Objective: To investigate how mobility is related to sexual risk behavior and HIV infection, with special reference to the partners who stay behind in mobile couples.
Methods: HIV status, sexual behavior and demographic data of 2800 couples were collected in a longitudinal study in Kisesa, rural Tanzania. People were considered short-term mobile if they had slept outside the household at least once on the night before one of the five demographic interviews, and long-term mobile if they were living elsewhere at least once at the time of a demographic round.
Results: Overall, whereas long-term mobile men did not report more risk behavior than resident men, short-term mobile men reported having multiple sex partners in the last year significantly more often. In contrast, long-term mobile women reported having multiple sex partners more often than resident women (6.8 versus 2.4%; P = 0.001), and also had a higher HIV prevalence (7.7 versus 2.7%; P = 0.02). In couples, men and women who were resident and had a long-term mobile partner both reported more sexual risk behavior and also showed higher HIV prevalence than people with resident/short-term mobile partners. Remarkably, risk behavior of men increased more when their wives moved than when they were mobile themselves.
Conclusions: More sexual risk behavior and an increased risk of HIV infection were seen not only in mobile persons, but also in partners staying behind. Interventions aiming at reducing risk behavior due to mobility should therefore include partners staying behind.
From the aTanzania Essential Strategies Against AIDS (TANESA), Mwanza, Tanzania
bDepartment of Public Health, Erasmus MC, University Medical Center Rotterdam, The Netherlands
cNational Institute for Medical Research, Mwanza, Tanzania
dCentre for Populations Studies, London School of Hygiene and Tropical Medicine, London, UK.
*Both authors contributed equally to this paper.
Received 30 August, 2005
Revised 30 November, 2005
Accepted 7 December, 2005
Correspondence to Debby Vissers, Department of Public Health, Erasmus MC, University Medical Center Rotterdam. PO Box 1738, 3000 DR Rotterdam, The Netherlands. E-mail: email@example.com
Mobility is one of the many factors that have contributed to the AIDS epidemic [1–3]. Several studies have shown that people who travel or who have recently migrated tend to be at higher risk for HIV and other sexually transmitted diseases (STD) [4–8]. The role of migration in the spread of HIV has been described primarily as the result of men who become infected while they are away from home, and infect their wives or regular partners when they return [1,4,9]. Married men often travel without their spouses. Being away from their families and communities, and thus from social and sexual control, may cause mobile men to change their behavior. They may have sex with more women than if they had stayed at home . On the other hand, due to differences in languages and culture, sexual partnerships with local women in the destination area may be difficult. This leads mobiles to have sex with commercial sex workers, who often have high rates of HIV and other STD infections.
Most studies tend to give a one-sided view which only takes mobile people into account and do not consider those who stay behind. Due to a number of factors, such as loneliness, peer pressure, and lack of financial support, partners who stay behind may also engage in riskier sexual behavior. Consequently, people are not only vulnerable to HIV infection by the risk behavior of their partners, but also by their own risk behavior when left behind.
A South African study investigated HIV infection among migrants and their partners staying behind and among non-migrant couples in which both partners stayed at home . This study showed that HIV discordance was 2.5 times more likely in migrant couples than in non-migrant couples. Men and women in a migrant couple were both more likely to be infected from outside the relationship than by their spouse. This study also found that in one-third of the couples with only one HIV-positive partner, the wife who stayed at home was infected . Therefore, understanding the sexual behavior of both partners within a couple is essential for the successful implementation of targeted interventions.
This study was carried out in Tanzania, where the HIV epidemic is spreading at an alarming pace since the first three AIDS cases were reported in 1983. In the period 1990 to 2000, surveillance reports indicated an increase in the HIV prevalence from 8.9 to 12.2% among pregnant women in antenatal clinics . By the end of 2003, it was estimated that about 1.6 million people in Tanzania were living with HIV/AIDS . HIV prevalence in the sexually active population (15–49 years) was estimated at 8.8% in 2003 . AIDS is now the major cause of illness and death in all economic sectors and at all social levels .
Our study aimed to establish whether men and women who are part of couples in which one of the partners is mobile show more sexual risk behavior and a higher HIV prevalence than continuously co-resident men and women. In particular, we were interested in whether absence of the mobile partner increased the risk behavior of the partner staying behind.
Study site and data collection procedure
Data were available from a longitudinal cohort study in Kisesa Ward in Mwanza Region, Tanzania [14,15]. The main objective of this study was to monitor the spread and impact of the HIV/AIDS epidemic in the Kisesa community and to identify possible risk factors (including mobility). It started in 1994 and is still ongoing. The ward has a population of about 28 000 people and is administratively divided into six villages, three of which are located along one of the main roads connecting Tanzania and Kenya. The main economic activity among the residents of the area is farming (97%).
Between 1994 and the end of 2003, 16 demographic surveillance rounds were completed. Per round, information was collected from each household (defined as a group of people who regularly eat together from the same pot) on the residence and survival status of all household members, on pregnancies in women, on mobility behavior and on new arrivals (migrants and new-borns). A new person was listed as a household member if the household respondent had indicated that this person was intending to stay in the household. Returning household members were re-listed, keeping their original line number on the household card.
Between 1994 and 2000, three rounds of epidemiological and behavioral surveys were conducted: the first in 1994/95, the second in 1996/97, and the third in 1999/2000. In all three surveys, participants were interviewed using standardized questionnaires (in Swahili). Sociodemographic characteristics, birth and marital history, family planning, sexual behavior, STD history, HIV/AIDS awareness, and risk perception were asked. Participants were also asked to provide a blood sample for HIV screening. Blood samples were tested using two independent enzyme-linked immunosorbent assay (ELISA) tests: Vironostika HIV-MIXT (Organon, Boxtel, The Netherlands) and Enzygnost HIV1/HIV2 (Behring, Marburg, Germany). Only samples with two positive ELISA results were considered to be HIV positive. All participants were given study numbers to guarantee anonymity.
Men and women in long-term relationships or marriages (marital units) were identified in 1996 and again in 2002. For all marital units, demographic interview information was used to identify the period of co-residence. Marital units were only included in our analyses if they had been living together before or during the studied demographic rounds and if at least one of the partners had been tested for HIV infection at survey 2 or 3. Those marital units with members who moved, separated or died before the identification were not included.
Two indicators, both asked in demographic surveillance rounds, were used to define the mobility status: ‘slept outside the household on the night before the demographic round’ and ‘lives in another household since the last demographic round’. We used data from the five demographic rounds between survey 1 and 2 to define the mobility status of people attending survey 2. For people in survey 3, we used the five rounds between survey 2 and 3. A person was considered to be short-term mobile if he or she had slept outside the household at least once the night before one of the five demographic rounds, and considered to be long-term mobile if he or she had been living elsewhere at least once. Residents were all people that did not sleep outside the household nor lived elsewhere at the time of the demographic rounds. Long-term mobiles were defined first, followed by short-term mobiles. This procedure led to exclusive mobility categories.
For men and women all analyses were done separately. Polygamous marriages with men married to two or three women were included in our study. However, in our analyses a couple is defined as one man married to one woman. Therefore, for the analyses of men we kept only one record in case of a polygamous marriage and we randomly chose one of his wives. For the analyses of women, we included all women of whom some share the same man.
For each of the three mobility strata, proportions were calculated for different indicators of sexual risk behavior and of STD/HIV status. The sexual risk indicators were: having a regular non-spousal partner in the last year (i.e. long-term relation with someone besides the husband/wife), and having a casual partner in the last year (i.e. short-term relation). In addition, the number of sex partners in the last year was asked. STD/HIV indicators were reported ulcers in private parts in the last year, and prevalence and incidence of HIV. To adjust for the confounding effect of age, proportions were standardized using the age distribution of the total number of men or women. To test whether short-term or long-term mobile people differed from residents, we used logistic regression modeling, adjusting for age (which was the only significant demographic confounder).
After these analyses based on the participant's own mobility, we studied the effects of the mobility status of their partners. We further stratified the groups of men and women according to the mobility status of their partners, and calculated age-adjusted proportions. As there was only one long-term mobile man with a short-term mobile wife, and because earlier analyses yielded only minimal differences between short-term and long-term mobile men, we decided to combine these two groups into one. In a second analysis, we formed groups based on the mobility status of both partners and studied the differences between these groups with logistic regression analysis adjusting for age. The groups were divided as follows: both resident; being short-term or long-term mobile yourself with a partner at home; being at home with a short-term or long-term mobile partner; and both being short-term or long-term mobile. We combined the groups of short-term and long-term mobile people in the analyses of women, because some categories did not have enough observations to perform logistic regression analyses.
We determined the HIV incidence using two successive surveys (i.e. survey 1 and 2, or survey 2 and 3). As not all people attended two successive surveys, the numbers of people in the analyses of HIV incidence were lower.
The results of survey 2 and 3 were combined to increase power. Therefore, individuals who attended both surveys were included twice. We applied generalized estimation equation (GEE) techniques for logistic regression to account for dependencies between repeated observations and for dependencies resulting from inclusion of polygamous marriages . All analyses were performed using Stata version 8.0 (Stata Corporation, College Station, Texas, USA).
In total, we identified 2800 marital units in which at least one of the partners was tested for HIV. These marital units consisted of 2614 monogamous relationships, 175 men with two wives, and 11 men with three wives. Basic demographic information was available for all of the individuals involved in these partnerships, but HIV status, sexual risk behavior, and mobility characteristics were known for 1675 out of 2800 men (59.8%) and 2185 out of 2997 women (72.9%). The analyses that also involved the mobility status of the partner were limited to 1541 men and 2157 women, respectively. Most long-term mobiles (60%) temporarily lived elsewhere and returned to their original households in the next demographic rounds. Others lived in a place nearby and were still able to attend the next survey round.
Of the men, 1049 were resident, 474 short-term mobile (i.e. slept outside the household at least once) and 152 long-term mobile (i.e. lived elsewhere at least once). Of the women, 1534 were resident, 444 short-term mobile, and 207 long-term mobile. Age distribution of men and women and the mobility status of their partners are shown in Table 1. Long-term mobile men were considerably younger than short-term mobile or resident men. The majority of resident men also had a resident partner (70.2%), and most long-term mobile men had a long-term mobile partner (81.6%). Half of the short-term mobile men had a resident wife, and about one-third had a short-term mobile wife. Women showed a similar age pattern as men. About two-thirds of the resident women and two-thirds of the long-term mobile women had a husband with the same mobility status. Almost half of the short-term mobile women had a resident husband and another 50% a short-term mobile one.
The proportion of men having sex with regular non-spousal or casual partners did not consistently differ between the three mobility groups (Table 2). However, short-term mobile men reported significantly more often than resident men that they had more than two sex partners in the last year (47.8 versus 40.0%; P = 0.006). The proportion reporting ulcers in private parts was somewhat higher for short-term (6.7%) and long-term mobiles (5.7%), but this did not differ significantly from resident men (4.6%). The HIV status did not differ significantly between the three mobility groups, although short-term and long-term mobile men had a slightly lower prevalence and incidence of HIV.
Sex with a regular non-spousal partner was more common among long-term mobile women than among resident women (8.3 versus 2.0%; P < 0.001) (Table 2). Long-term mobile women also reported sex with casual partners (5.2 versus 1.8%; P = 0.004) and with multiple partners in the last year (6.8 versus 2.4%, P = 0.001). The increased risk behavior among long-term mobile women was accompanied by a significantly higher HIV prevalence than in resident women (7.7 versus 2.7%; P = 0.02), and a slightly higher HIV incidence. There were no particular differences in sexual risk behavior and HIV status between short-term mobile and resident women.
After looking at participants' own mobility, we studied the effects of the mobility behavior of the partner. For resident men, the mobility status of their wives was strongly associated with their own sexual risk behavior and HIV/STD status (Tables 3 and 5). The overall test on the mobility status of couples was for most outcomes either significant or borderline significant (see P-values in Table 5). Resident men with long-term mobile wives reported significantly more regular non-spousal [30.9 versus 11.7%; odds ratio (OR) = 2.65], casual (36.3 versus 23.0%; OR = 2.15) or multiple sex partners (62.4 versus 36.7%; OR = 2.76) in the last year than resident men with resident partners. These men also had a higher proportion of ulcers in private parts (11.9 versus 3.8%; OR = 3.95) and a higher HIV prevalence (11.2 versus 5.5%; OR = 2.79) and incidence (6.0 versus 2.0%; OR = 4.22). Resident men with short-term mobile wives reported more casual sex partners, more regular non-spousal partners and multiple sex partners in the last year than those with resident wives, although the differences were not significant. Surprisingly, men's sexual behavior seemed to be more risky if their wives moved than if they were mobile themselves. Short-term mobile men with a partner at home had an odds ratio of 1.59, in comparison with resident men with resident wives, of reporting two or more sex partners in the last year. If both partners were mobile, men reported more casual partners [OR = 1.41; 95%confidence interval (CI), 1.03–1.92] and more ulcers in private parts (OR = 1.90; 95% CI 1.08–3.35) than if both partners were resident.
Table 4 shows sexual behavior data and STD prevalences of women with partners in different mobility groups. The overall test on the mobility status of couples was only significant for having a regular non-spousal partner in the last year (see P-values in Table 5). Resident women reported markedly more casual partners if their husbands were mobile (3.0 versus 1.3%; OR = 2.45); they also had a higher, although not significant, HIV prevalence and incidence than resident women with resident partners (Tables 4 and 5). In contrast to resident women, the mobility status of the partners did not have much influence on the sexual behavior or STD status of short-term mobile women. If both partners were mobile, women had an odds ratio of 2.93 compared with resident women with resident husbands to report sex with a regular non-spousal partner (Table 5). These women also reported more casual partners (OR = 2.31; 95% CI, 1.06–5.05). Overall, long-term mobile women with mobile partners showed the highest HIV prevalence (8.6%).
The mobility status of men did not greatly influence their sexual risk behavior or STD status. In contrast, long-term mobile women reported more often than resident women that they had non-spousal, casual or multiple partners in the last year and they had a higher HIV prevalence. The risk behavior of men was influenced more by the mobility of their partner than by their own mobility. Long-term mobile women with mobile partners reported more sexual risk behavior, which was also reflected in a higher HIV prevalence.
There are some limitations in our study design. The indicator ‘slept outside the household the night before the demographic round’ is a good indicator for population mobility, but a relatively poor indicator at the individual level. For instance, people who often sleep outside the household may, by chance, be found at home during a demographic round, and thus be classified in the study as a resident. A similar reasoning can be followed for people who rarely sleep elsewhere but were not found at the time of the demographic round, and consequently classified as short-term mobile. In spite of this, our results show that being short-term mobile or having a short-term mobile partner is a risk factor for increased sexual risk behavior and HIV infection. With the use of more precise indicators of mobility this pattern may well become stronger. Such an alternative indicator could be ‘the number of nights spent outside the household during the last month’, but this has the limitation of recall bias.
Another limitation was the couple identification, which occurred after the study period. It is possible that some couples were not included due to HIV/AIDS-related causes. Some people might have lost their partner due to AIDS. Other couples might have split up when one of the partners found out that their husband/wife was HIV positive. It is therefore possible that the number of HIV cases in our study is biased towards a lower value.
About 60% of the men and 70% of the women were included in our analyses. Men who were included were somewhat younger, but included women were slightly older than persons not included. Mobile people were relatively often not included in the analyses, because they did not attend the survey rounds.
Male circumcision could be a potential confounder in the association between mobility and HIV infection. The circumcision status was known for 80% of the men in our analyses and mobile men were slightly more often circumcised than resident men (24 versus 19%). Adjusting for male circumcision did not change the association between mobility and sexual risk behavior or HIV infection.
Our study demonstrated an association between people's mobility behavior, their sexual risk behavior and their HIV status. This is consistent with earlier findings that migration and travel are affecting the spread of HIV [4–8]. A study in South Africa showed that migrant women were significantly more likely than non-migrant women to have had two or more partners in the last year and to have had sexual contact with a partner other than the regular partner . This was accompanied by a higher HIV prevalence in migrant women . In addition to this, long-term mobile women in our study also showed increased sexual risk behavior and had a higher HIV prevalence than resident women.
Resident women with mobile partners reported more casual partners in the last year than those with resident partners. A possible explanation could be that women with mobile partners have more opportunity to engage themselves in sexual relationships with other men. This may not always be out of free choice. Sometimes men who travel or move do not leave money for their wives. In a South African study only half of the migrant men did send money back home . Some women who stay behind may be compelled to engage in transactional sex for food and other living expenses.
Most studies have focused on one type of mobility. Some studies looked at seasonal migration, circular migration or migration in relation to work [4,9,19]. Studies in Cameroon and Uganda investigated the relation between traveling, sexual behavior, and HIV infection [8,20]. We compared resident people with both short-term and long-term mobiles. This enabled us to study differences in sexual behavior and HIV status as a result of different types of mobility. As in an earlier study in South Africa, we followed both the partner moving away and the one staying behind [11,18]. The South African study was restricted to men, whereas we also examined mobile women and their partners. We saw that long-term mobile women, in particular, showed more risky sexual behavior and had a higher HIV prevalence. Since moving rates of women in Kisesa are high (70% lived elsewhere at least once in their life), these results indicate that long-term mobile women play an important role in the spread of HIV .
We conclude that both partners, namely the one moving away and the one staying behind, showed more sexual risk behavior and had a higher risk of HIV infection. Current interventions mainly target mobile people (e.g. miners) and places where they gather (e.g. truck stops), and consist of STD services, condom distribution and education [22,23]. This study shows that policy-makers should be aware of partners staying behind. They should aim their HIV interventions not only at mobiles, but also at partners at home. A first option could be health education or condom distribution focused at partners staying behind in rural areas. In the case that this is too costly, another option might be to encourage partners to move to the new area together, by creating the right circumstances for housing, employment and schooling opportunities for children. However, moving with your partner may not always be practical in Tanzania. Most people who move to seek casual employment can not risk bringing their family, because they will often share a room with relatives, who may tolerate one person but not a whole family. In other areas moving together could be a practical intervention. If partners move together, they may be less prone to engage in risky sexual behavior, and may therefore be less likely to acquire an HIV infection.
The authors wish to thank the directors of the National Institute for Medical Research and the TANESA program, Tanzania, for their support and assistance in carrying out this study.
Sponsorship: This work was supported by a grant from the European Union (Grant no. B7.6211/99/010).
1. Decosas J, Adrien A. Migration and HIV. AIDS 1997; 11(Suppl. A):S77–S84.
2. Mabey D, Mayaud P. Sexually transmitted diseases in mobile populations. Genitourin Med 1997; 73:18–22.
3. Quinn TC. Population migration and the spread of types 1 and 2 human immunodeficiency viruses. Proc Natl Acad Sci USA 1994; 91:2407–2414.
4. Pison G, Le Guenno B, Lagarde E, Enel C, Seck C. Seasonal migration: a risk factor for HIV infection in rural Senegal. J Acquir Immune Defic Syndr 1993; 6:196–200.
5. Nunn AJ, Wagner HU, Kamali A, Kengeya-Kayondo JF, Mulder DW. Migration and HIV-1 seroprevalence in a rural Ugandan population. AIDS 1995; 9:503–506.
6. Lagarde E, Pison G, Enel C. A study of sexual behavior change in rural Senegal. J Acquir Immune Defic Syndr Hum Retrovirol 1996; 11:282–287.
7. Barongo LR, Borgdorff MW, Mosha FF, Nicoll A, Grosskurth H, Senkoro KP, et al
. The epidemiology of HIV-1 infection in urban areas, roadside settlements and rural villages in Mwanza region, Tanzania. AIDS 1992; 6:1521–1528.
8. Lydie N, Robinson NJ, Ferry B, Akam E, De Loenzien M, Abega S. Mobility, sexual behavior, and HIV infection in an urban population in Cameroon. J Acquir Immune Defic Syndr 2004; 35:67–74.
9. Lurie M, Harrison A, Wilkinson D, Abdool Karim S. Circular migration and sexual networking in rural KwaZulu/Natal: implications for the spread of HIV and other sexually transmitted diseases. Health Transition Review 1997; 7(Suppl. 3):17–27.
10. Mbizvo MT, Machekano R, McFarland W, Ray S, Bassett M, Latif A, et al
. HIV seroincidence and correlates of seroconversion in a cohort of male factory workers in Harare, Zimbabwe. AIDS 1996; 10:895–901.
11. Lurie M, Williams BG, Zuma K, Mkaya-Mwamburi D, Garnett GP, Sweat MD, et al
. Who infects whom? HIV-1 concordance and discordance among migrant and non-mirant couples in South Africa. AIDS 2003; 17:2245–2252.
12. UNAIDS. Tanzania epidemiological fact sheet on HIV/AIDS and sexually transmitted infections
. Geneva: UNAIDS; 2004.
13. TACAIDS. National multi-sectoral strategic framework on HIV/AIDS 2003–2007
. Dar es Salaam, Tanzania: TACAIDS; 2003.
14. Boerma JT, Urassa M, Senkoro K, Klokke A, Ng'weshemi JZL. Spread of HIV infection in a rural area of Tanzania. AIDS 1999; 13:1233–1240.
15. Mwaluko G, Urassa M, Isingo R, Zaba B, Boerma JT. Trends in HIV and sexual behaviour in a longitudinal study in a rural population in Tanzania, 1994–2000. AIDS 2003; 17:2645–2651.
16. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. BMKA 1986; 73:13–22.
17. Zuma K, Gouws E, Williams B, Lurie M. Risk factors for HIV infection among women in Carletonville, South Africa: migration, demography and sexually transmitted diseases. Int J STD AIDS 2003; 14:814–817.
18. Lurie M, Williams BG, Zuma K, Mkaya-Mwamburi D, Garnett GP, Sturm AW, et al
. The impact of migration on HIV-1 transmission in South-Africa: a study of migrant men and non-migrant men and their partners. Sex Trans Dis 2003; 30:149–156.
19. Ramjee G, Gouws a EE. Prevalence of HIV among truck drivers visiting sex workers in KwaZulu-Natal, South Africa. Sex Transm Dis 2002; 29:44–49.
20. Morris M, Wawer MJ, Makumbi F, Zavisca JR, Sewankambo N. Condom acceptance is higher among travelers in Uganda. AIDS 2000; 14:733–741.
21. Isingo R, Urassa M, Kishamawe C, Knoops E, Voeten H, Mwaluko G, et al
. Mobility is an important risk factor for HIV infection in rural Tanzania. XIV International AIDS Conference
. Barcelona, Spain, July 2002 [abstract ThOrC1485].
22. Robinson ET. Reaching men: at work and in social settings. Network 1991; 12:15–16.
23. Nyamuryekung'e KM, Laukamm-Josten U, Vuylsteke B, Mbuya C, Hamelmann C, Outwater A, et al
. STD services for women at truck stop in Tanzania: evaluation of acceptable approaches. East Afr Med J 1997; 74:343–347.
Keywords:© 2006 Lippincott Williams & Wilkins, Inc.
mobility; sexual behavior; HIV; couples; Tanzania