWID who also engaged in malemale sex [1]. In 2010, PWID comprised 22% of adults and adolescents living with HIV infection in the United States [2]. PWID are estimated to comprise about 16% of persons with acute HBV infection [3]. A national probability survey, conducted from 1999 through 2002, showed that 48% of adults aged 2059 years who tested antibody positive for HCV reported a history of injection drug use [4].
The disparity in disease burden among PWID compared to their population size has been difficult to quantify. Although the Centers for Disease Control and Prevention (CDC) routinely uses population data from the Census Bureau to calculate disease rates by selected demographic categories (e.g., sex, race/ethnicity, and age at diagnosis) [3], [5] no census data are available for the number of PWID in the U.S. and rate calculations require this number for the denominator. Rates allow for comparison among subgroups and over time. Several methods have been used by various countries to measure the size of populations of PWID, including: 1) the capturerecapture method, using data collected from the population at risk; 2) the multiplier method, based on existing data; and 3) the network scaleup method, based on data collected from the general population [6]. For the U.S., multiple data sources have been compiled to estimate the population size of PWID among the nation as a whole and for large metropolitan areas [7][10]. While these estimates are informative, they are based on pastyear behavior, which is not the most relevant time period for calculating disease rates from national HIV surveillance data, which essentially measures lifetime behaviors [5], or for calculating rates from national hepatitis C survey data since ever use of injection drugs, even in the distant past, is a risk for HCV infection [11].
Recently, CDC used metaanalysis to estimate the proportion of the U.S. population who are men who have sex with men (MSM) and quantify the burden of HIV and sexually transmitted diseases among MSM [12]. Population size estimates together with census and surveillance data were used to calculate disease rates among MSM. Applying this established method, in this report we estimate the population proportion of PWID and quantify the burden of HIV and HCV infections among PWID. We conducted a metaanalysis of national surveys to estimate the proportion of persons in the U.S. who have injected drugs, used these estimates to calculate disease metrics for PWID using national surveillance data for HIV infection, and calculated rate ratios by gender, race/ethnicity, and age. We used other methods to estimate rates of HIV infection among PWID in Puerto Rico and rates of HCV infection among adult PWID in the U.S. The estimates of the number of PWID and rates of HIV and HCV infection among PWID are needed to effectively plan, implement at an appropriate scale, and evaluate programs that serve PWID with or at risk for bloodborne infections such as HIV, HBV, or HCV.






