Sex and drug risk behavior pre- and post-emigration among Latino migrant men in post-Hurricane Katrina New Orleans

Abstract

High rates of sex and drug risk behaviors have been documented among Latino migrant men in the U.S. Whether these behaviors were established in the migrants’ home countries or were adopted in the U.S. has not been described and has implications for prevention strategies. Quarterly surveys were conducted to gather information on selected sex and drug risk practices of Latino migrant men who arrived in New Orleans after Hurricane Katrina seeking work. Both kappa scores and McNemar’s tests were performed to determine if practice of these behaviors in home country was similar to practice post-emigration to the U.S. Female sex worker (FSW) patronage, same sex encounters (MSM), and crack cocaine use was more likely to occur post-rather than pre-emigration. Of those who ever engaged in these selected behaviors, most adopted the behavior in the U.S. (i.e. 75.8% of FSW patrons, 72.7% of MSM participants, and 85.7% of crack cocaine users), with the exception of binge drinking (26.8%). Men who were living with a family member were less likely to adopt FSW patronage OR=0.27, CI=0.10-0.76, whereas men who earned >$465 per week were more likely to adopt crack cocaine use OR=6.29 CI=1.29, 30.57. Interventions that facilitate the maintenance of family cohesion and provide strategies for financial management may be useful for reducing sex and drug risk among newly arrived migrants.

Background

Separation from family, cultural changes, and language barriers experienced by Latino migrant men in the United States can lead to increased substance use and risky sexual practices, ultimately placing these men at elevated risk of contracting HIV and becoming addicted.^1,2 Although risky sex behavior and substance use among migrant men in the US has been well documented, ^1-5 no studies to date have investigated whether these observed behaviors were adopted pre- or post-emigration.

Cross-sectional studies provide some indication that Latino migrant populations in the United States participate in sex and drug risk behaviors more frequently than men in their home countries^1,6-10. For example, while the rate of female sex worker (FSW) patronage was estimated at 5% in Mexico,⁶ rates among Latino migrant men in the US have ranged from 16%⁷ to 28%¹ in North Carolina to 42% in San Francisco⁸ to 52%⁹ - 74%¹⁰ in New Orleans. Furthermore, most migrants living with HIV in the U.S. acquired the disease post-emigration, suggesting that risky sex and drug behaviors were adopted in the U.S.^4,5

Latino migrants in new receiving communities

The contribution of new receiving environments to risk behaviors among Latino migrant men has been understudied. There are presently an estimated 8.8 million undocumented Latinos in the U.S., representing 3% of the population and 4.2% of the work force.¹¹ Latino migrants are increasingly moving to new receiving communities to seek greater economic opportunities. Between 1990 and 2000, the number of Latinos in selected southern states that had no or small Latino communities increased by more than 300% on average.¹²

Individuals entering new receiving communities are typically young, unmarried men, and may be more transient and isolated due to the lack of existing social infrastructure compared to those in established communities.^9,13 New receiving environments may foster riskier behaviors than those observed in similar cohorts living in frequent migrant destinations, since the normal social support and social controls are not present. For example, one outbreak of syphilis occurred among Latino migrants in Alabama, also a new receiving environment, which was mostly attributed to patronage of female sex workers and the use of crack cocaine¹⁴.

Latino Migrants in New Orleans

While New Orleans has had a small, long established Honduran population, New Orleans nevertheless is considered a new receiving community for Latino migrants.¹⁵ The sharp increase in Latino immigration to the area was due to the sudden and high demand for cheap debris-removal and reconstruction laborers following Hurricane Katrina which struck the city in August 2005¹⁶. Between 2005 and 2006 alone, the estimated Latino populations of Orleans and Jefferson Parish rose from 3.1% to 9.6% and 8.1% to 9.7% respectively.¹⁷

Understanding the context and dynamics of highly transient and isolated populations is important in assessing HIV/STI risks in the U.S. and abroad. Migrants practicing risky behaviors in the U.S. and returning to their country of origin may contribute to the observed increase of HIV rates in Mexico and other Latin American countries.^18-20

The purpose of this study was to examine whether four selected HIV and drug risk behaviors (i.e. patronage of FSW, same sex encounters (MSM), crack cocaine usage, and weekly binge drinking) were adopted pre- or post-emigration. A secondary goal was to examine the association of selected individual and socioeconomic factors with adoption of these behaviors post-emigration to identify potential intervention points.

Materials and Methods

This quantitative research was a sub-study of a larger cohort study of sex and drug HIV risk among Latino migrant men in post-Katrina New Orleans. Methods have been presented in depth elsewhere and are briefly described here.¹⁰ The cohort was assembled using respondent driven sampling (RDS).

Study Sample Selection and Informed Consent

Participants were eligible if they were: male, Spanish-speaking, born in Mexico or Central America, 18 or older, who arrived in metropolitan New Orleans area post-Hurricane Katrina (August 29, 2005) seeking work. The later inclusion criteria was included to limit the study sample to migrants who were newly arrived in the New Orleans area.

RDS was used to approximate simple random sampling²¹. Eight recruitment sites (a soccer stadium, a trailer park, a community based organization, select street corners, and the parking lots of home improvement stores) were identified after initial quantitative and qualitative studies. Eight initial recruits who fit the eligibility criteria, “seeds”, were each given 3 coupons to distribute to other eligible men in their social network. The coupons gave basic instructions on how to join the study, contact information for study personnel, dates and locations where the recruiters could be found, and an expiration date. The coupons were embossed to avoid falsification and were sequentially numbered to track chains of participants to their respective seeds in order to assess the RDS recruitment strategy. Any individual who contacted study personnel before the expiration date and fit the eligibility criteria was given an explanation of the study, consented, interviewed, and given 3 coupons to distribute; a total of 282 referral coupons were distributed by participants. The participants received a $25 incentive for every referral recruited.

Participants were interviewed over a span of 30 months. Data from interviews done at baseline, 3, 6, (for post-immigration data) and 15 months (when the information about pre-emigration was added) were used for this analysis. The surveys were completed between October 2007 and May 2009. Of the original 125 participants (117 referred + 8 seed participants), 93 (74.4%) had a follow-up visit at the 15 month visit and were included in the study.

Human Subjects

The Institutional Review Board of Tulane University approved the study, and a Certificate of Confidentiality (DA-09-216) was also obtained from the Department of Health and Human Services, National Institutes of Health. To assure confidentiality, participants were interviewed in private rooms in one of the study locations (either a house or a van). All databases were de-identified and all records with identifiers were secured in locked cabinets.

Follow-up Visits

Recruitment and follow-up visits were conducted on the weekends with interviews occurring in a location determined by the participant. All participants received a $30.00 equivalent incentive after each survey in the form of cell phone minutes, a store gift card, or an international calling card, depending on their preference. The majority of the surveys were conducted in person (85.6%) and the remainder conducted over the phone. Phone interviews only included individuals who had moved more than 60 miles away from the New Orleans metropolitan area and contained the same information as the face-to-face interviews.

Survey instrument

The survey instrument was informed by formative research and was translated and back translated by native Spanish speakers from Honduras and Mexico. The instrument was pilot tested on 20 men in an iterative test-revise-test manner to ensure content validity.^22,23 Interviews consisted of questions pertaining to individual (e.g. demographics, alcohol and drug use) and socio economic factors, and living arrangements. All interviews were conducted by trained staff members in Spanish. Data were collected using computer assisted personal interviews (CAPI). This method also allowed for interviewers to answer questions participants had and to clarify terms during the interview process.

Country of origin behavior compared to New Orleans behavior

The four behaviors of interest were: patronage of FSWs, MSM, use of crack cocaine, and weekly binge drinking. Quarterly surveys inquired about behavior in New Orleans during the preceding month and the country of origin questions encompassed lifetime pre-emigration. For post-emigration behavior, we examined data from baseline, 3 and 6 months. Multiple time points were used: 1) because behavior is episodic, and 2) to allow participants enough time to engage in the behavior. Figure 1 depicts how the behaviors pre-and post-emigration were compared using crosstabulations. The wording of the questions is described below.

Figure 1.

Definition of terms and cross-tabulation methods

Categorization of FSW patronage

Participants who answered “yes” to the question at baseline, “In [home country] did you have sex with sex workers?” were categorized as FSW patrons in home country. Participants were categorized as FSW patrons in New Orleans if their response to, “How many of [the women you had sex with in the past month] were sex workers?” was greater than zero at the baseline, 3, or 6 month surveys.

Categorization of MSM

Participants who responded “yes” to, “In [home country] did you ever have sex with a man?” or, “Have you ever had sex with a man in New Orleans?” at the baseline, 3, or 6 month surveys were categorized as having a same-sex encounter (MSM) in the respective locations.

Categorization of Drug Use

Subjects who gave an affirmative response to, “When you were in [home country] did you use crack cocaine?” were categorized as using crack cocaine in their country of origin. At the baseline, 3 and 6 month surveys, participants were asked, “In the past month, how many days did you use crack cocaine?”. Participants were categorized as using crack cocaine in New Orleans if their response was greater than zero.

Categorization of weekly binge drinking

Weekly binge drinking was calculated using responses to: “When you were in [home country] how many times a week would you drink alcohol?” and, “About how many drinks would you drink each day you drank?” Those who reported consuming greater than 5 alcoholic beverages on a given occasion at least once a week were categorized as weekly binge drinkers per the definition described by the National Institute on Alcohol Abuse and Alcoholism²⁴. To assess binge drinking behavior in New Orleans, participants were asked, “In the past month, how often did you have a drink containing alcohol” and, “How many drinks containing alcohol did you have on a typical day when you were drinking?” An average of the number of days they drank in the month preceding baseline and the first two quarterly surveys was calculated, as was the number of drinks consumed on those occasions. As with home country assessment, those who drank greater than 5 drinks on one occasion at least once a week on average (greater than 4 times per month) were categorized as weekly binge drinkers in New Orleans.

Statistical Analysis

Percentages, medians and ranges were calculated for descriptive information since many of the continuous variables were not normally distributed. Risk behaviors were categorized dichotomously (participated versus did not participate in a given behavior). Crosstabulations between behaviors pre- and post-emigration were conducted and subjects’ behaviors were categorized as continuation, termination, adoption, or abstention (Figure 1). The percentage of discrepant behavior accounted for by adopters was calculated by dividing the number of adopters by the sum of the adopters and terminators.

Kappa values were calculated for each variable to assess the degree of similarity between behaviors practiced in New Orleans and those reported in the country of origin. Variables with poor agreement (kappa values less than 0.40) indicated discrepant behavior between the two locations²⁵. McNemar’s test for paired measurements was then performed to test the hypothesis that the behaviors were more likely to occur post-emigration. A p-value of <0.05 indicated that pre- and post-emigration behaviors were statistically different and examination of the percentages indicated the direction of the difference.

Factors associated with risk behavior adoption post-emigration

For the three risk behaviors that were different pre- and post-emigration (i.e. patronage of FSWs, MSM, and crack cocaine use), adopters were compared to abstainers by selected individual and socioeconomic factors to explore potential influences affecting adoption. Factors examined included: age, marital status, location of a spouse or long term partner, birth country, migration origin (sending country versus other U.S. city), years of education, time lived in New Orleans, total time lived in the U.S., mobility, travel partners during migration to New Orleans, number and type of people in household (family member, women, and children), participation in a club or organization, job type, average weekly income, and use of the English language. Chi-squared statistics and Mantel-Haenszel unadjusted odds ratios were calculated to assess the significance of these factors. All statistical analyses were performed in Statistical Packages for Social Sciences (SPSS) (IBM, New York, New York) Version 19.

Results

Cohort characteristics

Ninety-three men were included in the cohort (i.e., enrolled in the parent cohort study and completed a 15 month survey). The median age was 28.0 (range 18 to 50) with a median of 6.0 (0 to 22) years of education. The ability to understand English “very well” or “somewhat” was reported by 49.5% of the participants while 50.5% could not understand English at all, and less (33.3%) reported speaking English “very well” or “somewhat”, while 66.7% could not speak English at all. The majority of men were Honduran (71.0%), who came to New Orleans directly from their country of origin (62.0%) and lived in the city a median of 17 months at baseline (range 1 to 27). Of the 35 men who arrived in New Orleans from other areas in the U.S., 24 provided information on how long they were in the U.S. before coming to New Orleans. Of those 24, the median time in the U.S. was 36 months (range 12 to 240) before migrating to New Orleans.

Few men (7.5%) lived with their wife or long term partner, 38.7% had a wife or long term partner living outside the New Orleans area, and 53.8% were single or divorced. Subjects lived in a household with a median of 4 individuals (range: 0-9). Less than half (40.2%) lived with at least one family member, 41.9% lived in a household with at least one woman (not necessarily a partner), and only 14.1% lived in a household with a child under 18 years of age. At baseline, 10.9% belonged to a club or other social organization (a church, a union, Alcoholics Anonymous, soccer club). The primary form of employment was construction with 74.2% of the subjects working in this field, and the median average weekly income calculated from the first three quarterly surveys was $463.33 (range $233.33-$1,116.67). (Table 1-3)

Table 1.

Study Population Demographics (n=93)

Characteristic	Median (Range)
Age in years	28 (18-50)
Education in years	6 (0-22)
People in household	4 (0-9)
Weekly Income*	$463.33 ($233.33-$1116.67)
Months lived in New Orleans	17 (1-27)
Months lived in U.S. before New Orleans (n=24)	36 (12-240)
Total months lived in the U.S.	24 (1-249)

^*Average $ per week at baseline, 3 and 6 months

Table 3.

Socio-economic characteristics of study population (n=93)

	N(%)
Marital status
Single	47 (50.5)
Married	35 (37.6)
Long term partner (unwed)	8 ( 8.6)
Separated/divorced	3 ( 3.2)
Lives with family member (n=87)
Yes	35 (40.2)
No	52 (59.8)
Women in household
Yes	39 (41.9)
No	54 (58.1)
Child in household (<18 yrs of age) (n=85)
Yes	12 (14.1)
No	73 (85.9)
Belongs to club/organization (n=92)
Yes	10 (10.9)
No	82 (89.1)
Works in Construction
Yes	69 (74.2)
No	24 (25.8)
Understands English
Very well/Somewhat	46 (49.5)
Not at all	47 (50.5)
Speaks English
Very well/ Somewhat	31 (33.3)
Not at all	62 (66.7)

Risky sex and drug behaviors in home country versus New Orleans, LA

Participation in risky sex and drug behaviors in New Orleans (patronage of FSWs, MSM, crack cocaine use, and weekly binge drinking) was dissimilar to the participants’ behaviors in their country of origin as measured by kappa values: -0.01, 0.27, -0.04, and 0.23 respectively). Post-emigration adoption accounted for 89.3%, 88.9%, 85.7% and 44.1% of the discrepant behavior, respectively. (Table 4)

Table 4.

Comparison of Country of origin and New Orleans risk behavior (n=93).

Risk Behavior		Kappa	McNemar’s p-valuea
FSW patronage		−0.01	<0.001
Continuation	10/93 (10.7%)
Termination	6/93 (6.5%)
Adoption	50/93 (53.8%)
Abstention	27/93 (29.0%)
MSM		0.27	0.039
Continuation	2/93 (2.1%)
Termination	1/93 (1.1%)
Adoption	8/93 (8.6%)
Abstention	82/93 (88.2%)
Crack cocaine use		−0.04	0.013
Continuation	0/93 (0.0%)
Termination	2/93 (2.2%)
Adoption	12/93 (12.9%)
Abstention	79/93 (84.9%)
Weekly binge drinking (n=91)		0.23	0.608
Continuation	20/91 (22.0%)
Termination	19/91 (20.9%)
Adoption	15/91 (16.5%)
Abstention	37/91 (40.6%)

^aExact binomial distribution used for McNemar’s test

The proportion of adopters was significantly higher than the proportion of those terminating FSW patronage, MSM, and crack cocaine use post-emigration as measured by McNemar’s test, (P= 0.001, 0.039, and 0.013, respectively). Of those who ever engaged in the selected behaviors, most adopted the behavior in the U.S. (i.e. 75.8% of FSW patrons, 72.7% of MSM participants, and 85.7% of crack cocaine users), with the exception of binge drinking (26.8%). (Table 4)

Individual and socioeconomic factors and behavior adoption

Several factors were less likely among those who adopted FSW patronage compared to those who abstained. These were living with a family member (nuclear or extended) (OR= 0.27 with 95%CI: 0.10, 0.76), having children (regardless of where they lived) (OR=0.28; 95%CI: 0.09, 0.95), and having a woman living in the home (not necessarily a partner) (OR=0.38; 95%CI: 0.14, 0.99). (Table 5).

Table 5.

Unadjusted Odds Ratios of individual and socioeconomic factors—comparing adopters versus abstainers

Factors	FSW patronage OR(95%CI)	MSM OR(95%CI)	Crack cocaine use OR(95%CI)
	n=77	n=90	n=91
Honduran	1.67 (0.61, 4.57)	3.79 (0.46, 31.58)	4.80 (0.59, 39.30)
Age≤28	1.39 (0.54, 3.58)	1.09 (0.28, 4.14)	1.51 (0.42, 5.44)
Education≤6	1.60 (0.60, 4.26)	0.50 (0.13, 1.89)	2.38 (0.47, 12.02)
Has spouse/long term partner	0.53 (0.21, 1.38)	1.13 (0.30, 4.21)	0.77 (0.23, 2.64)
Has children	0.28 (0.09, 0.95)†	1.00 (0.24, 4.20)	0.55 (0.16, 1.92)
Migrated from other U.S. area	2.74 (0.98, 7.66)	0.68 (0.16, 2.82)	1.21 (0.35, 4.15)
Migrated with family/friends	1.36 (0.45, 4.08)	1.31 (0.31, 5.54)	1.58 (0.43, 5.83)
Months lived in New Orleans ≤18	0.95 (0.37, 2.46)	0.72 (0.19, 2.70)	1.06 (0.31, 3.62)
Total months lived in U.S.≤24	0.71 (0.28, 1.82)	2.22 (0.54, 9.20)	0.86 (0.25, 2.89)
Cannot understand English	1.07 (0.66, 4.41)	1.75 (0.39, 7.81)	0.44 (0.12, 1.58)
Cannot speak English	2.39 (0.90, 6.33)	4.04 (0.47, 34.43)	1.56 (0.39, 6.23)
>4 people in household	0.62 (0.23, 1.67)	1.24 (0.31, 4.97)	0.73 (0.20, 2.62)
Lives with long term partner	0.24 (0.04, 1.41)	0 lived with LTP	0 lived with LTP
Lives with family member	0.27 (0.10, 0.76) *	0.39 (0.08, 2.02)	0.26 (0.05, 1.25)
Woman in household	0.38 (0.14, 0.99)†	0.54 (0.13, 2.22)	0.95 (0.28, 3.24)
Child (<18yrs old) in household	0.44 (0.12, 1.62)	0 lived with child	0.72 (0.08, 6.30)
Moved in first 6 months of study	0.54 (0.17, 1.71)	1.40 (0.33, 5.94)	0.59 (0.12, 2.92)
Works in construction	1.20 (0.42, 3.39)	1.40 (0.28, 7.13)	All users worked construction
Average income >$465/week for first 6 months of study	1.17 (0.46, 2.98)	1.03 (0.28, 3.81)	6.29 (1.29, 30.57) *
Belongs to a club/organization	1.09 (0.25, 4.76)	0 belonged to club	0 belonged to club

^*p-value <0.05

^†p-value <0.08

Those who adopted crack cocaine in the US were 6.29 times more likely than those who abstained (95%CI: 1.29, 30.57; P=0.028) to have earned more than $465 per week on average (the median income). All crack cocaine adopters worked in construction (Table 5).

No factors were found to be statistically associated with adoption of MSM behavior.

Discussion

In this analysis, we found that three of the four selected behaviors (FSW patronage, MSM, and crack cocaine use) were likely adopted in New Orleans. Post-Katrina New Orleans was a harsh environment for Latino migrants^16,26. The transient nature of migrant work, isolation, and lack of social structure observed in new receiving communities²⁷ compounded by the easy access to drugs and sex partners in New Orleans²⁴ provided an opportunity to engage in these behaviors.

Having a child and living with a woman or family member were found to be protective against adopting FSW patronage, suggesting that social influences are important in risk behavior. It was interesting that having own’s own child living in the home was not associated. This may be due to insufficient power to detect the association or some other mechanism. Studies have demonstrated the importance of context and environment on HIV and drug risk^28,29 and have suggested that policies that result in separation of families may increase risk, and interventions that promote connection have the potential reduce risk in this group.

Higher income (>$465 per week) was associated with crack cocaine adoption. It was interesting to note that all men who adopted crack cocaine use worked in construction. Since many migrant men are paid in cash³⁰, it is possible that having cash in hand facilitated this behavior. While the interpretation of these finding should be made with caution, given the small sample size, interventions that increase options for financial management, such as credit unions and worker’s rights, may reduce risk behaviors related to cash economies.

The finding that binge drinking was not as likely to be adopted in the U.S. corroborates that of others³¹ and was not surprising given that binge drinking among men is generally considered acceptable in many Latin American cultures.³² However, there was some indication that binge drinking behavior is episodic. While the overall proportion of binge drinkers remained relatively steady (a 4.4% difference between the country of origin and New Orleans) a substantial proportion of the cohort adopted weekly binge drinking post-emigration (20.9%) while a similar proportion terminated the behavior (16.5%) resulting in the low Kappa value.

Limitations

There are some limitations to this study. Since this was a pilot study and the sample size was small (n=93), thus associations should be assessed with care (especially those with large confidence intervals). Multivariable regression was not possible so we could not explore for confounding. Moreover, some factors may have been associated, but because of insufficient power, we were not able to detect associations. Larger studies are needed.

Generalizability may also be an issue. The study was conducted in a unique city with post-disaster circumstances. New Orleans has a high prevalence of STIs, HIV, and drug use, in addition to the vibrant adult entertainment industry²⁶. Thus, the findings may not be generalizable to all areas in the United States. Studies of this population in diverse settings are needed.

Over one-third (38%) of the cohort migrated to New Orleans from another area in the U.S. and did not come directly from their home country. Since we did not capture information on these behaviors in those areas we do not know if the men adopted them before arriving in New Orleans. However, it is still accurate to say they were adopted in the U.S.

Measurement of the behaviors had some limitations. Questions regarding behaviors practiced in the country of pre-emigration required recall exceeding one year. Furthermore, recall periods for comparisons pre-and post-emigration were not similar (i.e. lifetime for pre-emigration and six months for post-emigration). While some behaviors may be memorable (FSW patronage or an MSM encounter), others such as the amount of alcohol consumed in an average day or week may not be. Finally, the information gathered in this paper is self-reported and was face-to-face rather than self-administered. It is possible given the vulnerable nature of the population and the sensitivity of topics, the participants were not as forthcoming due to fear of arrest and deportation and social desirability.

By using the first six months of data, we had up to six interviews with the participant, allowing for sufficient time to develop rapport and possibly reduce reporter bias. Others have reported the importance of developing rapport in working with Latino migrants^7,33. Also it is possible that assessing the pre- and post-emigration behaviors at different times may have reduced telescoping. Moreover, since rates of adoption of these illicit behaviors, per self-report, was fairly high, we think that this potential source of error was minimal.

Conclusion

This study demonstrates that among newly arrived Latino migrant men, patronage of FSW, MSM, and crack cocaine use were behaviors likely adopted in the U.S. Given the high rates of the selected risk behaviors in this population, interventions to prevention adoption of these high risk behaviors among newly arrived migrant men are greatly needed. Some potential points of intervention that could help are reuniting migrants with their family or helping them to develop a sense of community. Other interventions could be to help migrants with assistance with financial management.

Table 2.

Migration Information (n=93)

	N (%)
Country of Birth
Honduras	66 (71.0)
Mexico	10 (10.8)
Guatemala	7 (7.5)
Nicaragua	5 (5.4)
El Salvador	5 (5.4)
Migration Origin
From home country	57 (62.0)
From other area of U.S.	35 (38.0)
Spouse/long term partner location
New Orleans	7 (7.5)
Outside New Orleans	36 (38.7)
Does not have spouse/long term partner	50 (53.8)
Migration group
Migrated with family or friends	23 (24.7)
Migrated alone	70 (75.3)
Family lived in New Orleans upon arrival (n=91)
Yes	33 (36.3)
No	58 (63.7)
Moved in first 7 months of study
Yes	22 (23.7)
No	71 (76.3)

Appendix

Continuation – behavior was reported in both home country and in New Orleans

Termination – behavior was reported in home country but not in New Orleans

Adoption – behavior was not reported in home country but reported in New Orleans

Abstention – behavior was not reported in home country or in New Orleans.