Previous studies have shown a link between migration of Mexicans to the United States and an increased risk of HIV infection.1–11 Mexican surveillance data from the early years of the HIV epidemic have linked the development of the epidemic in Mexico with migration to the United States.1 In 1983, all of the cases registered in Mexico were people who had lived in the United States.1 Subsequent surveillance data have suggested that migration is still likely to be associated with AIDS incidence, with 33% of AIDS cases in Mexico coming from the states with the highest number of migrant workers to the United States.2 These surveillance data have provided potential evidence of migrants becoming infected in the United States and subsequently returning to their local community in Mexico.3 Furthermore, results from a Los Angeles study suggested that the HIV-positive, Mexican-born study participants were more likely to have been infected after migration.4 With regard to HIV high-risk behaviors, a study in Mexico showed that migrants to the United States reported more HIV high-risk behavior than non-migrants.1 Ethnographic studies have described an association between Mexican migration to the United States and the adoption of HIV high-risk behaviors as well as an increased frequency of these behaviors among male Mexican migrants.5,6 Bronfman et al described the change of environment after migration to the United States, the new or increased level of exposure to high-risk behaviors, and the subsequent adoption of these practices.5 In addition, other studies of migrant farmworkers and high-risk migrant populations in the United States have shown that Mexican migrants engage in HIV high-risk behaviors; however, the effect of migration on these high-risk behaviors was not quantified.7–11
This study reports results from the first case-crossover study conducted at multiple migrant communities in the United States to directly examine the effect of migration on HIV high-risk behaviors among Mexican migrants.7 While previous studies have described this association, we utilized a case-crossover study design to actually quantify the effect.12 We directly estimated the effect of migration on HIV high-risk behaviors among Mexican migrants by comparing their individual HIV high-risk behaviors before and after migration.
Study Design and Subject Recruitment
To assess the effect of migration on HIV high-risk behaviors and test for HIV and sexually transmitted infections (STIs) among Mexican migrants, we implemented the California-Mexico Epidemiological Surveillance Pilot (CMESP) from July through November 2005 in both rural and urban areas of Fresno County and San Diego County in California. CMESP, a binational collaboration, combined targeted, venue-based sampling, household-dwelling sampling, and survey methods to study this hard-to-reach population.13 We identified sites frequented by Mexican migrants for enumeration and sampling based on key informant interviews and focus groups with members of the Mexican migrant community in both counties. CMESP study sites consisted of 3 types of venues where there was a high concentration of Mexican migrants. The male work venues included male migrant camps, shelters, or shantytowns for men and job pick-up sites. The community venues consisted of those sites where male and female migrants worked and lived or congregated regularly. Finally, the high-risk behavior venues consisted of bars and clubs, including those frequented by men who have sex with men (MSM), and parks and streets known for drug use or MSM activity. The sampling frame was dynamic. As the attendance patterns of our target population changed throughout the study, sites were added or deleted from the sampling frame in accordance with attendance criteria. Study participants were systematically sampled and recruited at each site in relative proportion to the volume of eligible migrants enumerated at that site. For the high-risk behavior venues, which had a smaller concentration of eligible migrants, the proportion sampled and recruited was increased to insure a sufficient sample size for this high-risk subpopulation. Sampling was conducted during peak attendance dates and times for our target population and when enrollment success was considered likely to be optimal.
Our Mexican migrant study participants were defined as those reporting being between the ages of 18 and 64 years, who were born in Mexico, and (1) had been living or working in the United States for 5 years or less or (2) had been in the United States for more than 5 years but had returned to Mexico at least every 2 years on average. In addition, those who had been in the United States for 30 days or less or did not speak either English or Spanish were excluded from the study.
Case-Crossover Study and Statistical Analysis
We conducted a case-crossover study analysis with eligible participants sampled in 2005 to directly compare individual HIV high-risk behaviors before and after migration. This study estimates the net changes in the odds of engaging in HIV high-risk behaviors after migration and shows how these changes vary for subgroups. By collecting data on HIV high-risk behaviors before and after migration, we created matched-pair data for each subject. The strength of this case-crossover design is that each person serves as his or her own control, with reported risk behaviors before migration serving as the control time. This approach controls for confounding due to those individual characteristics that remain relatively constant over time in adults, such as gender, educational level, and the proclivity to engage in HIV high-risk behaviors because those characteristics exert a similar influence during both periods measured. For those variables changing over time (“age” and “length of time living or working in the United States”), we controlled for confounding by including them in our model.14 We directly compared the odds of individuals engaging in HIV high-risk behaviors before and after migration using exact conditional logistic regression, which is valid for complex sampling designs such as our targeted, venue-based sampling.15,16 Using SAS 9.1, we estimated odds ratios with 95% confidence intervals, controlling for venue type and gender. To adjust for the discreteness of the exact conditional logistic distribution, mid-P values were used.17
Measures and Testing
A standardized questionnaire was administered in Spanish to collect information about demographic and HIV high-risk behavioral characteristics, migration patterns, access to health services, sexual behavior, and alcohol and drug use. Questions regarding HIV high-risk behaviors before and after migration were asked in similar formats. “Before migration” behaviors consist of the study participant's reported activities anytime before first coming to the United States to live or work. Similarly, “after migration” behaviors consist of the participant's reported activities anytime after coming to the United States to live or work. We selected behavior variables that had been previously cited in the literature concerning Mexican migrants and that were potentially measurable in a pre- and post-migration question format. These behaviors include 2 consistently interrelated HIV high-risk behaviors among male Mexican migrants, sex with sex workers and sex under the influence of alcohol.3,18,19 Additionally, the frequency of condom use among male Mexican migrants has been reported both as relatively high and low, depending on the study and the type of partner.20,21 Research in Mexico indicates that whether condom use is relatively high or low in the United States, Mexican migrants report more condom use than Mexican non-migrants.2,22 Another high-risk behavior that we selected, which has been tacitly associated with survival practices among some Mexican migrants, is performing sex work in exchange for money, food, shelter, protection, drugs, or anything else.23 Our final selected behavior, MSM, continues to be the predominant mode of HIV exposure among Latinos in California and has been associated with considerable HIV infection among Latinos at the Mexican border.11
In addition, blood and urine specimens were collected for testing for HIV, syphilis, and Chlamydia trachomatis. Gen-Probe Aptima Combo 2 assay was used to detect C. trachomatis in urine samples. Syphilis was detected in serum samples using Rapid Plasma Reagin and confirmed by Treponema pallidum particle agglutination. HIV was detected on serum samples by enzyme immunoassay and confirmed by Western blot. A follow-up disclosure session of test results included counseling and referrals to treatment and other services.
A total of 458 study participants were enrolled in California: 277 (60%) from San Diego County and 181 (40%) from Fresno County. Of the total, 364 (79%) were male and 94 (21%) were female. The study participants were recruited from 36 sampling sites: 19 in San Diego County and 17 in Fresno County. A total of 202 (44%) study participants were sampled from male work venues, which included 11 male migrant camps, shelters, or shantytowns for men and 5 job pick-up sites; 195 (43%) study participants were sampled from community venues, including 4 family migrant camps, an apartment complex, 2 laundromats, 3 parks, 2 churches, a food bank, and a legal assistance site; and 61 (13%) study participants were sampled from high-risk behavior venues, which included 2 MSM and 3 non-MSM bars or night clubs and 1 park with drug activity. Of the 617 eligible migrants offered enrollment across the sites, 74.2% participated in the survey. Of those who refused to participate in the survey, 79% provided basic demographic data during initial contact, before refusing participation in the study. This large subset of the refusal group was compared with the 458 survey participants, and the 2 groups did not differ significantly based on the basic demographic data collected.
Table 1 presents characteristics of the 458 study participants by gender. These descriptive results in Table 1 are unweighted and unstratified by sampled venue type. Hence, these descriptive results are not representative of our overall target population at these sampled venues, but rather, they summarize the study subpopulation used in this analysis. The median age at migration for the 364 men was 20 years, whereas for the 94 women, it was 22 years. For the men, 51.1% were currently married, with 80.2% of these spouses living in Mexico or elsewhere outside of California. In contrast, 68.1% of the women were married, with 26.6% of these spouses living outside of California. Of the men, 39.7% lived with family, whereas 92.0% of the women lived with family. The percentage of men and women returning to Mexico in the last 12 months was 69.3% and 78.3%, respectively. One man enrolled at an MSM bar (0.28% of the 364 men) tested positive for HIV. In addition, 4 men (1.10%) tested positive for early latent syphilis, of whom 3 were recruited from male work venues and 1 from an MSM bar. None of the women tested positive for HIV or syphilis. Twelve (3.33%) of the men tested positive for C. trachomatis, with only one case at a community venue.
Study participants who reported any sexual partners in the past 12 months (85.7% of the men and 88.3% of the women) were asked for the total number of partners, and more detailed information (eg, partner's gender and race/ethnicity, type of partner, and condom use at last sexual encounter) on each of these partners, up to the 6 most recent partners. The mean number of sexual partners for this period was 2.6 for the men and 1.2 for the women. Of the men, 5.8% reported one or more male sexual partners during this period, with an average of 2.1 partners. Most study participants reported having only Latino sexual partners in the past 12 months (90.1% of the men and 97.6% of the women). For the men and women, 62.7% and 65.5%, respectively, reported that they did not have access to condoms when needed, and 64.4% and 80.5%, respectively, reported having at least one vaginal or anal sexual encounter without a condom in the past 12 months. Study participants also reported whether a sexual partner had been a main, casual, one-time, or sex work partner. The percent of condom use at the most recent vaginal or anal sexual encounter with a main partner was 28.8% among the male study participants and 17.9% among the female study participants. Additionally, the percent of condom use at the most recent vaginal or anal sexual encounter with casual, one-time, and sex work partners was 58.2%, 67.2%, and 91.4%, respectively, among the male study participants. Finally, for the men, 52.8% reported having had more than 5 alcoholic drinks in one day in the last 30 days, whereas 19.0% reported methamphetamine or cocaine use in the last 12 months.
Results in Table 2 show a significant increase in the odds of the male Mexican migrants having a sexual partner who was a sex worker after migration [odds ratio (OR) = 2.64, P < 0.0001]. There were also significant increases in the odds of the male Mexican migrants having sexual relations with a partner while under the influence of drugs or alcohol after migration (OR = 5.00, P < 0.0001) and in the odds of the males reporting sex with a male sexual partner after migration (OR = 13.00, P = 0.001). Additionally, there was an increase after migration, although not significant, in the odds of the males exchanging sex for money, food, shelter, protection, drugs, or anything else (OR = 6.00, P = 0.070). In contrast, there was a significant decrease after migration in the odds of the males reporting low condom use (never, rarely, or sometimes use a condom) (OR = 0.21, P < 0.0001). Among the female migrants, although none of the changes after migration were found to be significant, the direction of the change after migration, either an increase or decrease, for each behavior was comparable to the results for the men.
Table 3 shows the changes after migration in these HIV high-risk behaviors among the 364 male Mexican migrants, stratified by the venues where they were recruited: male work venues, community venues, and high-risk behavior venues. The largest increase in the odds of engaging in sex with a sex worker after migration was among the men in the male work venues (OR = 2.94, P < 0.0001). With regard to the men engaging in sex under the influence of drugs or alcohol, the increase in the odds after migration was significant for all 3 venue types, with the most significant increase occurring among the men sampled in the high-risk behavior venues (OR = 25.47, P < 0.0001). There was a significant increase in the odds of engaging in sex with male sexual partners after migration among the men surveyed in the high-risk behavior venues (OR = 18.26, P = 0.0001), whereas this was not the case for men surveyed in the other venues. With regard to reporting low condom use, the most significant decrease in the odds after migration occurred among the men sampled in the male work venues (OR = 0.19, P = 0.0001).
To test for interaction, “age,” “age at migration,” “length of time living or working in the United States,” “housing,” “years of education,” and “living alone, with family, friends, or acquaintances” were all investigated. Both age and length of time living or working in the United States were determined to be effect modifiers for all 4 of the HIV high-risk behaviors that increased for the men after migration. Table 4 highlights the significant results from our testing for interaction. The most significant increase in the odds of engaging in sex with a sex worker after migration occurred among the men who had been in the United States for more than 5 years (OR = 6.75, P < 0.0001). The most significant increase in the odds of having sexual relations while under the influence of drugs or alcohol after migration occurred among the men between the ages of 18–29 years (OR = 7.83, P < 0.0001). Finally, the most significant increase in the odds of engaging in sex with a male partner after migration also occurred among the men between the ages of 18 and 29 years (OR = 12.49, P = 0.002).
Using a case-crossover study design, we estimated the effect of migration on HIV high-risk behaviors that could potentially lead to an increased risk for HIV infection among migrants. Our results indicated notable increases in the odds of male Mexican migrants adopting HIV high-risk behaviors after migration, even when migrants were well into adulthood. These behaviors included engaging in sex with a sex worker, engaging in sex while under the influence of drugs or alcohol, performing sex work, and having sex with a male partner. Furthermore, we were able to estimate particularly pronounced effects within relevant subgroups of the population, including male migrants, men who have been living or working in the United States for more than 5 years and men from the youngest age cohort.
We stratified by 3 venue types given that the housing conditions, job stability, and isolation from spouses and families for those sampled from these venues so distinctly differed by venue type. Poor housing conditions, job instability, and isolation from spouses and families were more common in the male work venues as compared with the community venues. We found that men enrolled at male work venues reported the greatest increase in having sex with a sex worker after migration. In many instances, this increased sexual activity with sex workers occurred at the actual male work site. For those enrolled from male migrant camps who reported sex with a sex worker in the last 12 months, 16 of the men specifically reported that they had met their sex work partner in their respective male migrant camp. Among married male migrants, 80.2% reported living away from their spouse and among all of the male migrants, only 39.7% reported living with family, as compared with 92.0% of the female migrants. Thus, we found that the majority of our male study participants were isolated from their family and exhibited a significantly increased level of sexual activity with sex workers, particularly men recruited from the exclusively male-populated venues. We found that excessive use of alcohol and methamphetamine or cocaine was common in this male population. With regard to sexual activity while under the influence of drugs or alcohol, we saw a significant increase in the odds after migration across all types of venues, but most significantly among the male-dominated venues. In summary, these results indicate that the male migrants found in these male-dominated venues are at the greatest risk for engaging in high-risk behaviors after migration. Thus, these male migrants are vulnerable to HIV and STI transmission through their work and housing environments and bars and clubs, particularly those settings where women are infrequently present.
Both having been in the United States for more than 5 years and being in the youngest age cohort (18–29 years) were determined to be effect modifiers for all 4 of the HIV high-risk behaviors that increased after migration, but age at time of migration was not an effect modifier. Among Mexican immigrants, the number of years in the United States has been correlated with acculturation, which in turn has arguably been associated with negative health behaviors, such as substance abuse.24,25 In our analysis, the measure of effect of migration on HIV high-risk behaviors was greatest for the subgroup of men who have been in the United States for the longest period of time (more than 5 years). This additional time in the United States potentially allowed for more opportunities to be exposed to environments and practices that can foster high-risk behavior. For many in this study population, the migratory experience included extended periods of time separated from family; poor living conditions in migrant camps, shelters, or shantytowns; constant mobility for the agricultural workers; and extended periods of unemployment. However, the association between acculturation and the levels of HIV high-risk behaviors still remains open to question. Both lower and higher levels of acculturation among Latinos have been reported in the literature as protective against or promoting increased HIV high-risk behaviors.26 Perhaps, this apparent inconsistency is due to the use of the term acculturation to refer to an individual migrant's vast array of experiences, both risk-enhancing and risk-reducing, that may not be interrelated, unidirectional, and are aggregated together because they all occurred while in the United States. Nevertheless, it is interesting that acculturation has been associated with length of time in the United States and that for our analysis, greater time in the United States was associated with increased odds of adopting HIV high-risk behaviors. The magnitude of the effect of migration on HIV high-risk behaviors was also greatest for the youngest subgroup of men. Additionally, a recent population-based survey of young men (18–35 years of age) in 5 counties in California found that Latino immigrants reported considerable HIV high-risk behavior.27 Similarly, the youngest age cohort in our study exhibited the greatest change in HIV high-risk behaviors after migration. Therefore, from an intervention perspective, it should be emphasized that this youngest male subgroup is at the greatest risk for adopting these HIV high-risk behaviors after migration.
In this study, we designed a sampling strategy to recruit participants from a broad range of settings, including work-based, community, and social venues, to survey a fair representation of the types of settings where Mexican migrants live, work, or congregate within San Diego and Fresno Counties. Both rural and urban areas within these counties were selected for this study given their high volume and diversity of Mexican migrants. Therefore, the study participants systematically sampled within the venues should serve as a fair representation of Mexican migrants in California as a whole, which is home to approximately 40% of Mexican migrants in the United States.28 Given both the diversity and transient nature of the Mexican migrant population, our targeted, venue-based sampling methodology captured one of the most comprehensive samples of the Mexican migrant population in the United States to date, but the probability of selection for any individual migrant is not known. Therefore, as in all case-crossover studies with unknown selection probabilities, the constant coefficient in the logistic regression model cannot be estimated without bias, so the target population frequencies cannot be estimated. However, the coefficients of the linear terms can be estimated without bias provided the logistic regression model describes the target population well.16 Because there was no evidence in our data analysis to suggest that the logistic regression model did not describe the target population well, this study provides valid estimates of the relative effect of migration on the odds of engaging in HIV high-risk behaviors.
One limitation of our study is that only 74.2% of the eligible Mexican migrants agreed to participate, and, thus, those who refused enrollment may differ with respect to various characteristics or HIV high-risk behaviors. Additionally, our study population may have underreported street drug use because of a fear of the association with illegal activities. In addition, high-risk sexual behaviors, such as same-sex sexual relations, may have been underreported, given the cultural stigma. Finally, a given participant's inability to accurately remember HIV high-risk behaviors engaged in either before or after migration could result in recall bias. However, 4 of the 5 high-risk behaviors that we were measuring were significant activities or events which most people would likely remember over time, including having sex with a sex worker, performing sex work, or having same-sex sexual relations. The fifth high-risk behavior, low condom use, however, is a significantly less memorable event and reporting may be far more subjective in nature and, consequently, potentially less accurate.
Although most participants reported low condom use both before and after migration, we observed a significant decrease in the odds of male migrants reporting low condom use after migration. However, counteractively, over this same time period, we observed increases in the odds of engaging in all 4 of the other high-risk sexual behaviors. Additionally, even if Mexican men are using condoms more frequently after migration, 62.7% of those men reported not having access to condoms when needed and 64.4% reported having at least one vaginal or anal sexual encounter over this time period where they did not use a condom. The increased condom use reported after migration suggests that Mexican migrants are susceptible to public health campaigns or interventions that promote protected safe sex practices. Although prevention efforts targeted at Latino migrants are sparse and not necessarily tailored to the population, there is some evidence that community-based organization (CBO) prevention programs in California may be successfully promoting the use of condoms among this population.26 In one study, CBOs were found to be more effective in reaching Latinos than federal and state agencies because of their greater number of bilingual staff and culturally sensitive approaches to service delivery.29 Thus, the observed decrease in the odds of male migrants reporting low condom use after migration in our study could be associated with effective CBO prevention messages coupled with access to condoms in the United States. Moreover, our sexual partner results suggest that this study population perceives different levels of risk for HIV/STIs by type of partner (main, casual, one-time, and sex work) and may use a condom accordingly. Nevertheless, the fact that most study participants reported low condom use is an indication that prevention programs need to be more prevalent, targeted, and integrated into the health services throughout migrant communities.
Our study provides substantial quantitative support for the hypothesis that the odds of Mexican men engaging in HIV high-risk behaviors increase after migration to the United States. Given our case-crossover study embedded in our targeted, venue-based survey, we have shown that migration serves as a good predictor of future HIV high-risk behavioral changes for the male Mexican migrant population. These results suggest that in the absence of effective interventions, the HIV epidemic may expand among this population in the coming years. Furthermore, given that male migrants are returning to Mexico frequently, there is the additional risk of migrants increasing HIV transmission in Mexico. Overall, the prevalence of HIV high-risk behaviors and apparent low HIV-infection rates in the general Latino migrant population remain a paradox that has been documented elsewhere.7,30 Results from our study show an almost exclusively Latino sexual network among the study participants. It is possible that the Latino sexual and needle-sharing migrant networks may currently lack the threshold number of HIV-infected individuals to propel the HIV epidemic, despite the prevalence of HIV high-risk behaviors. In addition, the condom use reported with higher-risk sexual partners (67.2% with one-time partners and 91.4 % with sex work partners) in our study may reflect the success of prevention messages and further explain this paradox. However, it is also possible that new HIV infections in Latino migrants are not being detected by current HIV surveillance programs. Non–US-born Latinos are currently at the forefront of populations who test late for HIV.31,32 Our analysis additionally identified subgroups of the migrant population in which the effects of migration are particularly pronounced. These subgroups include those male migrants who have been living or working in the United States for more than 5 years, those from the youngest age cohort, and those in male work venues and high-risk behavior venues. These results suggest targeting HIV prevention interventions and health and diagnostic services to these male subgroups and male-dominated venues where this hard-to-reach population is concentrated.
HUMAN PARTICIPANT PROTECTION
The questionnaire, consent forms and procedures, and study design were approved by the Committee on Human Research at the University of California, San Francisco and by the Committee for the Protection of Human Subjects, California Health and Human Services Agency. In addition, all procedures followed were in accordance with the Helsinki Declaration of 1975, as revised in 2000.
The authors would like to thank the individuals who participated and communities that participated in their study. This study is a product of the California-Mexico AIDS Initiative, a collaboration between federal, state, and local agencies in the United States and in Mexico. Primary partners are the California HIV/AIDS Research Program, University of California, Office of the President; the Office of AIDS, California Department of Public Health; and the Centro Nacional para la Prevención y el Control del VIH/SIDA e ITS, Secretaría de Salud, México. Other partner agencies include the Sexually Transmitted Disease Control Branch, California Department of Public Health; the Public Health Laboratory, County of San Diego; Vista Community Clinic; Bi-National AIDS Advocacy Project; Planned Parenthood Mar Monte; the Health and Human Services Agency, County of San Diego; the Department of Community Health, County of Fresno; the Latino Center for Medical Education and Research, University of California, San Francisco at Fresno; and the Health Initiative of the Americas, University of California, Berkeley.
1. Magis-Rodríguez C, Lemp G, Hernandez MT, et al.. Going North: Mexican migrants and their vulnerability to HIV. J Acquir Immune Defic Syndr. 2009;51:S21–S25.
2. Magis-Rodríguez C, Gayet C, Negroni M, et al.. Migration and AIDS in Mexico. An overview based on recent evidence. J Acquir Immune Defic Syndr. 2004;37:S215–S226.
3. Organista KC, Organista PB. Migrant laborers and AIDS in the United States: a review of the literature. AIDS Educ Prev. 1997;9:83–93.
4. Harawa NT, Bingham TA, Cochran SD, et al.. HIV prevalence among foreign- and US-born clients of public STD clinics. Am J Public Health. 2002;92:1958–1963.
5. Bronfman M, Minello N. Sexual habits of Mexican migrants in the United States of America. Risk practices for HIV infection. In: Bronfman M, Amuchástegui A, Martina R, et al., eds. AIDS in México: Migration, Adolescence and Gender. 2nd ed. México, DF: Colectivo Sol; 1999:1–90.
6. Apostolopoulos Y, Sonmez S, Kronenfeld J, et al.. STI/HIV risks for Mexican migrant laborers: exploratory ethnographies. J Immigr Minor Health. 2006;8:291–302.
7. Sanchez MA, Lemp GF, Magis-Rodríguez C, et al.. The epidemiology of HIV among Mexican migrants and recent immigrants in California and Mexico. J Acquir Immune Defic Syndr. 2004;37:S204–S214.
8. Magana JR. Sex, drugs, and HIV: an ethnographic approach. Soc Sci Med. 1991;33:5–9.
9. Ruiz JD, Da Valle L, Jungkeit M, et al.. Seroprevalence of HIV and Syphilis, and Assessment of Risk Behaviors Among Migrant and Seasonal Farmworkers in Five Northern California Counties. Sacramento, CA: California Department of Health Services, Office of AIDS; 1997.
10. Mishra SI, Sanudo F, Conner RF. Collaborative research toward HIV prevention among migrant farmworkers. In: Bowser BP, Mishra SI, Reback CJ, et al., eds. Preventing AIDS: Community-Science Collaborations. New York, NY: The Haworth Press; 2004.
11. California Department of Health Services. Prevalence of HIV Infection and Related Risk Behaviors Among Young Latino Men Who Have Sex with Men: San Diego-Tijuana Border Region. Sacramento, CA: HIV/AIDS Epidemiology Branch, Office of AIDS; 2006.
12. Janes H, Sheppard L, Lumley T. Case-crossover analyses of air pollution exposure data: referent selection strategies and their implications for bias. Epidemiology. 2005;16:717–726.
13. Watters JK, Biernacki P. Targeted sampling: options for the study of hidden populations. Soc Probl. 1989;36:416–430.
14. Lumley T, Levy D. Bias in the case-crossover design: implications for studies of air pollution. Environmetrics. 2000;11:689–704.
15. Kleinbaum DG, Kupper LL, Morgenstern H. Applications of modeling: conditional likelihood estimation. In: Kleinbaum DG, Kupper LL, Morgenstern H, eds. Epidemiologic Research: Principles and Quantitative Methods. 1st ed. New York, NY: Van Nostrand Reinhold; 1982:492–507.
16. Prentice RL, Pyke R. Logistic disease incidence models and case-control studies. Biometrika. 1979; 66:403–411.
17. Agresti A. A survey of exact inference for contingency tables. Statist Sci. 1992;7:131–153.
18. Parrado EA, Flippen CA, McQuiston C. Use of commercial sex workers among Hispanic migrants in North Carolina: implications for the spread of HIV. Perspect Sex Reprod Health. 2004;36:150–156.
19. Organista KC, Balls Organista P, García de Alba JE, et al.. Survey of condom-related beliefs, behaviors, and perceived social norms in Mexican migrant laborers. J Community Health. 1997;22:185–198.
20. Ibáñez GE, Van Oss Marin B, Villarreal C, et al.. Condom use at last sex among unmarried Latino men: an event level analysis. AIDS Behav. 2005;9:433–441.
21. Organista KC, Organista PB, Bola JR, et al.. Predictors of condom use in Mexican migrant laborers. Am J Community Psychol. 2000;28:245–265.
22. Fosados R, Caballero-Hoyos R, Torres-López T, et al.. Condom use and migration in a sample of Mexican migrants: potential for HIV/STI transmission. Salud Publica Mex. 2006;48:57–61.
23. Galvan FH, Ortiz DJ, Martínez V, et al.. Sexual solicitation of Latino male day laborers by other men. Salud Publica Mex. 2008;50:439–446.
24. Arcia E, Skinner M, Bailey D, et al.. Models of acculturation and health behaviors among Latino immigrants to the U.S. Soc Sci Med. 2001;53:41–53.
25. Lara M, Gamboa C, Kahramanian MI, et al.. Acculturation and Latino health in the United States: a review of the literature and its sociopolitical context. Annu Rev Public Health. 2005;26:367–397.
26. Organista KC, Carrillo H, Ayala G. HIV prevention with Mexican migrants: review, critique, and recommendations. J Acquir Immune Defic Syndr. 2004;37:S227–S239.
27. Levy V, Page-Shafer K, Evans J, et al.. HIV-related risk behavior among Hispanic immigrant men in a population-based household survey in low-income neighborhoods of Northern California. Sex Transm Dis. 2005;32:487–490.
28. Pew Hispanic Center. Hispanics of Mexican Origin in the United States. Washington, DC: Pew Hispanic Center; 2010.
29. Castañeda D, Collins BE. Structure and activities of agencies providing HIV and AIDS education and prevention to Latina/Latino communities. AIDS Educ Prev. 1997;9:533–550.
30. Painter TM. Connecting the dots: when the risks of HIV/STD infection appear high but the burden of infection is not known–the case of male Latino migrants in the southern United States. AIDS Behav. 2008;12:213–226.
31. Tang JJ, Levy V, Hernandez MT. Who are California's late HIV testers? An analysis of State AIDS surveillance data, 2000-2006. Public Health Rep. 2011;126:338–343.
32. Espinoza L, Hall HI, Selik RM, et al.. Characteristics of HIV infection among Hispanics, United States 2003-2006. J Acquir Immune Defic Syndr. 2008;49:94–101.