Vissers, Debby C.J. MSc*; Voeten, Hélène A.C.M. PhD*; Urassa, Mark MSc†; Isingo, Raphael MSc†; Ndege, Milalu†; Kumogola, Yusufu†; Mwaluko, Gabriel PhD‡; Zaba, Basia MSc†§; de Vlas, Sake J. PhD*; Habbema, J Dik F. PhD*
UNSAFE SEXUAL BEHAVIOR INCREASES the risk of getting HIV infected. In couples, absence of 1 of the spouses may lead to more risky sexual behavior in both partners. There are many reasons for absence: work such as seasonal migration or truck driving,1,2 visiting family and relatives, or attending funerals and other ceremonies. Polygamy is also a common reason in the area of Tanzania where this study took place, because the usual arrangement is that a polygamously married man lives alternately with his co-wives.
Most previous studies looking at separation of married partners focused on occupational migration or travel. For example, studies involving work-related migration of the male partner were done in South Africa and Bangladesh. In South Africa, HIV infection of one or both partners occurred twice as often in migrant couples than in nonmigrant couples.3 In Bangladesh, persons who had lived apart reported 2 to 3 times more often that they had extramarital sex than those who had not lived apart.4 In an earlier study in Tanzania, we found that not only mobile persons but also the partners left behind reported more sexual risk behavior.5 In Tanzania and Zimbabwe, women whose partners traveled frequently were more likely to be HIV-positive.6,7
To our knowledge there are only 2 studies that looked into the effect of polygamy, absence of the partner, and the related risk of HIV. In both Kenya and Tanzania, women in polygamous marriages were more likely to report multiple partners.8,9
Risky sexual behavior such as multiple partners and concurrent relationships are associated with an increased risk of HIV infection.10,11 Especially classic sexually transmitted diseases (STDs), which have a short natural history, may be transmitted through concurrent relationships. These STDs enhance the transmission of HIV.12,13 Condoms may decrease the per-contact probability of male-to-female transmission with 95%,14 but are often not used.
In this study, we investigated whether partner absence due to travel or living in separate households leads to more extramarital sex in couples in rural northwestern Tanzania. If so, more information about these specific groups may help to develop new strategies to prevent HIV infection. Implementation of new interventions for couples with an absent partner is feasible, because they are relatively easy to identify. Mobile persons, for example, can be identified along roads, at bus-stops or lodges, and maybe at special ceremonies, like marriages or funerals.
Materials and Methods
Data were collected in an HIV cohort study in the Kisesa ward in the Mwanza region of northwest Tanzania. Kisesa is situated 20 km east of Mwanza city, the regional capital and second largest city in Tanzania. The ward consists of 6 villages and had about 21,000 inhabitants in 2003. The ward includes Kisesa trading center located on the main road from Mwanza to Kenya.
The HIV cohort was established in 1994. Demographic data for all residents have been collected yearly. Every 3 years, adults have been invited to come to serological surveys to provide information about their sexual behavior and give blood for HIV testing.15,16 From July to September 2003, additional data were collected on a purposively selected subsample of the Kisesa cohort. The subsample for this “Couples Apart Study” consisted of couples living apart, and a comparison group of cohabiting couples.
The subsample was based on couples who had been coresident in 1996. In the demographic round in 2002, it was checked whether they still lived together. Couples who no longer lived together were designated as living apart. Couples for comparison were randomly selected from those who were still living together, distributed across the study villages in the same proportion as the original residences of the people living apart. In polygamous marriages, only 1 wife was selected for interview. The members of the couples living apart who had stayed in the marital home were interviewed in 2003. They were asked to provide the name and village of their formerly cohabiting spouse (i.e., partner who moved). Most of them were still living in Kisesa area, but some were living further away. Only those living within 60 km from Kisesa were followed up for interview. If partners who moved were found and agreed to be in the study, they were interviewed. The aim was to interview both partners in couples living apart and in coresident couples, but this was not always possible, because some partners were absent due to work.
Data collected in the Couples Apart Study included age, education, marital status, partner visiting details for those living apart, travel behavior, sexual risk behavior, and self-reported symptoms of STDs. HIV testing was not part of the Couples Apart Study. However, the data were linked to HIV status ascertained in the routine serological surveys within the Kisesa cohort. Since not all respondents in the Couples Apart Study attended the serological surveys, HIV status was only available for 65% of women and 56% of men.
The original aim was to interview 300 couples currently living apart and 300 cohabiting couples. However, fewer people in the group living apart could be interviewed because of difficulties in tracing them or refusal. Persons were only included in the analyses if they were part of a married couple (Fig. 1). Couples in which one or both of the spouses claimed to be an ex-spouse, (ex-) regular partner, or merely a boyfriend/girlfriend at the time of interview were excluded. Some persons classified as living apart according to the demographic round in 2002 actually turned out to have reunited, and were cohabiting with their partner at the time of the interview. These persons were also left out of the analyses. Because women who moved (N = 18) and men who stayed (N = 40) both concerned relatively small numbers, we focused only on men who moved and women who stayed.
Partners living apart were divided into 2 categories based on visiting frequency, in which visits by both the man and the woman were taken into account. Partners living apart who visited each other 2 or more times per week were defined as frequently seeing each other, whereas those who visited each other less than 2 times per week were defined as infrequently seeing each other. This cutoff point of 2 times per week was chosen to obtain groups of approximately equal size. Coresident persons were stratified into mobile and nonmobile. Mobile coresidents were members of coresident couples who slept outside the household more than 10 times in the last year. Nonmobile coresidents slept outside the household at most 10 times in the last year.
Analyses were done for men and women separately. For each category, we calculated proportions of persons per socio-demographic or travel characteristic. To test for differences we used a chi-square test or a Fisher exact test. Statistical significance was conventionally based on P values ≤0.05. We calculated age-adjusted proportions for unprotected extramarital sex in the last year, self-reported STD symptoms in the last year, and HIV prevalence. In polygamous marriages, extramarital sex was sex with a woman other than the wives in the marriage. Unprotected extramarital sex was defined as not using a condom during the last sex act with an extramarital partner. Differences were tested using logistic regression adjusting for age, because age distribution differed significantly between coresidents and people living apart (P <0.001). Finally, we used logistic regression analyses to determine whether category was associated with extramarital sex, adjusting for age, and type of marriage. All analyses were done using Stata version 8.0 (Stata Corporation, College Station, TX).
Our statistical analysis focused on 794 people: 95 men living apart, 85 women living apart, 283 male coresidents, and 331 female coresidents (Fig. 1). Men living apart were significantly older than coresident men (Table 1). The majority of men living apart were polygamously married (92%), whereas most coresident men were monogamously married (87%, P <0.001). Furthermore, men living apart traveled more often within Kisesa ward during their most recent travel than coresident men (38% vs. 21%, P = 0.001). The most common reason for the most recent travel was attending a funeral or other ceremony, e.g., marriage or traditional dancing after harvest season. Men living apart reported visiting a spouse/extramarital partner during their most recent travel more often than coresident men. This was related to men living apart being more often polygamously married. Mobile coresident men were significantly more often polygamously married than nonmobile coresident men (21% vs. 8%, P = 0.001). The destination of their most recent travel was significantly more often in Kisesa ward compared with nonmobile coresident men. There were no differences in socio-demographic and travel characteristics within the subgroups of men living apart.
Women living apart were significantly less educated and more often polygamously married than coresident women (Table 1). The duration of the most recent travel was significantly longer for coresident women than for women living apart. About 60% of the women, both coresident and living apart, traveled to attend a funeral or other ceremony, followed by 30% who traveled to visit relatives. Nonmobile coresident women did not differ from mobile coresident women with regard to socio-demographic and travel characteristics. Within the group of women living apart, those who saw their partner frequently were more often polygamously married than those who saw their partner less than twice a week (92% vs. 67%, P = 0.001).
More than 80% of the men living apart still lived in close proximity of their wife, i.e., within Kisesa ward. In women, the distance between the partners differed between the subgroups: 92% of the women living apart who frequently saw their partners lived close to their husbands, i.e., both in Kisesa ward, compared with 43% of the women living apart who saw their partners infrequently (P <0.001). The main reason for couples to live apart was that the man had a polygamous marriage and was living with another woman. Other reasons were work-related, including farming, or family-related (e.g., taking care of relatives).
Table 2 shows extramarital sex in the last year and STD/HIV status of coresidents and people living apart. After age-adjustment, there were no differences in unprotected extramarital sex, self-reported STD symptoms in the last year or HIV prevalence between coresident men and men living apart. Mobile coresident men reported significantly more extramarital sex in the last year than nonmobile men (35% vs. 15%, P <0.001). However, this difference in risk behavior did not lead to differences in self-reported STD symptoms or HIV prevalence between nonmobile and mobile coresident men. Overall, reported condom use with extramarital sex partners was low. Condom use in the last extramarital sex act varied from 0% in men living apart who frequently saw their spouse to 43% in nonmobile coresident men (data not shown).
In women, the pattern of sexual risk behavior and STDs is not consistent. Women living apart reported significantly more extramarital sex in the last year than coresident women (13% vs. 7%, P = 0.05) (Table 2), but they reported significantly fewer STD symptoms in the last year (7% vs. 17%, P = 0.04). The HIV pattern in women is similar to the pattern of extramarital sex: 10% in women living apart versus 5% in coresident women, although the difference is not significant (P = 0.16). Reported condom use during the last extramarital sex act varied from 20% in women living apart to 15% in coresident women (data not shown). There were no significant differences in sexual behavior or STD/HIV status within the different subgroups of women.
Table 3 gives the results of the multivariate analyses for extramarital sex in the last year. Mobile coresident men were significantly more likely than nonmobile coresident men to have had extramarital sex [odds ratio (OR) 3.7, 95% confidence interval (CI) 2.0–6.7]. The odds of having had extramarital sex was 3.2 times higher in women living apart who saw their partner infrequently than in nonmobile coresident women (95% CI 1.1–9.6). Men living apart in both subgroups and mobile coresident women had higher odds of having had an extramarital sex partner than nonmobile coresidents, but results were not significant. Polygamously married men were less likely to report extramarital sex (OR 0.7, 95% CI 0.3–1.6), whereas women in polygamous marriages were slightly more likely to report extramarital sex (OR 1.2, 95% CI 0.4–3.2).
In this rural area in northwestern Tanzania, most couples living apart were part of polygamous marriages, which led to a relatively low prevalence of reported extramarital sex in men. However, women living apart from their husbands reported more extramarital sex than women who lived with their husbands. More specifically, those living apart and infrequently seeing their spouse were most at risk. Being mobile was a risk factor for unprotected extramarital sex in men, but not in women.
Men and women may have different perceptions about their relationship and may sometimes report differently. Men who see their wives infrequently may report this relationship has ended, whereas the women may still report the marriage as ongoing because they are sometimes visited by their husbands. Women in Tanzania were more likely than men to report a relationship, that started a relatively long time ago, as ongoing.9 Because we were only interested in risky behavior of married couples, we restricted our analyses to couples in which both partners still reported to be married. In this way, dilution of the study effect due to misclassification of married people was not possible.
Furthermore, most polygamous couples were included in the group living apart. However, some polygamous individuals were by chance included as being coresident. Polygamous marriages in Tanzania consist of a man married to 2 or more women. In general, this man will alternately live with 1 of his wives, which means that whenever he is coresiding with one, he is living apart from the other(s). Whether polygamous individuals were classified as coresiding or living apart at the time of the interview was therefore accidental. This probably diluted the effects found in our study.
Men living apart outside Kisesa or even abroad were underrepresented due to tracing difficulties. It is plausible that men living further away will see their wives less frequently. We found that men living apart and seeing their wives infrequently reported considerably more extramarital sex, but results were not significant (OR 2.3, P = 0.16), due to small numbers and because the effect is most likely diluted because our sample did not include enough men living further away.
We only investigated reported extramarital sex in the last year, which has proven to be a good indicator of HIV infection.10,17 Other indicators such as the number of lifetime sex partners or involvement in commercial sex were not asked. However, these indicators are more subject to recall and reporting bias than extramarital sex. Moreover, commercial sex questions are limited to men.
HIV status was obtained by linking with the overall cohort study. Being mobile or absent is a risk factor for HIV infection, but also a reason for not attending HIV surveys. Therefore, we could have underestimated the HIV prevalence in these persons. STD symptoms were only asked in a subgroup who had heard of STDs, and symptoms were self-reported and not confirmed by a physical examination and/or laboratory testing. Women living apart had heard of STDs slightly less often than coresident women (76% vs. 85%, P = 0.05). Less knowledge about STDs and their symptoms may explain the conflicting results in women living apart reporting more extramarital sex, but less STD symptoms in the last year. Moreover, STD symptoms in women can be very asymptomatic and can also be related to other diseases. We would expect a higher STD prevalence in our study because high prevalences of STDs were found in both men and women in the Mwanza trial, which study area is near to Kisesa.18
Women living apart and seeing their husbands infrequently reported more extramarital sex than those who saw their partners frequently. It seems likely that women who are left alone for a longer period are more vulnerable to risky behavior. By seeing their husband less often their sexual needs may not be fulfilled.19,20 Furthermore, there is also less social control from their husbands. Both conditions facilitate women to actively start new sexual relationships. On the other hand, some women may not be supported financially by their absent husbands and may therefore need to engage in sex in exchange for money or food.21 Before effective prevention programs can be developed, more understanding of the determinants of sexual risk behavior of women left behind is needed.
Other studies reported more extramarital sex and more HIV infections in men living apart due to work-related migration.3,4 However, migration for work is not a common reason for living apart in this rural area in Tanzania. We found that being polygamously married was the most common reason for living apart, but for men this was not associated with more risky sexual behavior. Possible explanations are that polygamously married men need all their time and energy for their wives, have less opportunities to start other sexual relationships, and have less unmet sexual needs. Furthermore, most men living apart in our study still lived in close proximity to their wives and were able to see them often. In contrast, migrant men in South Africa and Bangladesh lived relatively far away and were not able to visit their rural homes often.3,4
Our finding that mobile coresident men more often report an extramarital sex partner than nonmobile coresident men is consistent with studies in Cameroon and Senegal. In Cameroon, mobile men were significantly more likely to report having nonspousal partners and one-off contacts than men who were not mobile.22 In Senegal, short-term mobility was associated with having more casual sex partners.23
We conclude that couples who are temporarily separated due to travel or living apart reported more risky behavior and are therefore at increased risk of acquiring an HIV infection. The main reasons for temporary separation from a spouse were attending funerals or other ceremonies, visiting relatives, and in particular being polygamously married. Changing these patterns is neither straightforward nor likely to happen in the near future. Therefore, the best prevention strategy might be to make sexual risk behavior as safe as possible. Because condom use within marriages is unpopular,24,25 condom use with extramarital partners should be emphasized. Individuals most at risk, namely mobile coresident men, and women living apart and infrequently seeing their husbands, which can readily be identified, need extra attention in HIV prevention campaigns.
1. 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.
2. Rakwar J, Lavreys L, Thompson ML, et al. Cofactors for the acquisition of HIV-1 among heterosexual men: Prospective cohort study of trucking company workers in Kenya. AIDS 1999; 13:607–614.
3. Lurie M, Williams BG, Zuma K, et al. Who infects whom? HIV-1 concordance and discordance among migrant and non-mirant couples in South Africa. AIDS 2003; 17:2245–2252.
4. Mercer A, Khanam R, Gurley E, et al. Sexual risk behavior of married men and women in Bangladesh associated with husbands’ work migration and living apart. Sex Transm Dis 2007; 34:265–273.
5. Kishamawe C, Vissers D, Urassa M, et al. Mobility and HIV in Tanzanian couples: Both mobile persons and their partners behind show increased risk. AIDS 2006; 20:601–608.
6. Mbizvo EM, Msuya SE, Stray-Pedersen B, et al. HIV seroprevalence and its associations with the other reproductive tract infections in asymptomatic women in Harare, Zimbabwe. Int J STD AIDS 2001; 12:524–531.
7. Msuya SE, Mbizvo E, Hussain A, et al. HIV among pregnant women in Moshi Tanzania: The role of sexual behavior, male partner characteristics and sexually transmitted infections. AIDS Res Ther 2006; 3:27.
8. Hattori MK, Dodoo FN. Cohabitation, marriage, and ‘sexual monogamy’ in Nairobi's slums. Soc Sci Med 2007; 64:1067–1078.
9. Nnko S, Boerma JT, Urassa M, et al. Secretive females or swaggering males? An assessment of the quality of sexual partnership reporting in rural Tanzania. Soc Sci Med 2004; 59:299–310.
10. Morris M, Kretzschmar M. Concurrent partnerships and the spread of HIV. AIDS 1997; 11:641–648.
11. Halperin DT, Epstein H. Concurrent sexual partnerships help to explain Africa's high HIV prevalence: Implications for prevention. Lancet 2004; 364:4–6.
12. Laga M, Nzila N, Goeman J. The interrelationship of sexually transmitted diseases and HIV infection: Implications for the control of both epidemics in Africa. AIDS 1991; 5:S55–S63.
13. Plummer FA, Simonsen JN, Cameron DW, et al. Cofactors in male-female sexual transmission of human immunodeficiency virus type 1. J Infect Dis 1991; 163:233–239.
14. Pinkerton SD, Abramson PR. Effectiveness of condoms in preventing HIV transmission. Soc Sci Med 1997; 44:1303–1312.
15. Boerma JT, Urassa M, Senkoro K, et al. Spread of HIV infection in a rural area of Tanzania. AIDS 1999; 13:1233–1240.
16. Mwaluko G, Urassa M, Isingo R, et al. Trends in HIV and sexual behaviour in a longitudinal study in a rural population in Tanzania, 1994–2000. AIDS 2003; 17:2645–2651.
17. Slaymaker E. A critique of international indicators of sexual risk behaviour. Sex Transm Infect 2004; 80(suppl 2):ii13–ii21.
18. Orroth KK, Korenromp EL, White RG, et al. Higher risk behaviours and rates of sexually transmitted diseases in Mwanza compared to Uganda may help explain HIV prevention trial outcomes. AIDS 2003; 17:2653–2660.
19. Tawfik L, Watkins SC. Sex in Geneva, sex in Lilongwe, and sex in Balaka. Soc Sci Med 2007; 64:1090–1101.
20. Kesby M. Participatory diagramming as a means to improve communication about sex in rural Zimbabwe: A pilot study. Soc Sci Med 2000; 50:1723–1741.
21. Lurie M, Williams BG, Zuma K, 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 Transm Dis 2003; 30:149–156.
22. Lydie N, Robinson NJ, Ferry B, et al. Mobility, sexual behavior, and HIV infection in an urban population in Cameroon. J Acquir Immune Defic Syndr 2004; 35:67–74.
23. Lagarde E, Schim Van Der Loeff M, Enel C, et al. Mobility and the spread of human immunodeficiency virus into rural areas of West Africa. Int J Epidemiol 2003; 32:744–752.
24. Bond V, Dover P. Men, women and the trouble with condoms: Problems associated with condom use by migrant workers in rural Zambia. Health Transit Rev 1997; 7(suppl):377–391.
25. Chimbiri AM. The condom is an ‘intruder’ in marriage: Evidence from rural Malawi. Soc Sci Med 2007; 64:1102–1115.