Female and male differences in AIDS diagnosis rates among people who inject drugs in large U.S. metro areas from 1993 to 2007.

Annals of Epidemiology 25 (2015) 218e225

Contents lists available at ScienceDirect

Annals of Epidemiology journal homepage: www.annalsofepidemiology.org

Original article

Female and male differences in AIDS diagnosis rates among people who inject drugs in large U.S. metro areas from 1993 to 2007 Brooke S. West PhD a, *, Enrique R. Pouget PhD b, Barbara Tempalski PhD b, Hannah L.F. Cooper ScD c, H. Irene Hall PhD d, Xiaohong Hu PhD d, Samuel R. Friedman PhD b a

Division of Global Public Health, School of Medicine, University of California, San Diego, La Jolla, CA National Development and Research Institutes, Inc., New York, NY Department of Behavioral Sciences and Health Education, Rollins School of Public Health, Emory University, Atlanta, GA d Centers for Disease Control and Prevention, Atlanta, GA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 July 2014 Accepted 9 January 2015 Available online 7 February 2015

Purpose: We estimated female and male incident AIDS diagnosis rates (IARs) among people who inject drugs (PWID) in U.S. metropolitan statistical areas (MSAs) over time to assess whether declines in IARs varied by sex after combination antiretroviral therapy (cART) dissemination. Methods: We compared IARs and 95% confidence intervals for female and male PWID in 95 of the most populous MSAs. To stabilize estimates, we aggregated data across three-year periods, selecting a period immediately preceding cART (1993e1995) and the most recent after the introduction of cART for which data were available (2005e2007). We assessed disparities by comparing IAR 95% confidence intervals for overlap, female-to-male risk ratios, and disparity change scores. Results: IARs declined an average of 58% for female PWID and 67% for male PWID between the pre-cART and cART periods. Among female PWID, IARs were significantly lower in the later period relative to the pre-cART period in 48% of MSAs. Among male PWID, IARs were significantly lower over time in 86% of MSAs. Conclusions: IARs among female PWID in large U.S. MSAs have declined more slowly than among male PWID. This suggests a need for increased targeting of prevention and treatment programs and for research on MSA level conditions that may drive differences in declining AIDS rates among female and male PWID. Ó 2015 Elsevier Inc. All rights reserved.

Keywords: Injection drug use AIDS Gender Health status disparities

Introduction After the introduction of combination antiretroviral therapy (cART), there was a substantial reduction in the number of AIDS diagnoses in the United States, particularly among people who inject drugs (PWID) [1]; however, national trajectories of change in AIDS diagnoses have not been the same for female and male PWID [1,2]. In 1995, PWID represented seven percent of persons This study was supported by grants from the U.S. National Institute on Drug Abuse (R01 DA013336 and R01 DA037568). The authors acknowledge the NIHfunded Center for Drug Use and HIV Research (P30 DA121041) for its support and assistance. The findings and conclusions in this study are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention. The authors have no conflicts of interest to disclose. * Corresponding author. Division of Global Public Health, School of Medicine, University of California, San Diego, 9500 Gilman Drive 0507, La Jolla, CA 92093. Tel.: þ1 858-822-4393; fax: þ1 858-534-7566. E-mail address: [email protected] (B.S. West). http://dx.doi.org/10.1016/j.annepidem.2015.01.006 1047-2797/Ó 2015 Elsevier Inc. All rights reserved.

diagnosed with AIDS among females and 19% among males [2]. In 2011, 20% of new AIDS diagnoses among females were injection related, whereas only 10% of AIDS diagnoses among males were from PWID [1]. During this time, we have estimated that PWID prevalence per 10,000 adult population in large U.S. metropolitan statistical areas (MSAs) also fell from 157 in the period from 1993 to 1995 to 133 in the period from 2005 to 2007 among males, and from 82 to 75 among females [3]. These data raise questions about changes in epidemiologic patterns of AIDS among female PWID and how these patterns may vary across geographic areas. This article focuses on differences between female and male PWID in incident AIDS diagnosis rates (IARs) across MSAs. Analysis at the MSA level provides a useful lens through which to understand the historical and social factors that drive HIV burden in different locales. Several articles have found differences in HIV prevalence and AIDS mortality among PWID across MSAs, and this variation has been linked to MSA-level factors [4e6]. Previous research has also demonstrated that racial and/or ethnic disparities

B.S. West et al. / Annals of Epidemiology 25 (2015) 218e225

in AIDS diagnoses among PWID varied significantly across MSAs [7]. Although there is ample evidence demonstrating the success of cART in reducing morbidity and mortality in the United States [8,9], the diffusion of cART and other important services may not have been uniform across MSAs and may have reached male and female injectors at different times. Disparities in the availability of cART and HIV care, as well as its accessibility to female PWID, may have resulted in significant sex differences in AIDS diagnoses across MSAs among injectors. In this article, we assess changes in IARs among female and male PWID across large MSAs between a precART period (1993e1995) and a period well after cART was available (2005e2007). Methods and materials The unit of analysis in this article is the MSA, which is defined by the U.S. Census Bureau as contiguous counties containing a central city of 50,000 people or more and that form a socioeconomic unity [10]. Studying HIV among PWID at the MSA level [11] is useful because, as noted, each MSA has its own epidemic history and HIV prevalence rate. Analyses were conducted on a cohort of 95 of the largest MSAs in the continental United States. Preliminary analysis of trend lines over time from 1992 to 2007 (not shown) demonstrated that rates of AIDS diagnosis fell for both male and female PWID over time but fell more rapidly beginning in 1996 when cART was first introduced. In this article, we focus on long-term change between the period before cART and a time when cART was more widely available, treating these two time periods as summary end points of linear change. To minimize the potential impact of small numbers of diagnoses in some MSAs, we combined data from the 3years immediately preceding cART dissemination (1993e1995) and compared this to a 3-year period (2005e2007) with the most recent available data. We estimated metropolitan IARs among adult (aged 15e64 years) female and male PWID using Centers for Disease Control and Prevention’s (CDC) National HIV Surveillance System data for 95 MSAs. Female PWID included females who reported injection drug use as a risk factor. Males who reported injecting drugs and males who were both PWID and men who have sex with men (MSM) were included as male PWID. For each MSA, period and sex, we calculated IARs by dividing the total number of AIDS diagnoses during that period by the estimated number of male or female PWID without AIDS. Methods for creating annual estimates of PWID for each MSA for females and males have been reported in detail elsewhere but involve calculating the number of PWID in the United States and then apportioning estimates to MSAs using multiplier methods [3,11e13]. These estimates were based on data on HIV counseling and testing, drug treatment, AIDS diagnoses, and estimates from published national and MSA-specific research studies [3,13]. PWID living with AIDS were excluded from the denominators to calculate incidence for the PWID population at risk for AIDS. PWID at risk for AIDS could be HIV infected or HIV uninfected. These estimates were then scaled (multiplied by 10,000) to provide IARs per 10,000 PWID for both females and males. At the time we conducted these analyses, our estimates of PWID prevalence were only available through 2007. Data on PWID AIDS diagnoses stratified by both sex and race were not available because of the suppression of CDC data with small cell sizes. To facilitate comparisons, we calculated 95% confidence intervals (CIs) for each IAR. CIs were calculated using standard formulas when the number of new diagnoses in an MSA was 100 or more and using tabled values when the number of diagnoses was smaller [14]. The tabled values formula assumes that AIDS diagnoses are infrequent events relative to the size of the population and so can be viewed as occurring in a Poisson-distributed function

219

[14]. As a result, the size of the CIs reflects both the rates and numbers of diagnoses so that MSAs with few diagnoses have relatively low rates and wide CIs and MSAs with many diagnoses have higher rates and more narrow CIs. We also present the dispersion of IARs across MSAs for female and male PWID using a measure of the coefficient of variation (CV). We used various methods to assess absolute and relative disparities in IARs for female and male PWID. First, to assess important differences in IARs within a given MSA and time period, we compared the extent to which CIs overlapped for females and males: if the CI overlapped, the difference was considered to be not significant at the 95% level [14,15]. The CIs are based on standard error estimates that reflect only random error in the numerator. Comparing CIs for overlap is an absolute comparison that accounts for the precision of the IARs, allowing us to use all available data without excluding MSAs with small numbers of AIDS diagnoses and providing a way to assess the frequency and importance of sex differences across MSAs. Second, we assessed the magnitude of sex differences in IARs among PWID in each period by calculating the rate ratio for female PWID compared with male PWID (female IAR/male IAR) for each MSA [16]. It is important to note that rate ratios are sensitive to small values so differences may appear to vary greatly within an MSA with small but changing numbers of diagnoses; therefore, we also provide a calculation of relative disparity, the percent difference between groups, which is more appropriate for MSAs with lower diagnosis rates [16]. Third, we present disparity change scores (DCSs) for each MSA to track differences in IARs for female and male PWID over time. The DCS indicates the change in the female-to-male rate ratio in each MSA between the early and late period by subtracting the rate of the health outcome at the later time point from the rate of the health outcome at baseline using the following formula: DCS ¼ jRRT1 1j jRRT2 1j [16]. A negative DCS indicates that the gap in IARs between females and males is narrowing, with a larger DCS value indicating a greater change in the disparity [16]. We included the DCS as an additional tool with which to identify possible shrinking or growing disparities and to draw attention to those MSAs in which the service needs of female or male PWID may be greater. The geographic distribution of differences was assessed by comparing IARs across region (Northeast, South, Midwest, and West). All analyses were conducted using IBM SPSS Statistics, version 19 [17]. Results AIDS diagnosis rates, 95% CIs, and significance of CIs for female and male PWID in each MSA and time period are listed in Table 1. Across all 95 MSAs, in the pre-cART period, IARs among females (per 10,000 female PWID) averaged 103 (SD ¼ 124.1). In the later period, after the introduction of cART, the average IAR among females was 43 (SD ¼ 44.9). Among males, the average IAR (per 10,000 male PWID) was 163 (SD ¼ 138.6) in the pre-cART period and 53 (SD ¼ 37.4) in the later period. IARs were more dispersed for female PWID (CV ¼ 120%) than male PWID (CV ¼ 85%) in the precART period. Similarly, in the cART period, IARs were more dispersed for female PWID (CV ¼ 105%) compared with male PWID (CV ¼ 71%). Differences between female and male IARs were present in both the pre-cART and cART periods. As seen in Table 1, in 94 of 95 MSAs, IARs in the pre-cART period were lower among female PWID than male PWID; differences between male and female IARs were significant in 65 of these 94 MSAs as indicated by nonoverlapping CIs. In the cART era, however, IARs among females were significantly

220


Table 1 IARs per 10,000 PWID and CIs for the periods from 1993 to 1995 and 2005 to 2007 by MSA in the United States MSA

Mean (SD) Median (range) CV* % Akron, OH AlbanyeSchenectadyeTroy, NY Albuquerque, NM AllentowneBethlehemeEaston, PA Ann Arbor, MI Atlanta, GA AustineSan Marcos, TX Bakersfield, CA Baltimore, MD Bergene Passaic, NJ Birmingham, AL Boston, MAeNH BuffaloeNiagara Falls, NY CharlestoneNorth Charleston, SC CharlotteeGastoniaeRock Hill, NCeSC Chicago, IL Cincinnati, OHeKYeIN ClevelandeLoraineElyria, OH Columbus, OH Dallas, TX DaytoneSpringfield, OH Denver, CO Detroit, MI El Paso, TX Fort Lauderdale, FL Fort WortheArlington, TX Fresno, CA Gary, IN Grand RapidseMuskegoneHolland, MI GreensboroeWinstonSalemeHigh Point, NC GreenvilleeSpartanburgeAnderson, SC HarrisburgeLebanoneCarlisle, PA Hartford, CT Honolulu, HI Houston, TX Indianapolis, IN Jacksonville, FL Jersey City, NJ Kansas City, MOeKS Knoxville, TN Las Vegas, NVeAZ Little RockeNorth Little Rock, AR Los AngeleseLong Beach, CA Louisville, KYeIN Memphis, TNeAReMS Miami, FL MiddlesexeSomerseteHunterdon, NJ MilwaukeeeWaukesha, WI MinneapoliseSt. Paul, MNeWI MonmoutheOcean, NJ Nashville, TN NassaueSuffolk, NY New HaveneMeriden, CT New Orleans, LA New York, NY Newark, NJ NorfolkeVirginia BeacheNewport News, VAeNC Oakland, CA Oklahoma City, OK Omaha, NEeIA Orange County, CA Orlando, FL Philadelphia, PAeNJ PhoenixeMesa, AZ Pittsburgh, PA

Pre-cART era (1993e1995)

cART era (2005e2007)

IAR per 10,000 female PWID: rate (95% CI)

IAR per 10,000 male PWID: rate (95% CI)

103.5 (124.1) 57.0 (3.8e648.4) 119.9

163.0 (138.6) 107.4 (19.8e731.4) 85.1

18.2 202.6 3.8 65.2 19.0 149.3 63.0 7.3 210.1 341.3 89.3 121.8 92.5 141.1 64.4 211.3 29.8 58.3 14.5 43.4 30.4 25.6 99.3 8.0 372.9 40.7 22.1 33.7 44.8 57.5

(5.9e42.5) (153.9e261.9) (1.0e9.8) (45.9e89.9) (3.9e55.4) (131.7e167.0) (50.4e77.8) (3.6e13.1) (195.7e224.4) (299.8e382.8) (58.9e130.0) (111.0e133.0) (68.6e121.9) (89.5e211.7) (47.3e85.7) (195.0e227.5) (18.7e45.2) (44.5e75.1) (7.9e24.3) (35.8e51.0) (16.2e52.0) (19.1e33.7) (86.9e111.6) (2.9e17.4) (328.5e417.2) (33.0e49.6) (15.5e30.4) (18.4e56.6) (23.1e78.2) (40.3e79.6)

36.9 313.4 40.0 70.9 85.0 298.2 94.2 95.7 333.3 415.9 140.5 180.2 168.7 288.4 182.1 299.3 70.3 99.3 72.0 130.6 66.1 70.7 129.3 19.8 414.6 77.8 34.9 63.7 54.7 126.0


101.2 80.0 256.8 27.4 53.9 32.3 92.2 453.7 25.9 13.5 46.3 24.3 32.4 18.2 72.3 393.6 338.9 79.5 38.8 239.9 58.9 333.8 253.0 81.7 648.4 558.3 83.6

(63.4e153.3) (53.6e114.8) (222.2e291.4) (14.2e47.9) (47.4e60.4) (21.3e47.0) (74.6e112.8) (402.3e505.0) (16.2e39.1) (5.0e29.5) (35.5e59.4) (11.1e46.2) (28.8e36.0) (10.4e29.6) (50.6e100.1) (359.3e427.9) (289.0e388.9) (56.3e109.1) (26.5e54.7) (201.6e278.1) (42.1e80.2) (294.1e373.4) (226.1e279.9) (65.7e97.7) (631.1e665.8) (522.8e593.8) (67.0e103.1)

214.4 112.1 303.1 54.0 105.6 76.8 178.9 562.3 115.8 39.8 103.6 40.7 82.8 65.8 145.4 482.9 370.1 90.3 89.6 350.9 138.6 395.0 295.0 139.7 731.4 643.5 183.5


57.0 24.5 41.1 28.4 131.9 154.7 33.7 43.7

(48.3e65.7) (14.5e38.8) (19.7e75.6) (22.0e36.1) (106.1e162.2) (142.3e167.1) (25.6e43.5) (32.6e57.3)

97.1 74.9 78.5 63.0 285.6 195.0 81.1 73.2

(88.0e106.1) (60.4e91.7) (55.8e107.3) (55.5e70.5) (256.1e315.0) (186.4e203.6) (71.9e90.4) (61.8e84.6)

Significance of CIy

z z

z z z z

z z z z z z z z z z z

z

z z

z z z z z z z z z z

z z z

z z z z z z z z z z z



42.6 (44.9) 27.0 (2.3e272.7) 105.2

52.8 (37.4) 41.0 (6.8e240.8) 70.8

14.9 65.9 5.9 32.4 12.0 84.7 32.4 3.4 58.2 87.5 25.2 25.0 45.5 83.0 69.0 66.7 13.7 28.0 14.4 22.0 6.4 18.7 35.9 8.9 165.8 26.0 17.2 18.1 26.9 39.5


25.1 63.8 17.8 41.6 30.2 96.3 54.1 47.2 70.4 82.1 57.6 23.0 37.8 86.7 75.6 109.5 21.8 38.3 30.1 37.9 23.9 32.0 42.5 35.5 166.3 33.4 17.9 17.9 37.4 40.4


51.1 21.6 83.3 27.0 80.1 25.7 61.8 165.4 21.6 10.5 22.5 15.8 22.8 27.2 45.2 207.1 63.6 33.3 27.0 42.5 36.3 60.7 80.6 61.1 272.7 155.6 27.8


69.0 35.7 64.3 25.7 49.7 42.9 132.7 143.1 86.0 26.3 36.3 35.0 35.4 26.3 58.4 240.8 63.5 25.2 30.5 58.0 35.1 47.3 87.4 74.4 122.5 136.2 43.9


(13.4e26.0) (6.6e23.5) (20.9e64.0) (8.8e22.3) (51.6e82.4) (52.4e66.3) (17.9e31.5) (5.3e14.7)

32.5 64.3 35.8 22.7 131.1 74.2 36.3 14.9

(26.5e38.6) (49.0e83.0) (21.5e55.9) (17.6e28.8) (113.3e148.9) (68.5,79.8) (31.5e41.1) (10.6e20.3)

18.9 13.1 38.2 14.5 65.7 59.4 24.0 9.2


z z

z

z

z

z z z

z

z

z z

z z z

(continued on next page)


221

Table 1 ( continued) MSA

PortlandeVancouver, OReWA ProvidenceeFall Rivere Warwick, RIeMA RaleigheDurhameChapel Hill, NC RichmondePetersburg, VA RiversideeSan Bernardino, CA Rochester, NY Sacramento, CA St. Louis, MOeIL Salt Lake CityeOgden, UT San Antonio, TX San Diego, CA San Francisco, CA San Jose, CA SarasotaeBradenton, FL ScrantoneWilkes-BarreeHazleton, PA SeattleeBellevueeEverett, WA Springfield, MA StocktoneLodi, CA Syracuse, NY Tacoma, WA TampaeSt. PetersburgeClearwater, FL Toledo, OH Tucson, AZ Tulsa, OK Ventura, CA Washington, DCeMDeVAeWV West Palm BeacheBoca Raton, FL Wichita, KS WilmingtoneNewark, DEeMD YoungstowneWarren, OH * y z

Pre-cART era (1993e1995)

cART era (2005e2007)




13.6 (9.4e18.9) 151.1 (120.7e186.8)

50.2 (43.1e57.2) 189.6 (163.8e215.4)

z

4.8 (2.6e8.2) 43.2 (29.9e60.3)

74.1 74.8 28.9 328.9 20.1 24.0 29.3 16.4 33.0 75.6 21.2 91.2 43.8 15.4 171.4 17.8 134.5 24.3 89.1 13.8 10.1 19.3 16.1 224.0 151.9 37.3 217.6 24.9

145.4 155.6 96.8 293.3 65.2 68.0 52.8 57.4 107.4 264.2 53.5 117.0 162.0 76.7 208.1 40.5 255.9 53.3 178.2 45.5 48.7 62.5 34.8 337.9 185.0 118.1 340.8 88.7

z z z

73.9 31.5 18.9 93.9 9.3 25.1 4.5 16.7 9.6 26.6 9.8 16.3 35.6 8.7 46.2 4.3 83.0 4.7 47.5 3.1 10.5 13.2 2.3 99.2 89.0 12.0 52.7 8.8

(54.1e99.2) (56.6e96.9) (23.2e34.5) (277.3e380.5) (14.5e27.1) (16.4e33.8) (18.4e44.4) (11.1e23.4) (26.7e40.3) (65.7e85.6) (14.3e30.3) (61.1e130.9) (17.6e90.1) (10.9e21.2) (137.8e210.7) (10.2e28.9) (89.3e194.3) (15.0e37.1) (73.8e104.4) (3.8e35.4) (5.0e18.0) (9.2e35.4) (7.7e29.5) (206.2e241.8) (124.7e179.1) (15.0e76.9) (177.1e258.1) (6.8e63.6)


z z z z z z z z z z z z z z z z z



IAR per 10,000 male PWID: rate (95% CI) 24.4 (19.8e29.9) 27.1 (19.8e36.3) 56.2 48.8 36.6 59.9 20.2 40.1 14.1 43.0 40.9 68.9 42.5 23.5 53.0 26.8 27.4 18.1 83.6 24.2 71.0 24.7 16.7 44.0 6.8 106.3 84.0 49.6 63.8 27.9


Significance of CIy z

z z z z z z z

z z z z

z

z

Absolute value of CV. Significance determined by there not being an overlap of 95% CI for IARs among females relative to males. P < 0.05.

lower than rates among male PWID in only 27 MSAs and significantly higher in two MSAs (Houston and New York). On average, across all 95 MSAs, IARs decreased by 67% for male PWID between the pre-cART and cART era, but only by 58% for female PWID (not shown in the table). When we look within MSAs, we see that diagnosis rates over time among female PWID were significantly lower in only 48% of MSAs (n ¼ 46) and significantly higher in Houston, TX. In the remaining 48 MSAs, there was no significant change in IARs among female PWID over time. Among male PWID, IARs in the cART period were significantly lower in 86% of MSAs (n ¼ 82) and showed no significant change in the remaining 13 MSAs. Disparities between female and male IARs across MSAs and within and across time periods, including female-to-male rate ratios, relative difference in IAR by sex, and DCSs, are listed in Table 2. Over time, the female-to-male IAR ratio increased in a number of MSAs: in the period from 1993 to 1995, on average across all 95 MSAs, the female-to-male PWID rate ratio was 0.5, but this increased to 0.7 in the period from 2005 to 2007. In other words, in the pre-cART period, IARs among female PWID were about 47% lower than for male PWID, but in the cART era, they were only about 26% lower than among males. These results are supported by the DCS as well. On average, across all 95 MSAs, the DCS was 0.2, indicating that the gap between female and male PWID IARs narrowed by 21% over time, with the magnitude of dispersion across MSAs being 1.4 times smaller than the mean (CV ¼ 141%). In 15 of 95 MSAs, the gap between female and male IARs increased by 10% or more over time as a result of declining female-to-male rate ratios, which suggests that

female PWID were faring better relative to male PWID. For 59 MSAs, however, the difference between IARs among female and male PWID narrowed by 10% or more; in 21 of these 59 MSAs, the femaleto-male rate ratio in the cART period was 1.00 or greater, and in 12 MSAs the disparity narrowed by 50% or more, as indicated by the DCS. This “narrowing” of the female-to-male disparity indicates that female’s IARs decreased less over time and became more similar to rates among males. As seen in Table 3, regional analysis suggests that sex differences in IARs among PWID may vary by region. In the pre-cART period, IARs for male and female PWID were highest in the Northeast (male ¼ 300; female ¼ 239) followed by the South (male ¼ 179; female ¼ 104). These two regions also had the largest female-tomale rate ratios of 0.8 in the Northeast and 0.5 in the South. During the cART era, in the period from 2005 to 2007, IARs among female PWID remained the highest in the Northeast (75) followed by the South (54), but among male PWID, rates were higher in the South (70) than the Northeast (64). Although female-to-male rate ratios were greater in all regions in the later period, it is important to note that in the Northeast, where the PWID epidemic started, the sex ratio reversed, such that female IARs were larger relative to male rates (1.12). The DCS over time ranged from about 9% in the West to 36% in the Northeast, which indicates that female and male IARs are becoming more closely aligned. Discussion AIDS diagnosis rates among both female and male PWID decreased after the introduction of cART; however, declines were

222


Table 2 Differences in female-to-male IAR ratios among PWID, relative differences, and DCS over time for the periods 1993 to 1995 and 2005 to 2007 by MSA in the United States MSA

Mean (SD) Median (range) CV* % Akron, OH AlbanyeSchenectadyeTroy, NY Albuquerque, NM AllentowneBethlehemeEaston, PA Ann Arbor, MI Atlanta, GA AustineSan Marcos, TX Bakersfield, CA Baltimore, MD BergenePassaic, NJ Birmingham, AL Boston, MAeNH BuffaloeNiagara Falls, NY CharlestoneNorth Charleston, SC CharlotteeGastoniae Rock Hill, NCeSC Chicago, IL Cincinnati, OHeKYdIN ClevelandeLoraineElyria, OH Columbus, OH Dallas, TX DaytoneSpringfield, OH Denver, CO Detroit, MI El Paso, TX Fort Lauderdale, FL Fort WortheArlington, TX Fresno, CA Gary, IN Grand RapidseMuskegone Holland, MI GreensboroeWinstonSalemeHigh Point, NC GreenvilleeSpartanburge Anderson, SC HarrisburgeLebanoneCarlisle, PA Hartford, CT Honolulu, HI Houston, TX Indianapolis, IN Jacksonville, FL Jersey City, NJ Kansas City, MOeKS Knoxville, TN Las Vegas, NVeAZ Little RockeNorth Little Rock, AR Los AngeleseLong Beach, CA Louisville, KYeIN Memphis, TNeARdMS Miami, FL MiddlesexeSomerseteHunterdon, NJ MilwaukeeeWaukesha, WI MinneapoliseSt. Paul, MNeWI MonmoutheOcean, NJ Nashville, TN NassaueSuffolk, NY New HaveneMeriden, CT New Orleans, LA New York, NY Newark, NJ NorfolkeVirginia BeacheNewport News, VAeNC Oakland, CA Oklahoma City, OK Omaha, NEeIA Orange County, CA Orlando, FL Philadelphia, PAeNJ

Pre-cART era (1993 to 1995)

cART era (2005 to 2007)

Female-to-male rate ratio

Relative difference (%)



0.53 (0.22) 0.50 (0.08 to 1.12) 40.94

47.14 (21.64) 49.98 (92.38 to 12.12) 45.91

0.74 (0.39) 0.67 (0.07 to 2.23) 52.91

26.38 (38.95) 33.06 (92.89 to 122.58) 147.68

DCS over time

0.21 (0.29) 0.18 (1.33 to 0.26) 141.49

0.49 0.65 0.10 0.92 0.22 0.50 0.67 0.08 0.63 0.82 0.64 0.68 0.55 0.49 0.35

50.68 35.35 90.42 8.01 77.70 49.91 33.13 92.38 36.98 17.93 36.42 32.38 45.19 51.08 64.62

0.59 1.03 0.33 0.78 0.40 0.88 0.60 0.07 0.83 1.07 0.44 1.08 1.21 0.96 0.91

40.58 3.27 67.17 22.27 60.22 12.03 40.03 92.89 17.30 6.53 56.18 8.49 20.51 4.27 8.74

0.10 0.39 0.23 0.14 0.17 0.38 0.07 0.01 0.20 0.25 0.20 0.41 0.66 0.47 0.56

0.71 0.42 0.59 0.20 0.33 0.46 0.36 0.77 0.40 0.90 0.52 0.63 0.53 0.82

29.42 57.57 41.27 79.87 66.78 54.00 63.78 23.23 59.72 10.07 47.67 36.80 47.01 18.11

0.61 0.63 0.73 0.48 0.58 0.27 0.59 0.84 0.25 1.00 0.78 0.96 1.01 0.72

39.05 37.34 26.71 52.19 41.97 73.41 41.49 15.54 74.84 0.26 21.99 3.94 1.16 27.97

0.10 0.20 0.15 0.28 0.25 0.19 0.22 0.08 0.15 0.10 0.26 0.33 0.48 0.10

0.46

54.38

0.98

2.22

0.52

0.47

52.77

0.74

25.89

0.27

0.71 0.85 0.51 0.51 0.42 0.52 0.81 0.22 0.34 0.45 0.60 0.39 0.28 0.50 0.81 0.92 0.88 0.43 0.68 0.42 0.84 0.86 0.59 0.89 0.87 0.46

28.68 15.27 49.22 48.96 57.92 48.45 19.32 77.67 66.04 55.28 40.32 60.81 72.33 50.29 18.51 8.41 11.94 56.71 31.64 57.50 15.51 14.24 41.49 11.35 13.24 54.45

0.61 1.30 1.05 1.61 0.60 0.47 1.16 0.25 0.40 0.62 0.45 0.65 1.03 0.77 0.86 1.00 1.32 0.89 0.73 1.03 1.28 0.92 0.82 2.23 1.14 0.63

39.42 29.59 5.00 61.40 40.04 53.45 15.59 74.92 60.15 38.17 54.78 35.38 3.35 22.60 13.98 0.22 32.08 11.47 26.70 3.38 28.25 7.73 17.95 122.58 14.25 36.63

0.11 0.45 0.54 1.10 0.18 0.05 0.35 0.03 0.06 0.17 0.15 0.25 0.76 0.28 0.05 0.09 0.44 0.45 0.05 0.61 0.44 0.07 0.24 1.34 0.27 0.18

0.59 0.33 0.52 0.45 0.46 0.79

41.27 67.22 47.62 54.89 53.79 20.64

0.58 0.20 1.07 0.64 0.50 0.80

41.77 79.60 6.61 36.21 49.94 19.92

0.01 0.12 0.54 0.19 0.04 0.01 (continued on next page)


223

Table 2 ( continued) MSA

PhoenixeMesa, AZ Pittsburgh, PA PortlandeVancouver, OReWA ProvidenceeFall Rivere Warwick, RIeMA RaleigheDurhameChapel Hill, NC RichmondePetersburg, VA RiversideeSan Bernardino, CA Rochester, NY Sacramento, CA St. Louis, MOeIL Salt Lake CityeOgden, UT San Antonio, TX San Diego, CA San Francisco, CA San Jose, CA SarasotaeBradenton, FL ScrantoneWilkesBarreeHazleton, PA SeattleeBellevueeEverett, WA Springfield, MA StocktoneLodi, CA Syracuse, NY Tacoma, WA TampaeSt. Petersburge Clearwater, FL Toledo, OH Tucson, AZ Tulsa, OK Ventura, CA Washington, DCeMDeVAeWV West Palm BeacheBoca Raton, FL Wichita, KS WilmingtoneNewark, DEeMD YoungstowneWarren, OH *


cART era (2005 to 2007)




DCS over time


0.42 0.60 0.27 0.80

58.49 40.28 72.99 20.33

0.66 0.62 0.20 1.59

33.89 38.35 80.39 59.23

0.25 0.02 0.07 0.80

0.51 0.48 0.30 1.12 0.31 0.35 0.56 0.29 0.31 0.29 0.40 0.78 0.27

49.03 51.96 70.18 12.12 69.24 64.77 44.43 71.46 69.29 71.38 60.37 22.06 72.99

1.31 0.64 0.52 1.57 0.46 0.63 0.32 0.39 0.23 0.39 0.23 0.69 0.67

31.44 35.60 48.36 56.75 54.00 37.49 67.98 61.26 76.54 61.37 76.90 30.72 32.81

0.80 0.16 0.22 0.45 0.15 0.27 0.24 0.10 0.07 0.10 0.17 0.09 0.40

0.20 0.82 0.44 0.53 0.46 0.50

79.86 17.65 56.12 47.47 54.44 49.98

0.32 1.69 0.24 0.99 0.19 0.67

67.71 68.58 75.97 0.69 80.62 33.06

0.12 0.86 0.20 0.47 0.26 0.17

0.30 0.21 0.31 0.46 0.66 0.82 0.32 0.64 0.28

69.60 79.37 69.16 53.82 33.71 17.87 68.42 36.15 71.98

0.13 0.63 0.30 0.34 0.93 1.06 0.24 0.83 0.32

87.40 37.42 70.00 66.00 6.71 6.00 75.80 17.37 68.36

0.18 0.42 0.01 0.12 0.27 0.24 0.07 0.19 0.04

Absolute value of CV.

slighter for female PWID relative to male PWID. Although average IARs among males declined 67% across the 95 MSAs, they only declined an average of 58% among females. This is evident in our analysis of overlap in CIs, which showed that between the pre-cART and cART periods, IARs among males declined significantly in 86% of MSAs. For female PWID, IARs declined significantly in only 48% of MSAs. Additional measures of disparity, including female-to-male rate ratios and DCSs, also demonstrate this trend. This pattern of slower declines in IARs among female PWID is particularly apparent in the Northeast and in some other larger MSAs. For example, in Baltimore, Boston, Chicago, Houston, New York, Newark, Philadelphia, and Washington, DC, IARs among female PWID declined 53% compared to 71% among male PWID. Given that females may be at higher epidemiologic and sociobehavioral risk for HIV [18] and that PWID are less likely than persons in other transmission categories to seek early HIV counseling, testing, and treatment [19], to receive CD4 and viral load testing while under care [20], and have a greater risk for HIV-related morbidity and mortality [20,21], these findings warrant public health attention. Although the results of this study show that IARs among female PWID typically remain lower than those of males, the slighter declines among females may suggest that advances in HIV prevention and care over time might be reaching or benefiting male PWID to a greater extent. At a structural level, poorer outcomes for PWID relative to noninjectors have been linked to barriers to antiretroviral access and adherence that stem from sociocultural issues, such as drug policies and stigma, and provider-based barriers, including limited knowledge about substance abuse and physician perceptions [22]. A social determinants of health perspective also points to

sex as an important determinant of health; roles, norms, and inequality shape HIV transmission, access to health services, and how health systems respond to the different needs of females and males [18,23]. At the individual level, studies suggest that a number of factors might be underlying differences in AIDS diagnoses among female and male PWID. In the years soon after cART became widely available, females were less likely than males to receive cART [24e29]. Females infected with HIV were also less likely than males to receive Pneumocystis carinii pneumonia (PCP) prophylaxis [27]. A study assessing causes of mortality among people living with HIV in care during the cART era (1998e2005) found that the percentage of time receiving cART while in care was lower for females and that female sex was the only factor associated with mortality [30]. Furthermore, progression to AIDS and deaths among females and PWID were greater than that for males or non-PWID after the introduction of cART [31]. The combination of being female and a drug injector, however, may make female PWID markedly more vulnerable to poor health outcomes. For instance, a recent study in Canada assessing sex differences in virologic responses to antiretroviral therapy and mortality among PWID and noninjectors, found that female PWID were the least likely to “virologically suppress, mostly likely to experience rebound, and had a greater risk of mortality” [21]. Other studies have shown that being a female PWID in Vancouver is associated with less CD4 monitoring, [32] and compared with females exposed through heterosexual contact, females in 33 U.S. states who injected had lower 3-year survival after HIV diagnosis [19]. Sociopolitical characteristics, stigma, and sex subordination at

224


Table 3 Regional differences in IARs per 10,000 PWID by sex and ratio of female-to-male diagnosis rates for the periods 1993 to 1995 and 2005 to 2007 in the United States Region

Northeast (n ¼ 21) Midwest (n ¼ 20) South (n ¼ 32) West (n ¼ 22)


cART era (2005 to 2007)

DCS over time

IAR per 10,000 female PWID

IAR per 10,000 male PWID


IAR per 10,000 female PWID

IAR per 10,000 male PWID

Mean female-to-male rate ratio

238.8 44.8 104.1 26.8

299.5 90.7 178.6 75.8

0.76 0.46 0.53 0.37

74.9 22.9 54.1 13.1

64.1 38.3 70.5 29.3

1.12 0.64 0.74 0.46

the MSA level may also serve to exacerbate disparities in AIDSrelated outcomes, shaping female PWID’s access to key health interventions, such as cART and syringe exchange programs; however, research on these factors needs to be conducted. The studies discussed previously would suggest that the disparities we find in AIDS diagnoses between male and female PWID stem from differences in rates of successful access to and use of care. Another possible mechanism, which is not analytically explored in this article given our lack of sound data on HIV incidence, is that differences in IARs stem from overall changes in the sex distribution of rates of new HIV infections. This could happen if the rate of new HIV infections was declining more slowly for female PWID than for males. Collectively, our results suggest that greater attention should be paid both to sex differences in PWID HIV diagnoses and access to and utilization of services at the MSA level that may be making female PWID vulnerable to AIDS. More research is needed on the specific factors that may be driving observed sex disparities in AIDS diagnoses across MSAs.

0.36 0.18 0.21 0.09

Conclusions Geographical analysis of MSA-level differences in IARs can illuminate where there have been successes and also where interventions and services are most necessary and with which populations. In this analysis, we found that although there has been substantial progress in reducing IARs among PWID over time, declines among males appear to be greater than declines among females. These findings indicate that female PWID may not be accessing and receiving health services to the same extent as male PWID and therefore suggest that efforts be made to improve their access to and retention in such services. Given potential sociocultural barriers to equality for females in HIV treatment and care, greater efforts are needed in more MSAs to reduce sex differences in AIDS diagnoses; specifically, targeted and comprehensive HIV treatment and prevention programs may be useful for getting female PWID into care more often and sooner. Finally, these results point to a need for research on MSA-level epidemiologic, socioeconomic, and political conditions that may increase risk of HIV infection and/or progression to AIDS for female PWID.

Limitations As our data were aggregated at the MSA level and represent an “ecological cohort” of MSAs, our results and the mechanisms underlying observed disparities should not be assumed to be appropriate for interpretation at the individual level. Because the method of comparing CIs for overlap is a conservative test of significance, some differences may be statistically significant even when the CIs overlapped had other tests been used [15]. We were unable to exclude PWIDeMSM from our analyses as our PWID estimates include PWID who are also MSM, which may skew male results. Furthermore, our PWID estimates, although validated, may include some error. Inaccuracy may also be introduced to the IARs because the time frame for defining someone as a PWID varies between the numerator and the denominator: The denominator is based on drug use in the last year, whereas the surveillance data on AIDS diagnoses classify PWID as ever injection drug use. It is also important to note that the CDC’s AIDS definition changed in 1993; however, these changes were implemented rapidly nationwide [33]. To account for potential bias, we aggregated data for 1993 to 1995 in this early period. We were also unable to obtain data on AIDS diagnoses stratified by sex and race and transmission route simultaneously. Although racial and/or ethnic disparities may intersect with sex in important ways, previous work by Pouget et al. demonstrated that racial disparities changed little across the same time period, suggesting that sex disparities grew in a time when racial disparities remained relatively stable [7]. Finally, we are limited by the unavailability of data on antiretroviral therapy use and adherence rates among PWIDs in these MSAs, as well as estimates on HIV incidence or prevalence among female and male PWID for these time periods. However, we note that changes in HIV incidence and prevalence take time to get reflected in IARs (because of the time between infection and AIDS), which may somewhat mitigate our lack of data on HIV.

References [1] Center for Disease Control and Prevention. Diagnoses of HIV infection in the United States and dependent areas. HIV Surveill Rep 2011;Vol. 23:1e84. [2] Center for Disease Control and Prevention. AIDS associated with injectingdrug useeUnited States, 1995. MMWR 1996;45(19):392e8. [3] Tempalski B, Pouget E, Cleland C, Brady J, Cooper H, Hall H, et al. Trends in the population prevalence of people who inject drugs in US metropolitan areas 1992-2007. PLoS One 2013;8(6):e64789. [4] Tempalski B, Lieb S, Cleland CM, Cooper H, Brady JE, Friedman SR. HIV prevalence rates among injection drug users in 96 large US metropolitan areas, 1992-2002. J Urban Health 2009;86(1):132e54. [5] Friedman SR, Lieb S, Tempalski B, Cooper H, Keem M, Friedman R, et al. HIV among injection drug users in large US metropolitan areas, 1998. J Urban Health-Bulletin of the New York Academy of Medicine 2005;82(3):434e45. [6] Friedman SR, West BS, Pouget ER, Hall HI, Cantrell J, Tempalski B, et al. Metropolitan social environments and pre-HAART/HAART era changes in mortality rates (per 10,000 adult residents) among injection drug users living with AIDS. PLoS One 2013;8(2):e57201. [7] Pouget ER, West BS, Tempalski B, Cooper HLF, Hall HI, Hu X, et al. Persistent racial/ethnic disparities in AIDS diagnosis rates among people who inject drugs in US metropolitan areas, 1993-2007. Public Health Rep 2014;129(3):267e79. [8] Kaplan JE, Hanson D, Dworkin MS, Frederick T, Bertolli J, Lindegren ML, et al. Epidemiology of human immunodeficiency virus-associated opportunistic infections in the United States in the era of highly active antiretroviral therapy. Clin Infect Dis 2000;30(Suppl. 1):S5e14. [9] Walensky RP, Paltiel AD, Losina E, Mercincavage LM, Schackman BR, Sax PE, et al. The survival benefits of AIDS treatment in the United States. J Infect Dis 2006;194(1):11e9. [10] Office of Management and Budget. Standards for defining metropolitan and micropolitan statistical areas. Fed Regist 2000;65:8228e82238. [11] Friedman SR, Tempalski B, Cooper H, Perlis T, Keem M, Friedman R, et al. Estimating numbers of injecting drug users in metropolitan areas for structural analyses of community vulnerability and for assessing relative degrees of service provision for injecting drug users. J Urban Health 2004;81(3):377e400. [12] Holmberg SD. The estimated prevalence and incidence of HIV in 96 large US metropolitan areas. Am J Public Health 1996;86(5):642e54. [13] Brady JE, Friedman SR, Cooper HL, Flom PL, Tempalski B, Gostnell K. Estimating the prevalence of injection drug users in the U.S. and in large U.S. metropolitan areas from 1992 to 2002. J Urban Health 2008;85(3):323e51. [14] Martin J, Hamilton B, Sutton P, Ventura S, Menacker F, Kirmeyer S, et al. Births: final data for 2006. Natl Vital Stat Rep 2009;57(7):1e104.

B.S. West et al. / Annals of Epidemiology 25 (2015) 218e225 [15] Kleinman J, Kiely J. Infant mortality. Health People 2000 Stat Notes 1991;1(2):1e10. [16] Hoover K, Bohm M, Keppel K. Measuring disparities in the incidence of sexually transmitted diseases. Sex Transm Dis 2008;35(12 Suppl):S40e4. [17] IBM Corp. IBM SPSS statistics for Windows. Version 20.0 ed. Armonk, NY: IBM Corp.; 2010. Released. [18] Quinn TC, Overbaugh J. HIV/AIDS in women: an expanding epidemic. Science 2005;308(5728):1582e3. [19] Grigoryan A, Hall HI, Durant T, Wei X. Late HIV diagnosis and determinants of progression to AIDS or death after HIV diagnosis among injection drug users, 33 US States, 1996-2004. PLoS One 2009;4(2):e4445. [20] Barash ET, Hanson DL, Buskin SE, Teshale E. HIV-infected injection drug users: health care utilization and morbidity. J Health Care Poor Underserved 2007;18(3):675e86. [21] Cescon A, Cooper C, Chan K, Palmer A, Klein M, Machouf N, et al. Factors associated with virological suppression among HIV-positive individuals on highly active antiretroviral therapy in a multi-site Canadian cohort. HIV Med 2011;12(6):352e60. [22] Wood E, Kerr T, Tyndall MW, Montaner JS. A review of barriers and facilitators of HIV treatment among injection drug users. AIDS 2008;22:1247e56. [23] Zierler S, Krieger N. Reframing women’s risk: social inequalities and HIV infection. Annu Rev Public Health 1997;18(1):401e36. [24] Eisenman D, Bogart LM, Bird CE, Collins RL, Golinelli D, Fremont A, et al. Differential diffusion of HIV technologies by gender: the case of highly active antiretroviral therapy. AIDS patient care and STDS 2007;21(6): 390e9. [25] Gebo KA, Fleishman JA, Conviser R, Reilly ED, Korthuis PT, Moore RD, et al. Racial and gender disparities in receipt of highly active antiretroviral therapy

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

225

persist in a multistate sample of HIV patients in 2001. J Acquir Immune Defic Syndr 2005;38(1):96e103. Giordano TP, White Jr CA, Sajja P, Graviss EA, Arduino RC, Adu-Oppong A, et al. Factors associated with the use of highly active antiretroviral therapy in patients newly entering care in an urban clinic. J Acquir Immune Defic Syndr 2003;32(4):399e405. Hirschhorn LR, McInnes K, Landon BE, Wilson IB, Ding L, Marsden PV, et al. Gender differences in quality of HIV care in Ryan White CARE Act-funded clinics. Womens Health Issues 2006;16(3):104e12. King WD, Minor P, Ramirez Kitchen C, Ore LE, Shoptaw S, Victorianne GD, et al. Racial, gender and geographic disparities of antiretroviral treatment among US Medicaid enrolees in 1998. J Epidemiol Community Health 2008;62(9):798e803. Reif S, Whetten K, Thielman N. Association of race and gender with use of antiretroviral therapy among HIV-infected individuals in the Southeastern United States. South Med J 2007;100(8):775e81. Lemly DC, Shepherd BE, Hulgan T, Rebeiro P, Stinnette S, Blackwell RB, et al. Race and sex differences in antiretroviral therapy use and mortality among HIV-infected persons in care. J Infect Dis 2009;199(7):991e8. Poundstone KE, Chaisson RE, Moore RD. Differences in HIV disease progression by injection drug use and by sex in the era of highly active antiretroviral therapy. AIDS 2001;15(9):1115e23. Wood E, Kerr T, Zhang R, Guillemi S, Palepu A, Hogg R, et al. Poor adherence to HIV monitoring and treatment guidelines for HIV-infected injection drug users. HIV Med 2008;9(7):503e7. Center for Disease Control and Prevention. Current trends update: impact of the expanded AIDS surveillance case definition for adolescents and adults on case reportingdUnited States, 1993. MMWR 1994;43(09):160e1. 167e170.

ethnic disparities in AIDS diagnosis rates among people who inject drugs in U.S. metropolitan areas, 1993-2007.

AIDS-associated mortality among people living with HIV who inject drugs.

Australia's response to HIV among people who inject drugs.

Undiagnosed HIV among people who inject drugs in Manipur, India.

Gender differences in planning ability and hepatitis C virus among people who inject drugs.

Sex work and HIV incidence among people who inject drugs.

How "hidden" are unobserved networks among people who inject drugs?

Increasing Infectious Endocarditis Admissions Among Young People Who Inject Drugs.

Male-to-male sex among men who inject drugs in Delhi, India: overlapping HIV risk behaviours.

Gender Differences in Factors Related to HIV Risk Behaviors among People Who Inject Drugs in North-East India.

Community reentry challenges after release from prison among people who inject drugs in St. Petersburg, Russia.

High-risk behaviors after release from incarceration among people who inject drugs in St. Petersburg, Russia.

AIDS in Puerto Rican People Who Inject Drugs: Policy Considerations.

Predictors of Injection Cessation and Relapse among Female Sex Workers who Inject Drugs in Two Mexican-US Border Cities.

Going into the groin: Injection into the femoral vein among people who inject drugs in three urban areas of England.

Differences in HIV risk behaviors among people who inject drugs by gender and sexual orientation, San Francisco, 2012.

Estimating the number of people who inject drugs in Australia.

Response among People Who Inject Drugs.

Managing physical health problems in people who inject drugs.

No significant HTLV seroprevalence in German people who inject drugs.

Geographic Differences in Temporal Incidence Trends of Hepatitis C Virus Infection Among People Who Inject Drugs: The InC3 Collaboration.

Depression Among People Who Inject Drugs and Their Intimate Partners in Kazakhstan.

Prevalence and correlates of neck injection among people who inject drugs in Tijuana, Mexico.

HCV in people who inject drugs: a neglected epidemic.