Migration and HIV Risk Behaviors: Puerto Rican Drug Injectors in New York City and Puerto Rico

Abstract

Objectives. We compared injection-related HIV risk behaviors of Puerto Rican current injection drug users (IDUs) living in New York City and in Puerto Rico who also had injected in the other location with those who had not.

Methods. We recruited Puerto Rican IDUs in New York City (n = 561) and in Puerto Rico (n = 312). Of the former, 39% were “newcomers,” having previously injected in Puerto Rico; of the latter, 14% were “returnees,” having previously injected in New York. We compared risk behaviors within each sample between those with and without experience injecting in the other location.

Results. Newcomers reported higher levels of risk behaviors than other New York IDUs. Newcomer status (adjusted odds ratio [OR] = 1.62) and homelessness (adjusted OR = 2.52) were significant predictors of “shooting gallery” use; newcomer status also predicted paraphernalia sharing (adjusted OR = 1.67). Returnee status was not related to these variables.

Conclusions. Intervention services are needed that target mobile populations who are coming from an environment of high-risk behavior to one of low-risk behavior.

The relationship between migration and HIV-related risk behaviors has been reported among diverse populations, including Hispanic populations in the Caribbean and Mexico who travel to New York1–3 and migrant workers in southern Africa.4 Studies have found that travel to an AIDS epicenter is related to HIV infection among drug users from low-prevalence communities.5,6

High rates of HIV/AIDS and HIV-related risk behaviors have been found among Puerto Rican injection drug users (IDUs).7–9 An “air bridge” has been identified between New York and Puerto Rico that represents high reported rates of travel and migration between Puerto Rico and New York,2 and higher levels of risk behaviors have been reported for Puerto Rican IDUs who reside on the island of Puerto Rico compared with those who reside in New York City.10–12 In addition, services that have been found to reduce risk behaviors (e.g., needle exchange programs [NEPs], methadone treatment) are more readily available in New York City.12,13 The extensive migration/travel reported between the 2 communities2 provides an opportunity to compare differences in levels of risk behaviors between IDUs at each location who have experienced injecting in the other location and thus to examine risk behaviors of those who have been socialized in injection behaviors in a community with higher- or lower-risk behaviors than their current community. This information can be helpful in developing and targeting appropriate intervention programs for IDUs who are immigrants.

This study was based on data from a dual-site study of Puerto Rican IDUs recruited in New York City and in Puerto Rico.11 We compared injection-related risk behaviors for IDUs in New York between those who had previously injected in Puerto Rico and those who had not and for IDUs in Puerto Rico between those who had previously injected in New York and those who had not.

METHODS

The Alliance for Research in El Barrio and Bayamón project was funded to assess differences in HIV risk behavior determinants between Puerto Rican IDUs and crack smokers living in New York and those living in Puerto Rico. We conducted recruitment in East Harlem, New York, and Bayamón, Puerto Rico (part of the greater San Juan metropolitan area). We established teams of researchers, including qualitative and survey researchers in each site and used multiple methods to ensure comparability of procedures for data collection, including jointly developing instruments and conducting pilot tests of the interviews in both locations.11,14 The interviews were developed in English and translated into Spanish. We established the fidelity of translation and comparability of terms through back-translation, review of terms by both research teams, and pilot tests.

Recruitment

The East Harlem community is an urban community of approximately 2.5 square miles, and the municipality of Bayamón is 45 square miles of mixed urban and rural environments. From detailed ethnographic maps of these 2 communities, we identified recruitment sectors that contained drugselling and drug-using locations as well as health and other social welfare services.15 Five contiguous sectors were identified in East Harlem, and 3 noncontiguous sectors were identified in Bayamón, with total areas that were similar in size. We randomly selected sectors for daily recruitment within each site.

We recruited subjects between January 1998 and July 1999. Eligible subjects were aged 18 years or older, identified themselves as Puerto Rican, and had injected drugs (usually heroin, cocaine, or speedball [a combination of heroin and cocaine]) or smoked crack within the last 30 days. After recruitment, subjects were brought to the field site in each location, where recent heroin and/or cocaine use was verified through urinalysis with the OnTrak system (Roche Diagnostics Corp, Indianapolis, Ind) and an informed consent was obtained. This article concerns only those participants who reported injecting drugs within the past 30 days.

Interviewing

We conducted baseline interviews and provided subjects with HIV counseling and testing. Interviews included sections on demographic variables (including birthplace), as well as detailed information on risk behaviors, including those engaged in during the 30 days before the interview. We also asked subjects whether they had lived in the other location (i.e., New York or Puerto Rico) since they had begun using drugs on a regular basis and whether they had injected drugs in the other location.

Variables Examined

To identify influences of changes in location, we developed a definition of “newcomers” to New York and “returnees” to Puerto Rico. Newcomers were those individuals recruited in New York who had previously lived in Puerto Rico for at least 1 year since they had begun using drugs on a regular basis and who had injected in Puerto Rico (this group accounted for 39% of the New York sample; n = 221). Returnees were those individuals recruited in Puerto Rico who had previously lived in New York for at least 1 year since they had begun using drugs on a regular basis and who had injected in New York (this group accounted for 14% of the Puerto Rico sample; n = 44). We also examined other sociodemographic predictors of risk, including sex, age, and homelessness (defined as living in a shelter or on the streets/in abandoned buildings).

Four injection-related variables were selected as dependent measures: (1) frequency of injection (which has been related to HIV seroincidence16); (2) injecting in “shooting galleries,” settings where others are injecting and where resources for injection are provided (which has been associated with increased needle and paraphernalia sharing17); (3) sharing of syringes; and (4) sharing of other injection-related paraphernalia that also may transmit HIV, including cookers, cottons, and rinse water.18

Analysis

We used multiple logistic regression analysis for the dichotomous dependent variables of shooting-gallery use and sharing of injection equipment (syringes and other paraphernalia) and multiple regression analysis for the continuous dependent variable of injection frequency. Because the distribution of injection frequency was skewed, we used logarithmic transformation. All regression models included the same independent variables of sex, age, homelessness, and newcomer/returnee status.

RESULTS

We recruited a total of 873 IDUs: 561 in New York City and 312 in Puerto Rico. Demographic and migration characteristics indicated that the majority of IDUs in both sites were male and had been born in Puerto Rico (56% of those recruited in New York and 87% of those recruited in Puerto Rico) (Table 1 ▶). Whereas 40% of the New York–recruited IDUs had been born in New York, 10% of the Puerto Rico–recruited IDUs had been born in New York (P < .001). We applied the term returnees to the IDUs recruited in Puerto Rico who met the criterion of having previously injected in New York, because the majority of them—approximately 70%—reported spending most of their childhood in Puerto Rico. In contrast, only 19% of the New York newcomers reported spending their childhood in New York. The New York IDUs were older—for example, 28% were older than age 42, compared with 10% of the sample in Puerto Rico (mean ages were 39 and 33, respectively; P < .001).

TABLE 1.

—Sociodemographic and Injection-Related Characteristics of Puerto Rican Injection Drug Users Recruited in East Harlem, New York, and Bayamón, Puerto Rico

	Recruitment Site
	New York (n = 561)	Puerto Rico (n = 312)
Sociodemographic characteristics
Sex, % male*	79	86
Age in years, %**
< 31	16	43
31–37	31	33
38–42	25	14
> 42	28	10
Homeless, %	23	22
Birthplace, %**
New York	40	10
Puerto Rico	56	87
Other	4	3
Injection-related characteristics
Mean number of years injected**	18 (SD = 10.3)	13 (SD = 9.5)
Shared syringes, %a**	10	37
Shared other paraphernalia, %a**	33	71
Drugs injected, %a
Heroin**	86	67
Cocaine**	46	67
Speedball**	52	90

Note. SD = standard deviation.

^aBased on behaviors during the 30 days prior to the interview.

*P < .01; **P < .001. All P values are 2-tailed.

The New York sample had been injecting for a longer period (18 years vs 13 years in Puerto Rico; P < .001), were less likely to report sharing of injection equipment (for example, 10% shared syringes vs 37% in Puerto Rico; P < .001), and were less likely to have injected speedball during the past 30 days (52% vs 90% in Puerto Rico; P < .001).

There were significant sociodemographic differences between the newcomers and other New York recruits and between the returnees and other Puerto Rico recruits (Table 2 ▶). The newcomers to New York were more likely than the other New York recruits to be male, and the returnees to Puerto Rico were older than the other Puerto Rico recruits. The newcomers to New York were more likely than other New York recruits to be homeless, and most of those who were homeless resided in shelters. (In Puerto Rico, all homeless participants lived on the streets because no shelters existed in Bayamón at the time of recruitment.)

TABLE 2.

—Sociodemographic and Injection-Related Characteristics of Newcomers and Returnees^a

	Recruitment Site
	New York		Puerto Rico
	Newcomers (n = 221)	Others (n = 340)	Returnees (n = 44)	Others (n = 268)
Sociodemographic characteristics
Sex, % male	91	71***	84	86
Age in years, %
< 31	19	14	14	48***
31–37	31	31	41	31
38–42	22	27	18	13
> 42	28	28	27	8
Homeless, %
Total	32	18***	14	24
Live on street	10	7	14	24
Live in shelter	23	11***	…	…
Injection-related characteristics
Mean number of years injected	20	16***	19	11***
Mean frequency of injections per monthb	87	69*	158	187
Used shooting galleries, %b	32	18***	68	81
Shared syringes, %b	10	10	39	37
Shared other paraphernalia, %b	41	28***	68	72

^aNewcomers were those individuals recruited in New York who had previously lived in Puerto Rico for at least 1 year since they started using drugs on a regular basis and who had injected in Puerto Rico. Returnees were those individuals recruited in Puerto Rico who had previously lived in New York for at least 1 year since they started using drugs on a regular basis and who had injected in New York.

^bBased on behaviors during the 30 days prior to interview.

*P < .05; **P < .01; ***P ≤ .001. All P values are 2-tailed.

HIV Risk Behaviors

The newcomers to New York reported more years of injecting and more frequent injecting than other New York IDUs (Table 2 ▶). In addition, they reported riskier injection practices than other New York IDUs. For example, use of shooting galleries was reported by 32% of newcomers compared with 18% of other New York IDUs (P < .001), and sharing of injection paraphernalia was reported by 41% of newcomers versus 28% of other New York IDUs (P ≤ .001). There were no statistically significant differences (P < .05) in these behaviors between the returnees to Puerto Rico and the other IDUs recruited in Puerto Rico. The significant differences in risk behaviors between Puerto Rican IDUs recruited in New York and those recruited in Puerto Rico have been reported elsewhere.10,11

We conducted multivariate analyses for the 4 risk behaviors, controlling for 3 sociodemographic variables (sex, age, and homelessness), to assess the influence of being a newcomer in New York and a returnee in Puerto Rico (Table 3 ▶). After control for the influence of sociodemographic factors on injection frequency, newcomer/returnee status was not a significant predictor of injection frequency in either site. In New York, being a newcomer was significantly related to shooting-gallery use (adjusted odds ratio [OR] = 1.62; 95% confidence interval [CI] = 1.06, 2.48) and sharing of injection paraphernalia other than syringes (adjusted OR = 1.67; 95% CI = 1.15, 2.43). In Puerto Rico, returnee status was not significantly related to these 2 variables. In New York, homeless newcomers reported the highest rates of shooting-gallery use (45%), and the lowest use rate was among those who were not homeless and not newcomers (14%).

TABLE 3.

—Multivariate Analysis for 4 Injection Risk Behaviors

	Recruitment Site
	New York (n = 561)	Puerto Rico (n = 312)
Injection frequencya (standardized coefficient)
Sexb	– 0.07	0.03
Age, yc
< 31	0.08	0.31**
31–37	0.07	0.21*
38–42	0.03	0.15
Homelessd	0.06	0.26***
Newcomer/returneee	0.06	0.03
Used shooting galleries (adjusted OR [95% CI])
Sexb	0.42** (0.21, 0.83)	0.72 (0.33, 1.55)
Age, yc
< 31	1.52 (0.81, 2.86)	1.89 (0.73, 4.89)
31–37	1.21 (0.70, 2.08)	0.94 (0.38, 2.34)
38–42	0.95 (0.52, 1.73)	1.39 (0.47, 4.11)
Homelessd	2.52*** (1.61, 3.94)	2.01 (0.92, 4.41)
Newcomer/returneee	1.62* (1.06, 2.48)	0.64 (0.31, 1.35)
Shared syringes (adjusted OR [95% CI])
Sexb	0.99 (0.48, 2.06)	0.60 (0.29, 1.24)
Age, yc
< 31	0.94 (0.42, 2.09)	1.00 (0.44, 2.28)
31–37	0.73 (0.36, 1.46)	0.98 (0.43, 2.24)
38–42	0.61 (0.28, 1.33)	0.51 (0.19, 1.39)
Homelessd	1.74 (0.94, 3.22)	0.92 (0.52, 1.66)
Newcomer/returneee	0.91 (0.51, 1.63)	1.11 (0.55, 2.23)
Shared other paraphernalia (adjusted OR [95% CI])
Sexb	0.90 (0.56, 1.45)	0.67 (0.34, 1.33)
Age, yc
< 31	1.08 (0.62, 1.87)	1.09 (0.45, 2.68)
31–37	0.79 (0.49, 1.26)	0.98 (0.40, 2.41)
38–42	1.08 (0.66, 1.77)	0.66 (0.24, 1.79)
Homelessd	1.48 (0.97, 2.25)	1.08 (0.58, 2.02)
Newcomer/returneee	1.67** (1.15, 2.43)	0.90 (0.44, 1.86)

Note. OR = odds ratio; CI = confidence interval.

^aLogarithmic transformations were used owing to a skewed distribution.

^b1 = male, 2 = female.

^cReference group is persons older than age 42 years.

^d0 = no, 1 = yes.

^e0 = other, 1 = newcomer/returnee.

*P < .05; **P < .01; ***P < .001. All P values are 2-tailed.

DISCUSSION

The newcomer IDUs who had injected in Puerto Rico and moved to New York City—a location with lower levels of injection-related risk behaviors and more tools available for risk reduction—evidenced higher levels of risk behaviors than other Puerto Rican IDUs in New York. However, these newcomers’ levels of risk behaviors were still lower than those found in the Puerto Rico sample. Shooting-gallery use in New York was influenced by newcomer status as well as homelessness, with almost half of all IDUs who were homeless newcomers reporting shooting-gallery use. The returnee IDUs who had injected in New York and were now injecting in Puerto Rico reported levels of risk behaviors that were not significantly different from the high levels shown by other IDUs recruited in Puerto Rico.

Other research suggests that environmental structural differences, including the greater availability of NEPs and drug treatment programs in New York, may account for some of the differences in level of risk behaviors in the 2 communities.14,19 In addition, focus groups conducted for the current project with newcomers in New York indicated that many of the newly arrived IDUs reported coming to New York because of the greater ease of getting into methadone maintenance treatment programs as well as access to welfare, Medicaid, and other services. Nonetheless, in-depth interviews indicated that some new arrivals may have difficulties in accessing these services (D. Oliver-Velez, personal communication, December 2000). Focus groups with returnees in Puerto Rico and newcomers in New York confirmed these environmental differences as well as differences in drug treatment availability (with participants expressing concerns about the lack of treatment access in Puerto Rico) and the greater availability of shooting galleries in Puerto Rico.20 Other qualitative research in New York involving observations in the community provided reports of homeless newcomers engaging in high-risk injection behaviors (D. Oliver-Velez, personal communication, December 2000).

These survey and qualitative results point to the need for services specifically targeting new arrivals from Puerto Rico that address language issues, homelessness, and the need for a variety of services, including drug treatment. Research is needed into the attempts of newcomers to access services and the barriers to utilizing services. In addition, our finding that new arrivals to New York may be practicing riskier behaviors than other New York IDUs should be incorporated into prevention/intervention messages to help these new arrivals in accessing risk-reduction tools (such as new needles from NEPs and pharmacies) and to encourage their injection partners in New York to assist them with risk reduction. Providers of drug treatment, NEPs, and related services can be alerted to help assess risk behaviors of individuals who may be newcomers so that appropriate risk-reduction messages and referrals can be provided.

The finding that returnees to Puerto Rico appear to take on the risk behaviors local to Puerto Rico also can be incorporated into prevention efforts in both locations, although this may be difficult in an environment of reduced resources and higher risks. It may be possible, nonetheless, to train IDUs who travel to Puerto Rico to help influence IDUs in the new location to reduce their risks within the environmental constraints. For example, if new needles are not available, these trained “change agents” can encourage fellow IDUs to follow recommended bleach-cleaning practices before sharing needles. Research on how these emissary efforts can be successfully undertaken should be conducted, and the impact on the emissaries themselves as well as other IDUs should be assessed.

There are several limitations of these data. As with all studies of “hidden populations” such as drug users, random sampling of the population is not possible. Nonetheless, targeted sampling based on ethnographic mapping within each community helped ensure broad-based recruitment of the population. The possibility for multiple back-and-forth visits, referred to as “circular migration,” was not taken into account, although the criteria of having injected in the other location and having lived there for at least 1 year since initiating regular drug use helped ensure that participants identified as newcomers or returnees had substantial experience in injecting in the other location. Additionally, some of those participants identified as newcomers actually may have lived in New York for many years and thus may not be newcomers in terms of being recent arrivals. Finally, the relatively small sample size of returnees to Puerto Rico may limit the power to detect significant differences between them and other IDUs there.

HIV infection among injection drug users had been reported in 114 countries by mid-1999,21 and increases in emigration from many countries to the United States have been reported, based on the 2000 US Census.22 In addition, we found that immigrants with limited resources who may be homeless or only temporarily domiciled appear to be particularly at risk of engaging in high-risk behaviors. These trends point to the need to implement effective prevention efforts to reduce HIV transmission in mobile populations.

Our findings have implications for HIV prevention efforts geared to mobile IDU populations. Efforts targeting these populations should be based on knowledge of the behaviors and risk-reduction resources in the communities from which individuals are emigrating as well as in the new host communities. Efforts to encourage the use of available risk-reduction services and to make these services more accessible to the migrant population (e.g., by offering services in appropriate languages and developing culturally appropriate educational materials) are needed to maximize their utility. In addition, incorporating information about the new risks or potential risk-increasing factors that may affect immigrants (e.g., for IDUs who may be immigrating to areas where shooting galleries are widely available) can be helpful in reducing HIVrelated risks and HIV transmission.