Applying the Dynamic Social Systems Model to HIV Prevention in a Rural African Context

Abstract

This study applied the Dynamic Social Systems Model (DSSM) to the issue of HIV risk among the Maasai tribe of Tanzania, using data from a cross-sectional, cluster survey among 370 randomly selected participants from Ngorongoro and Siha Districts. A culturally appropriate survey instrument was developed to explore traditions reportedly coadunate with sexual partnership, including “wife sharing”, fertility rituals, and various traditional dances. One dance, esoto, accounted for more than two thirds of participants’ lifetime sexual partners (n = 10.5). The DSSM, combining structural and systems theories, was applied to systematize complex multilevel factors regarding esoto practice. Participants reported multifaceted beliefs regarding esoto; a majority viewed the dance as exciting and essential, yet most men feared social stigma and three quarters of women had experienced physical punishment for nonattendance. In multivariate logistic regression, esoto attendance was predicted by female gender (adjusted odds ratio [AOR] = 4.67, 95% confidence interval [CI] = 1.6-13.2), higher positive beliefs regarding esoto (AOR = 2.84, 95% CI = 1.9-4.2), and Maasai life cycle events (AOR = 0.06, 95% CI = 0.01-0.47). The DSSM proved useful for characterizing esoto and for revealing feedback loops that maintain esoto, thus indicating avenues for future interventions.

Keywords

behavioral theories HIV/AIDS international health measurement issues sex behavior

Cultural and environmental contexts structure dyadic sexual partnerships. Behavioral interventions to prevent HIV should reflect these influences. On a national level, the Centers for Disease Control and Prevention (2009) has disseminated 26 tailored behavioral HIV prevention programs with proven effectiveness for various at-risk groups, ranging from gay African American men to inner-city girls to Latino drug users to youth recently released from prison. Internationally, numerous efficacious HIV prevention interventions have been developed and implemented in many countries, including Tanzania, the site of the present research. Programs that have been implemented and found efficacious in Tanzania include youth education interventions (Gallant & Maticka-Tyndale, 2004; Hayes et al., 2005; Klepp et al., 1994; Ross et al., 2007), mass media programs (Vaughan, Rogers, Singhal, & Swalehe, 2000), biomedical interventions (Dickson et al., 2011; Kilewo et al., 2008; Kilewo et al., 2009), voluntary counseling and testing as HIV prevention (Sweat et al., 2011; The Voluntary HIV-1 Counseling and Testing Efficacy Study Group, 2000), and interventions for most at-risk populations such as sex workers (Riedner et al., 2006). Some programs have been tailored for particular populations, such as truck drivers (Laukamm-Josten et al., 2000), but there are other groups within Tanzania whose risk profiles have not been fully addressed by these successful HIV prevention programs. One reason is that there is a dearth of quantitative, culture-specific information related to concurrent sexual partnerships among many groups (Leclerc-Madlala, 2009); this is the case for the Maasai of Tanzania.

The Maasai are a seminomadic tribe living in Kenya and Tanzania, with approximately 1,000,000 members (Phillips & Bhavnagri, 2002). Compared with most Swahili-speaking tribes in Tanzania, the Maasai are not well integrated into contemporary Tanzanian customs, language, and social institutions. No proper epidemiological studies of HIV prevalence among the Maasai have been conducted. One clinic-based HIV prevalence estimate was 8% (Sikar & Hodgson, 2006); analysis of a small sample (n = 89) of Kenyan Maasai sampled in a Demographic and Health Survey indicated a prevalence of 4% (Siegler, 2006). There is no comparable information for Tanzanian Maasai; estimated regional HIV prevalence for areas where some Maasai communities reside are 1.6% for Arusha Region and 1.9% for Kilimanjaro Region (Tanzania Commission for AIDS, Zanzibar AIDS Commission, National Bureau of Statistics, Office of the Chief Government Statistician, & Macro International Inc. 2008).

Theoretical Model

The present investigation into Maasai tradition and HIV prevention is structured by the Dynamic Social Systems Model (DSSM; Latkin, Weeks, Glasman, Galletly, & Albarracin, 2010). The DSSM is a structural model focused on behavioral interventions, and is unified by the concept that understanding distal structural factors is necessary to understand circumstances in which individuals can change their behavior. The DSSM builds up Bronfenbrenner’s ecological model, assessing the macro, meso, and micro levels.

What differentiates the DSSM from other structural models is its emphasis on dynamic linkages between broader structural components and the more proximal social factors that are tied to specified health behaviors. These dynamic linkages are emphasized through a systems theory component. In this framework, the ties between elements and structural factors vary in strength, and interact through positive and negative feedback loops. By exploring multiple levels and their interactions, the DSSM seeks to clarify “black boxes” where inputs and outputs are observed, but the intermediate processes are unknown.

HIV-Related Maasai Customs

Qualitative studies have indicated that there are several realms of Maasai custom that can lead to concurrent partnership, including “wife sharing” (Saitoti, 1986; Spencer, 1988; Talle, 1994), the esoto dance (Spencer, 1988; Talle, 2007), and emaho ceremonies (Coast, 2001; Llewellyn-Davies, 1978, 1990; Talle, 2007). “Wife sharing” has been described in the qualitative literature as consisting of unhindered sexual access to age-mates’ wives. We could not find a specific native Maa word, or customary phrase, to describe this custom. The esoto dance has been reported as a common place for warriors (aged 15-30 years) and girls (aged 8-15 years) to find sexual partners. Numerous other ceremonies (emaho) entailing sexual partnership have also been reported, such as oloip (daytime partnerships in the shade), manyata (warrior village), inkipot (gifting milk to selected lover), and olamal (fertility ritual).

Despite providing rich descriptions, qualitative studies on the Maasai have provided a picture that is neither fully coherent nor sufficiently clear in relative risk magnitude to prioritize targets for HIV prevention interventions. One example of disagreement within this literature can be seen in the observations of Jacobs (1979), who noted that “European mythology” has exaggerated descriptions of traditions related to sexual partnership among the Maasai. Sindiga (1987) also argued for this position, using as evidence the relatively low fertility rates of Maasai groups. Moreover, even if the qualitative literature were consistent in identifying traditions related to sexual risk, it could not describe the magnitudes of such risks.

Only two quantitative studies of sexual partners among the Maasai have been conducted; both were small scale and addressed only Maasai males. They reported markedly different findings, echoing the dissension in the qualitative literature; respectively, the studies reported averages of 2 and 12 partners per year (Coast, 2006; Morley, 1991).

This study seeks to fill a gap in the present literature by assessing the intersection of sexual partnership and Maasai ritual by (a) quantitatively identifying overall number of sexual partners and the number attributable to each custom, (b) exploring descriptive statistics and correlates of ritual participation, and (c) incorporating study findings into the rubric of the DSSM.

Method

Overview

This mixed qualitative–quantitative methods study was conducted in areas randomly selected from Siha and Ngorongoro Districts in northern Tanzania. A series of semistructured interviews was conducted with community members in two pilot subvillages, Karansi and Mondorosi. The aim of this data collection was to inform development of a quantitative survey instrument intended to measure culturally situated risks for HIV transmission. The 32 semistructured interviews were conducted with purposive sampling, leading to four in-depth interviews in each of eight categories created by sampling across age (18-30 and 31-50 years), gender, and education levels (no education vs. some education). Domains addressed in the interviews included relationship with main sexual partner, relationships with nonspousal/nonmain partners, HIV/AIDS knowledge and stigma, condom-related beliefs, traditional community celebrations, and cultural influences on sexual identity and partnership norms.

The survey instrument was translated directly from English into Maa by a team of two native Maa speakers, and then back-translated by a second tandem. The face validity of the instrument was improved through comments obtained in 15 cognitive interviews, a survey item analysis methodology suggested by Beatty and Willis (2007) to improve question design, clarity, and cultural compatibility.

Data were obtained from face-to-face interviews conducted by native Maa-speaking, gender-matched interviewers. The sample frame consisted of a set of local leaders (balozi wa nyumba kumi), who represented the smallest administrative unit of government in Tanzania. These units were theoretically sized at 10 households, but in our sample they were generally larger, with a mean of 12.1 households. From 955 units, representing an estimated population of 29,000 adults, 37 were randomly selected. From each of these, 11 households were randomly selected (for areas with <11 households, we sampled nearest neighboring unit), and one adult member was randomly selected from each household.

Measures

Demographic measures

Demographic variables not specific to the Maasai were adapted from the Tanzanian Demographic and Health Survey (National Bureau of Statistics Tanzania & ORC Macro, 2005). These included gender, age, marital status, number of wives, religion, cell phone ownership, radio ownership, education and literacy of any household member. Items were also developed to assess livestock wealth, age set (Maasai male life stages, each approximately 15 years), and tribal section.

HIV-related measures

A scale assessing knowledge of HIV transmission was adapted from the Tanzanian Demographic and Health Survey, consisting of true/false identification of 14 potential pathways for HIV transmission, such as “sharing needles” and “kissing.” A seven-item dichotomously coded stigma scale, adapted from the Tanzania Stigma Indicators survey (UNAIDS, 2005), included items such as “HIV is punishment for bad behavior” and “I would feel ashamed if I were infected with HIV.”

Sexual partnership

Number of lifetime sexual partners by type, a key variable for descriptive analyses, was ascertained using emic partnership categories (Jewkes, Nduna, Jama, Dunkle, & Levin, 2002), a method designed to make partner reporting more accurate. This resulted in seven categories: “premarital boyfriend/girlfriend,” “extramarital lover,” “esoto partner,” “emaho partner,” “partner met at bar,” “shared spouse,” and “sex worker.” These seven categories populated the question, “Did you have vaginal sex with ___ in your lifetime? [IF YES] How many different ___ ?” Emaho, or traditional ceremony, was used to assess partnership stemming from oloip, manyata, inkipot, and olamal; field testing indicated that these practices vary by locale, and emaho provided a universally understood concept in their stead. Formative research for our project found that the esoto dance is particularly important for development of Maasai sexuality, so we developed questions to assess weekly frequency of esoto, esoto as location of first coitus, punishment for esoto nonattendance, and fear of esoto punishment.

Dependent and independent variables

For the logistic regression model, esoto attendance was assessed with the binomial question, “Did you ever attend an esoto that ended with sexual partnership between warriors and girls?” Independent variables for the regression included general and Maasai-specific demographic variables described above. The regression also included a three-item positive esoto beliefs scale we developed. Dichotomous scale items were as follows: “Girls should attend esoto because they need the sperm to develop breasts,” “Attending esoto is an essential part of being a Maasai youth,” and “I was enthusiastic to attend esoto because it meant I was becoming a man [woman].” The average scale interitem correlation was .49. The Kuder–Richardson coefficient, a dichotomous version of Cronbach’s alpha, was .67, an expected low value because shorter scales have lower coefficient alpha values (DeVellis, 2003).

Data Analysis

Data analyses corrected for the hierarchical nature of the data by weighting and statistically adjusting error terms for correlations within housing clusters and natural variation in household sizes. Exploratory analyses were conducted assessing proportions, means, medians, and their error terms. Univariate analyses were conducted to build a regression model assessing determinants of esoto participation. Model-building techniques for the logistic regression that predicted esoto participation were implemented based on methodology suggested by Hosmer and Lemeshow (2000). All analyses were conducted in Stata 11.2 for Mac OS X (StataCorp, 2011). Each participant was consented by a trained interviewer, a process approved by Emory University’s Institutional Review Board and Tanzania’s National Institute for Medical Research.

Application of Theoretical Model

We applied the DSSM framework to the esoto dance to reveal factors that could be leveraged in future HIV prevention interventions. The DSSM seeks to elucidate “black boxes” between model inputs and outputs; in our case, this involved inputs of Maasai culture/governance, Tanzanian governance, and environmental settings. The primary model output was esoto attendance, as it entailed unprotected sex with multiple partners. The DSSM was selected because it is a structural model (requisite for the multilevel nature of esoto practice) specified for behavioral interventions. The full model includes six dimensions that represent possible arenas for theory building, but not all dimensions are considered pertinent for each environment (Latkin et al., 2010). In applying the model to Maasai traditions, we assessed three dimensions: Informal Social Influences, Formal Social Control, and Settings. To explicitly incorporate the impact of traditional Maasai culture, we divided the Formal Social Control category into Maasai Formal Social Control and Governmental Formal Social Control.

Quantitative survey findings were the primary source of information for applying the DSSM, and were supplemented by qualitative observations. During 9 months of fieldwork, numerous conversations were held between the principal investigator and traditional leaders, female heads of households hosting esoto, traditional medical/ritual practitioners, and representatives of various nongovernmental organizations and hospitals working with the Maasai on HIV prevention. Participant observation of the esoto dance (researchers left prior to the sexual portion of the ritual) also informed DSSM development. Field notes from interviews and observations were entered into maxQDA (2012), coded, and analyzed for emergent themes; this information was triangulated, along with quantitative survey data, to explore esoto in the DSSM framework. When possible, field observations were verified through literature, such as information regarding the local Tanzanian governmental system (United Republic of Tanzania President’s Office, 2003) and traditional Maasai leadership structure (Spencer, 1993). Field observations of esoto customs were corroborated through correspondence with Maasai leadership (N. O. Kin’gori, personal communication, October 26, 2009).

Results

Demographics

A sample of 370 individuals, mean reported age 29 years and 53% female, agreed to participate in the survey, a 91% response rate; the sample represented seven Maasai tribal sections (Table 1). The majority (78%) was married, with roughly half in a nonpolygynous marriage, another quarter in a polygynous marriage with one husband and two wives, and the remaining quarter in a polygynous marriage with one husband and three to six wives. Most (59%) of the population held traditional Maasai religious beliefs; a sizable minority subscribed to Lutheran or Catholic beliefs. Educational attainment of this sample was low, with only 3% having attained education beyond the seventh year, and only 39% having ever attended school. Females were more likely than males to have received no education (69% vs. 52%). Around 40% of participants were very poor, owning fewer than 20 cattle and goats.

Table 1.

Demographic Variables

	Total		Female		Male
	n	% [95% CI]	n	% [95% CI]	n	% [95% CI]
General demographic
Gender	370		194		176
Age in years; mean [95% CI]	NA	29 [28, 30]	NA	27.6 [26.4, 28.8]	NA	30.5 [28.5, 32.4]
Age set^a			NA	NA
Ilayiok (Boys)	15	9 [5, 17]			15	9 [5, 17]
Ilmurran (Warriors)	79	44 [35,52]			79	44 [35, 52]
Ilpayani (Junior elders)	59	35 [26, 45]			59	35 [26, 45]
Ilmoruak (Elders)	22	13 [8, 20]			22	13 [8, 20]
Married	291	78 [72, 83]	161	83 [77, 89]	130	72 [62, 81]
Number of wives
1	179	56 [48, 64]	82	44 [36, 53]	97	72 [60, 81]
2	60	24 [18, 31]	43	30 [22, 39]	17	15 [9, 26]
3-6	56	20 [15, 27]	40	26 [19, 35]	16	13 [7, 22]
Religion
Traditional	218	59 [50, 68]	114	60 [50, 70]	104	58 [46, 69]
Lutheran	94	26 [18, 36]	56	28 [19, 40]	38	24 [16, 34]
Catholic	39	10 [7, 15]	16	8 [4, 13]	23	13 [8, 21]
Other	17	5 [3, 7]	8	4 [2, 8]	9	5 [3, 10]
Education
None	223	61 [54, 68]	131	69 [61, 77]	92	52 [43, 61]
Standard 1-7^b	125	36 [29, 43]	50	28 [21, 37]	75	44 [36, 53]
>Standard 7^b	12	3 [2, 5]	5	3 [1, 6]	7	3 [1, 8]
Literate household member	249	69 [63, 75]	101	54 [44, 64]	148	86 [79, 91]
Wealth
Cattle <20	204	51 [40, 61]	105	51 [40, 62]	99	51 [38, 63]
Goats <20	197	52 [41, 63]	109	54 [43, 65]	88	50 [37, 63]
Cell phone	124	36 [29, 43]	63	35 [27, 44]	61	37 [27, 49]
Radio	153	41 [33, 50]	67	33 [23, 46]	86	50 [41, 60]
HIV-related
Heard of HIV	354	99 [97, 100]	186	98 [94, 100]	168	100 [99, 100]
Met someone with HIV	82	25 [18, 34]	50	28 [19, 39]	32	22 [14, 33]
HIV Knowledge Scale^c; mean [95% CI]	NA	9.0 [8.7, 9.3]		8.9 [8.5, 9.2]		9.1 [8.5, 9.7]
PLHIV Stigma Scale^c; mean [95% CI]	NA	5.2 [5.0, 5.4]		4.9 [4.5, 5.4]		5.5 [5.3, 5.8]
Ever condom use	7	3 [1, 6]	3	2 [1, 7]	4	4 [1, 10]
Willing to try condoms	107	31 [22, 42]	39	24 [15, 36]	68	40 [28, 53]

Note. CI = confidence interval; NA = not applicable; PLHIV = people living with HIV.

Age sets are Maasai male life stages, each stage = 15 years. ^bStandards in Tanzanian education approximate U.S. grades. ^cScales total 14 and 7 points for knowledge and stigma, respectively, with higher values indicating greater levels.

HIV Knowledge and Condom Use

Nearly all respondents (99%) were aware of HIV, yet many had a limited understanding of HIV transmission. Transmission pathways of blood, vaginal sex, and injections were each identified correctly by more than 90% of the population. Yet false pathways were also posited, with kissing, mosquitoes, and saliva each incorrectly identified by more than 40% as sources of HIV transmission. This resulted in low scores on the HIV knowledge scale. Stigma toward people living with HIV was common (mean 5.2 out of 7-point scale), although only 25% of respondents reported ever meeting someone they believed to have HIV. Regarding prophylactics, less than 3% ever used a condom, yet nearly one third said they would be willing to use condoms if available.

Premarital Sexual Partnership and Maasai Traditions

Table 2 displays mean lifetime nonmarital sexual partners by partnership category. With a mean of 10.2 lifetime partners in all categories, 6.7 came from traditional sources, with 6.3 from esoto, 0.4 from emaho, and <0.1 from wife sharing. Partners not encountered at Maasai traditional venues accounted for a mean of 3.5 total partners, with 2.4 from premarital boyfriends/girlfriends, 0.9 from long-term extramarital partnerships, and 0.2 from all other sources. The esoto dance accounted for more than twice the partnerships of all other modes of nonmarital lifetime partners. Among those who attended esoto, a range of 0 to 280 partners was reported, with a mean of 10.5 and a median of 4 esoto-specific partners. Outliers were included in our calculations, as reports of large partner numbers were consistent with field reports of popular males termed dance heroes. Males reported significantly more sexual partners than females did for the categories of esoto, emaho, and premarital boy/girlfriends. In survey studies across Africa, males nearly always reported significantly more sexual partners than females (Clark, Kabiru, & Zulu, 2011; Glynn et al., 2011; Nnko, Boerma, Urassa, Mwaluko, & Zaba, 2004). Despite different absolute levels of sexual partners reported, patterns within gender were similar; esoto accounted for the majority of lifetime partners for both males and females.

Table 2.

Means of Lifetime Sexual Partners in Emic Maasai Categories^a

	All	Males	Females	Esoto Attenders	Nonattenders
Maasai specific
Esoto	6.27 (1.9)	9.68 (3.9)	3.26 (0.92)	10.54 (3.2)	0 (NA)
Emaho	0.39 (0.13)	0.77 (0.26)	0.06 (0.02)	0.61 (0.21)	0.07 (0.02)
Spouse sharing	0.06 (0.02)	0.11 (0.05)	0.01 (0.01)	0.08 (0.04)	0.03 (0.02)
Non-Maasai specific
Premarital boyfriend/girlfriend	2.35 (0.41)	4.86 (0.83)	0.15 (0.06)	3.17 (0.60)	1.16 (0.23)
Postmarital boyfriend/girlfriend	0.91 (0.14)	1.20 (0.31)	0.70 (0.14)	1.16 (0.19)	0.31 (0.06)
Sex worker	0.03 (0.02)	0.04 (0.03)	0.01 (0.01)	0 (NA)	0.06 (0.04)
Partner met at bar	0.09 (0.03)	0.12 (0.05)	0.06 (0.03)	0.10 (0.04)	0.07 (0.03)
Other partner	0.14 (0.04)	0.02 (0.01)	0.24 (0.07)	0.14 (0.04)	0.14 (0.06)
Total	10.2 (2.2)	16.45 (4.5)	4.38 (0.92)	15.8 (3.6)	1.8 (0.2)

Note. All values are mean (SE).

Esoto Experience

Interviews indicated that esoto was closely linked to sex, with 80% of those surveyed reporting that the dance always ended in sexual partnership (Table 3). The esoto dance was attended by 61% of the sample; this number increased to 82% when the Ilarussa tribal section was excluded. All 93 members of the Ilarussa section reported never attending esoto and were thus excluded from all subsequent analyses regarding esoto. In follow-up interviews with Ilarussa elders, they could not recall a time when esoto was practiced in their section.

Table 3.

Esoto Dance Attendees’ Beliefs and Sexual Experiences Regarding Esoto by Gender

	Total; n	Total Agreement; % [95% CI]	Male; n	Male Agreement; % [95% CI]	Female; n	Female Agreement; % [95% CI]
Sex occurs at every esoto	145	80 [76, 84]	36	59 [49, 68]^a	109	91 [81, 96]
First coitus occurred at esoto	154	83 [76, 89]	42	66 [49, 79]^a	112	92 [80, 97]
Number of days per week^b esoto is held; mean [95% CI]	NA	6.0 [5.5, 6.4]	NA	6.6 [6.3, 6.9]^a	NA	5.7 [5.1, 6.2]
Number of days per week^b esoto is attended; mean [95% CI]	NA	4.5 [4.2, 4.8]	NA	4.8 [4.2, 5.5]	NA	4.3 [3.9, 4.7]
Felt enthusiasm in attending esoto for the first time	193	88 [81, 93]	83	88 [80, 93]	110	88 [77, 94]
Attending esoto is essential to being Maasai	199	89 [82, 93]	87	87 [77, 93]	112	90 [80, 95]
Semen provided at esoto is necessary for female breast development	185	85 [78, 90]	76	77 [67, 85]	109	90 [81, 95]
Fear social stigma if not attending esoto	178	82 [74, 88]	75	79 [69, 86]	103	84 [74, 90]
Ever beaten for not attending esoto	95	75 [61, 86]	NA^c	NA^c	95	75 [61, 86]
Would attend esoto less if there was no punishment for nonattendance	99	80 [64, 90]	NA^c	NA^c	99	80 [64, 90]

Note. CI = confidence interval; NA = not applicable.

Significant difference between males and females at the p = .05 level. ^bEsoto frequencies are for times of the year when the dance is held, during the rainy seasons lasting 4 to 6 months. ^cMale warriors may beat girls for not attending esoto, but are not themselves subject to such punishment.

Esoto was a common locale for first coitus (92% females, 66% males). Support for the esoto tradition in the abstract was strong, yet participants had mixed feelings about their own experiences. Agreement ranged from 85% to 89% for statements in the positive esoto beliefs scale, including, “I was enthusiastic to attend esoto,” “esoto is essential to being Maasai,” and “semen is necessary for girls to develop breasts.” Despite agreement with these abstract statements, 80% of males and females feared the social consequences of nonattendance. In addition to potential social ramifications, females faced physical ramifications; three quarters of females reported experiencing physical punishment, by beating with a stick, for nonattendance. Moreover, 80% of females reported that they would have attended esoto less if there were no threat of physical punishment.

Correlates of Esoto Participation

A multivariate logistic analysis predicting ever attending esoto showed that positive beliefs regarding esoto, female gender, and adult life cycle rituals were significant positive predictors (Table 4). The overall model was highly significant, with the global F test (F = 8.7e32, prob > F = .000) indicating good model fit. For each point higher score on the positive esoto beliefs scale, the likelihood of a community member ever attending esoto was 2.84 times higher (adjusted odds ratio [AOR] 2.84, 95% confidence interval [CI] = 1.9-4.2). Average female likelihood of attending esoto was more than 4 times higher than males (AOR = 4.67, 95% CI = 1.6-13.2). Maasai who had not undergone adult life cycle rituals, either male circumcision or female excision, had a 94% lower chance of attending esoto (AOR = 0.06, 95% CI = 0.01-0.47). The final regression model had no significant interactions among independent variables.

Table 4.

Logistic Regressions Predicting Attendance of Esoto Dance With Maasai Sociodemographic and Esoto Factors

	Unadjusted OR [95% CI]	Adjusted OR [95% CI]
Esoto Positive Beliefs Scale	2.73 [1.9, 4.0]	2.84 [1.9, 4.2]
Female gender	2.92 [1.1, 8.0]	4.67 [1.6, 13.2]
Has not received adult genital cutting ritual	0.07 [0.01, 0.32]	0.06 [0.01, 0.47]
Any education	0.25 [0.11, 0.56]	NA^a

Note. OR = odds ratio; CI = confidence interval; NA = not applicable.

Education is fully mediated by the Esoto Positive Beliefs Scale.

Although male circumcision and female excision was a significant predictor of esoto attendance, different causes seemed to underlie why males had not been circumcised and females not excised. For males, most (15/18) were ineligible for circumcision by Maasai ritual rules because of their age-group (boy). All females were of sufficient age to be eligible for excision (experience of menarche). Predictors of women’s avoidance of excision included higher levels of any education (65% vs. 24%) and Christianity (65% vs. 34%) than their peers.

Education’s impact on esoto attendance was fully mediated by the positive esoto beliefs scale, based on Baron and Kenny’s (1986) criteria. We hypothesized that the relation between education (independent variable) and esoto attendance (dependent variable) was mediated by the positive esoto beliefs scale (MED), and tested this with a single mediation model. The Sobel value was −2.51 (p = .01), indicating a significant mediator effect. The indirect, mediated, effect accounted for 40% of the total effect, suggesting that positive esoto beliefs accounted for 40% of education’s effect on esoto attendance.

Dynamic Social Systems Model of the Esoto Dance

The DSSM application to the esoto dance can be seen in Table 5, exploring the esoto system by detailing macro, meso, and micro levels of factors that eventuate a positive, stabilized feedback loop in which esoto and attendant unprotected sex is practiced by most (82%) of the eligible population. This section details the esoto feedback loop, and potential areas for disruptive negative feedback, by addressing informal social influences, Maasai formal social control mechanisms, governmental formal control mechanisms and settings.

Table 5.

Dynamic Social Systems Model of the Esoto Dance

	Macro	Meso	Micro
Informal social influences	Esoto is normative location of first coitus (92% females, 66% males), normative source of multiple sexual partners (mean = 10.5, median = 4). Customary social stigma for nonattendance (82% fear stigma) and nearly universal sex without condoms (3% ever condom use)	Population is aware of HIV (99%), yet only one fourth believe they have met someone with HIV. HIV knowledge is low; more than 40% believe mosquitoes and kissing transmit HIV. Stigma toward people living with HIV is high (mean = 5.2 on a 7-point scale)	Those with higher scores on the positive esoto beliefs scale are more likely to participate (adjusted odds ratio = 2.84) (i.e., believing esoto is essential to being Maasai, being enthusiastic to attend, or believing girls need sperm to develop)
Maasai formal social control mechanisms	Normative beatings for nonattendance (75% women beaten). The Ilarussa section does not practice esoto. Influential cultural leaders include age-group heads (Warrior, Junior Elder, Elder), traditional healers and circumcisers, and the Maasai Council	Frequent esoto dance availability (mean = 6 days per week). Customs of hosting household, local leaders and parents influence participants. Implicit local sanctioning of beatings allows practice to continue	Females are more likely to participate than males (adjusted odds ratio = 4.67). Individuals enact norms: Esoto “heroes” have many sexual partners. Visiting warriors access nonlocal esoto
Governmental formal social control mechanisms	Low reach of education programs (61% no education, 97% less than eighth grade education), government leaders and medical officers influence local life at district, village and subvillage levels. Low reach of national prevention programs (i.e., condom social marketing and VCT)	Tanzanian legal framework establishes sex with minors as illegal (female esoto participants are all prepubescent). Moreover, all physical coercion is illegal under Tanzanian statutes	Those with any education are less likely to participate (odds ratio = 0.25); education is mediated by Esoto Positive Beliefs Scale in multivariate model. Local formal mechanisms and political will for law enforcement are limited
Settings	Nearly half the population is very poor (51% own less than 20 cattle). Seminomadic population is influenced by droughts, seasonality, and political allocations for cattle herding	Availability of esoto sites in walking distance influences attendance, seasonality (dry vs. rainy) influences migration of cattle, and thus the presence of men who herd	Those who have not undergone adult rituals (female excision or male circumcision) are less likely (adjusted odds ratio = 0.06) to have participated in esoto

Source. All quantitative findings are from survey data detailed in the present study. Sources on the Tanzanian governmental system (United Republic of Tanzania President’s Office, 2003) and traditional Maasai leadership structure (Spencer, 1993) informed the model. Description of customs surrounding esoto practice is based on field observations by the principal investigator, which were subsequently verified through correspondence with Maasai leadership (Kin’gori, personal communication, October 26, 2009).

Informal social influences serve as key reinforcing mechanisms for esoto practice. Entrenched as the normative source of first coitus, esoto is further bulwarked by norms of attendance, with most Maasai fearing stigma for lapses in attendance. HIV risk is standard when attending esoto, with multiple sexual partners and unprotected sex being the current practice. Those who equate the importance of esoto to Maasai identity, or who subscribe to the traditional belief that women need sperm to develop breasts, are most likely to participate.

Maasai formal social control mechanisms reinforce esoto practice, yet simultaneously provide potential inroads for destabilizing negative feedback to be added. Physical coercion of females by warriors is normative. Without explicitly endorsing this, Maasai tradition and current leadership implicitly sanction the beatings, as warriors face no repercussions. Females, unable to opt out of esoto, are more likely to participate than males.

Esoto is conducted at local host compounds, generally in groups of 15 to 30 participants. The “hosting” family controls the environment in numerous ways; for instance, we observed one host sending girls home whom they deemed too young to participate. Maasai formal social control mechanisms could provide inroads for leveraging change; at the meso level, hosting households are influential, whereas at the macro level, traditional leaders such as age-group heads and traditional healers can instigate change.

Governmental formal control mechanisms have the potential to disrupt the esoto feedback loop. Educated participants were less likely to participate in esoto; basic investment in early education could yield a multitude of benefits. Expanding the government’s legal presence could also provide utility. Both sex with minors (girls attending are prepubescent) and physical coercion are illegal in Tanzania (Interpol, 2003), yet no enforcement mechanisms exist to address known Maasai practices.

Settings frame the esoto dance, proscribing both limitations on its practice and on any programs seeking to alter it. Esoto is practiced only during the rainy seasons, as warriors periodically leave during the dry season in search of water. Cattle herding also influences population density, with Maasai populations and local leadership being relatively dispersed based on herding needs. Because of low population density, transportation is limited. Few roads, and low road quality, are the norm. Any programs working with the Maasai must overcome transport, geographical and seasonal challenges to gain sufficient intervention penetration away from population centers.

One method of applying the DSSM involves identifying systems-based feedback loops across structural levels, with interventions designed to alter ties or build new disruptive loop elements. For instance, one positive reinforcing feedback loop involves esoto norms and physical beatings that coerce female attendance. The loop starts with individual esoto participation (micro), which strengthens esoto as normative Maasai rite of life passage (macro), contributing to stigma surrounding nonattendance (macro), which manifests in values and customs that sanction physical beatings (macro), normatively enacted at host households (meso), leading to greater likelihood of individual esoto attendance (micro). This loop may be susceptible to macrolevel interventions that develop structural barriers to prevent beatings, such as enforcement of governmental or Maasai regulations. Alternately, meso-focused interventions could involve parents, host households, and local community leaders in reaching out to warriors to change behaviors surrounding nonattendance.

A different application of the DSSM could seek to increase connections between informal norms at the macro level, such as norms of sex without condoms, and formal Maasai social control mechanisms that currently do not engage these norms, such as Maasai age-group leaders. This would be similar to a popular opinion leader approach (Kelly et al., 1991), adapted to a Maasai context. Future structural interventions addressing esoto will likely be most effective if they involve direct collaboration with Maasai leadership.

Discussion

At a proximal level, this study has important implications for the design of future HIV prevention programs for the Maasai. In meeting the first study objective, to quantify the number of sexual partners attributable to each emic category, we found that the esoto ritual was responsible for the vast majority of premarital partners for attendees. Given the high numbers of partners and the concurrent, overlapping nature of partnership that the dance entails, esoto is likely a principal source of HIV risk among the Maasai.

The second study objective, to describe characteristics and correlates of ritual participation, revealed a complex picture. Esoto was ubiquitous in its practice, and sex was strongly tied to the ritual. Participants held positive beliefs about it, but most feared stigma for nonattendance, and most females would have attended less in the absence of physical coercion. Although esoto was highly attended, the 18% of nonattenders (among practicing Maasai sections) were more likely to be male, have higher education, hold more negative beliefs toward esoto, and not have experienced male circumcision or female excision.

The third study objective, to incorporate the findings into the rubric of the DSSM, yielded several implications for future interventions seeking to address HIV-risk entailed by esoto. The DSSM does not indicate any single intervention, but instead provides a flexible framework for potential interventions. In designing an intervention to address esoto, program designers must determine whether to take an approach that aims to (a) eliminate esoto or (b) reduce potential hazards and maximize the right of self-determination of participants.

The current practice of esoto constitutes two areas of human rights violations: Female participants are too young to provide consent and are frequently physically coerced to participate. A programmatic approach seeking to eliminate esoto is attractive, in that it provides a clear ethical stance on these issues. This strategy could be paralleled to the current World Health Organization strategy regarding female genital excision, which solely emphasizes eradication and explicitly discourages harm-reduction approaches (UNAIDS et al., 2010). Yet female excision is different from esoto in that, no matter how it is adapted, it fundamentally violates human rights. Esoto is more nuanced; if age of participating girls was increased, and coercion was ended, esoto would be no more a human rights violation than some nightclubs in the West.

The DSSM framework indicates that esoto practice is deeply embedded in Maasai culture, having strong support from both elders and performing youth. Thus, a harm-reduction approach may be more likely to succeed than an elimination approach. Harm-reduction approaches have been applied to various groups, including injection drug users (Sterk, Theall, Elifson, & Kidder, 2003), commercial sex workers (Rekart, 2005), and “rave dance” (all-night parties frequently entailing Ecstasy drug use) attenders (Weir, 2000).

DSSM findings suggest several underlying factors that should be addressed in any harm-reduction approach to interventions regarding esoto. First, Maasai institutions at multiple levels should be engaged in developing and implementing interventions. This translates to collaborating with the Maasai Council of Tanzania and age-group leaders (ilaiguenak) in the project area (macro level), working directly with hosting households (meso level), and developing projects that communicate directly to warriors and girls practicing esoto (micro level).

Second, DSSM analysis indicates the centrality of physical coercion of females in perpetuating esoto and esoto values, and also shows pathways to end or limit this custom. Preventing physical coercion should start with interventions targeting changes in the rules established by Maasai leadership. Following a rule change, campaigns to prevent coercion should establish behavioral change programs targeting host households, parents of young girls, and younger age-group leaders who hold influence over their peers. Allowing girls to opt out of esoto will not only affect those girls but also the entire cultural framework supporting esoto.

A complementary harm-reduction strategy suggested by the DSSM would involve increasing the age at which girls are eligible to participate. In this scenario, rather than pregnancy being prevented by the young age of girls, condoms would be the factor that prevents pregnancy, and moreover would provide protection against HIV and other sexual transmitted infections. This strategy would likely be welcomed by warriors, as warriors prefer more physically developed sexual partners (Talle, 2007). Hosting households already assist in the selection of girls deemed too young for esoto, so they could be key stakeholders in implementing this change.

Navigating age eligibility changes might be difficult due to potential opposition from older age-sets, who traditionally have held marital rights to girls postmenarche. Yet age-group roles have begun to change, so this scenario is a possibility. For instance, warriors are traditionally strictly forbidden to have sex with women (Talle, 2007), yet the principal investigator attended a marriage ceremony between a warrior and a woman in rural Loliondo village. Maasai traditions are evolving, and the DSSM provides a flexible framework within which to approach Maasai leadership and negotiate a harm-reducing HIV prevention strategy for esoto that in turn protects human rights.

At a more distal level, this study demonstrates the utility of combining and tailoring qualitative and quantitative approaches to understand HIV-related practices in underserved rural areas. Given solely quantitative information, such as from a general knowledge, attitudes, and practices survey, the source of high levels of partnership and surrounding norms would have been obscured. Given solely qualitative information, the magnitude of different potential sources of risk would be unknown, making it nearly impossible to design targeted programs. By establishing a qualitative foundation that informs quantitative assessment of magnitude, future behavioral HIV prevention studies can avoid these pitfalls.

The DSSM provides a flexible framework that can serve as a tool for community-based, structural interventions. By overtly characterizing structural factors, multifaceted interventions can be designed to alter connection patterns based on a systems approach. It may be possible to leverage DSSM flexibility by using it as a tool to engage communities in developing locally appropriate structural interventions. This study is the first we know of that applies the DSSM framework to developing country settings. Further exploration of model implications for program planning and community participation is needed.

The present study has several limitations. One is that face-to-face interviews tend to increase levels of socially desirable reporting, compounding potentially unbiased recall error. Another is measurement error in translating a survey across languages and cultures, a process that is fraught with difficulty despite efforts to ameliorate this problem through reverse-translation and cognitive interviewing. Another limitation is that the study was exploratory and did not fully address the impact of seasonality and migration patterns on sexual networks during migratory periods; future studies should be conducted to more fully explore this area.

This study also has several strengths. It is the first study to quantitatively measure the intersection of ritual and HIV-related behaviors among the Maasai. It also provides a novel example of the utility of developing locally appropriate quantitative survey items based on qualitative work. Finally, the present study provides a foundation for future research and interventions aimed at preventing HIV transmission among the Maasai, while simultaneously highlighting areas for future research into application of the DSSM.

Footnotes

Acknowledgements

We would like to thank Mr. Robert Porokwa, Ms. Naini Leeshwell, Mr. Moses Ndiyaine, and Ms. Elivester Wilson for their work in data collection.

Authors’ Note

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Mental Health or the National Institutes of Health.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article:

This study was supported by Grant Number F31MH082647 from the National Institute for Mental Health and was facilitated by the Center for AIDS Research at Emory University (P30AI050409).

References

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. doi:10.1037/0022-3514.51.6.1173

Beatty

P. C.

Willis

G. B.

(2007). Research synthesis: The practice of cognitive interviewing. Public Opinion Quarterly, 71, 287-311. doi:10.1093/poq/nfm006

Centers for Disease Control and Prevention. (2009). DEBI: Diffussion of Effective Behavioral Interventions. Retrieved from http://www.effectiveinterventions.org/en/HighImpactPrevention/Interventions.aspx

Clark

Kabiru

Zulu

(2011). Do men and women report their sexual partnerships differently? Evidence from Kisumu, Kenya. International Perspectives on Sexual and Reproductive Health, 37, 181-190. doi:10.1363/3718111

Coast

(2001). Maasai demography. (Doctoral dissertation, University of London). Retrieved from http://eprints.lse.ac.uk/archive/00000264/

Coast

(2006). Local understandings of, and responses to, HIV: Rural-urban migrants in Tanzania. Social Sciene & Medicine, 63, 1000-1010.

DeVellis

R. F.

(2003). Scale development: Theory and applications (2nd ed.). Thousand Oaks, CA: Sage.

Dickson

K. E.

Tran

N. T.

Samuelson

J. L.

Njeuhmeli

Cherutich

Dick

Hankins

C. A.

(2011). Voluntary medical male circumcision: A framework analysis of policy and program implementation in eastern and southern Africa. PLoS Medicine, 8(11), e1001133. doi:10.1371/journal.pmed.1001133

Gallant

Maticka-Tyndale

(2004). School-based HIV prevention programmes for African youth. Social Science & Medicine, 58, 1337-1351. doi:10.1016/s0277-9536(03)00331-9

10.

Glynn

J. R.

Kayuni

Banda

Parrott

Floyd

Francis-Chizororo

French

(2011). Assessing the validity of sexual behaviour reports in a whole population survey in rural Malawi. PLoS One, 6(7), e22840. doi:10.1371/journal.pone.0022840

11.

Hayes

R. J.

Changalucha

Ross

D. A.

Gavyole

Todd

Obasi

A. I.

Grosskurth

(2005). The MEMA kwa Vijana project: Design of a community randomised trial of an innovative adolescent sexual health intervention in rural Tanzania. Contemporary Clinical Trials, 26, 430-442. doi:10.1016/j.cct.2005.04.006

12.

Hosmer

D. W.

Lemeshow

(2000). Applied logistic regression (2nd ed.). Hoboken, NJ: Wiley.

13.

Interpol. (2003). Legislation of Interpol member states on sexual offences against children: Tanzania. Retrieved from http://www.interpol.int/public/Children/SexualAbuse/NationalLawsold/csaTanzania.asp

14.

Jacobs

(1979). Maasai inter-tribal relations; belligerent herdsmen or peaceable pastoralists? Senri Ethnological Studies, 3, 33-52.

15.

Jewkes

Nduna

Jama

Dunkle

Levin

(2002, July). Steadys, roll-ons and hit and runs: Using indigenous typology to measure number of sexual partners. Paper presented at the 14th International Conference on AIDS, Barcelona, Spain.

16.

Kelly

J. A.

St. Lawrence

J. S.

Diaz

Y. E.

Stevenson

L. Y.

Hauth

A. C.

Brasfield

T. L.

Andrew

M. E.

(1991). HIV risk behavior reduction following intervention with key opinion leaders of population: An experimental analysis. American Journal of Public Health, 81, 168-171.

17.

Kilewo

Karlsson

Massawe

Lyamuya

Swai

Mhalu

Biberfeld

(2008). Prevention of mother-to-child transmission of HIV-1 through breast-feeding by treating infants prophylactically with lamivudine in Dar es Salaam, Tanzania: The Mitra study. Journal of Acquired Immune Deficiency Syndromes, 48, 315-323. doi:10.1097/QAI.0b013e31816e395c

18.

Kilewo

Karlsson

Ngarina

Massawe

Lyamuya

Swai

Biberfeld

(2009). Prevention of mother-to-child transmission of HIV-1 through breastfeeding by treating mothers with triple antiretroviral therapy in Dar es Salaam, Tanzania: The Mitra Plus study. Journal of Acquired Immune Deficiency Syndromes, 52, 406-416. doi:10.1097/QAI.0b013e3181b323ff

19.

Klepp

K. I.

Ndeki

S. S.

Seha

A. M.

Hannan

Lyimo

B. A.

Msuya

M. H.

Schreiner

(1994). AIDS education for primary school children in Tanzania: An evaluation study. AIDS, 8, 1157-1162.

20.

Latkin

Weeks

M. R.

Glasman

Galletly

Albarracin

(2010). A dynamic social systems model for considering structural factors in HIV prevention and detection. AIDS and Behavior, 14(Suppl. 2), 222-238. doi:10.1007/s10461-010-9804-y

21.

Laukamm-Josten

Mwizarubi

B. K.

Outwater

Mwaijonga

C. L.

Valadez

J. J.

Nyamwaya

Nyamuryekung’e

(2000). Preventing HIV infection through peer education and condom promotion among truck drivers and their sexual partners in Tanzania, 1990-1993. AIDS Care, 12, 27-40. doi:10.1080/09540120047440

22.

Leclerc-Madlala

(2009). Cultural scripts for multiple and concurrent partnerships in southern Africa: Why HIV prevention needs anthropology. Sexual Health, 6, 103-110. doi:10.1071/SH08032.

23.

Llewellyn-Davies

(1978). Two contexts of solidarity among pastoral Maasai women. In Caplan

Bujra

J. M.

(Eds.), Women united, women divided: Cross-cultural perspectives on female solidarity (pp. 206-238). London, England: Tavistock.

24.

Llewellyn-Davies

(Writer). (1990). The women’s Olamal: The organization of a Maasai Fertility Ceremony [Videocassette].

25.

maxQDA. (2012). Software for qualitative data analysis (Version 10). Berlin, Germany: VERBI Software.

26.

Morley

(1991). Kenya: Maasai warriors and their sexual partners. Lancet, 337(8742), 667-668. doi:10.1016/0140-6736(91)92471-D

27.

National Bureau of Statistics Tanzania & ORC Macro. (2005). Tanzania Demographic and Health Survey 2004-05. Dar es Salaam, Tanzania: Author.

28.

Nnko

Boerma

J. T.

Urassa

Mwaluko

Zaba

(2004). Secretive females or swaggering males? An assessment of the quality of sexual partnership reporting in rural Tanzania. Social Science & Medicine, 59, 299-310. doi:10.1016/j.socscimed.2003.10.031

29.

Phillips

Bhavnagri

(2002). The Maasai’s education and empowerment: Challenges of a migrant lifestyle. Childhood Education, 78, 140-146.

30.

Rekart

M. L.

(2005). Sex-work harm reduction. Lancet, 366(9503), 2123-2134. doi:10.1016/s0140-6736(05)67732-x

31.

Riedner

Hoffmann

Rusizoka

Mmbando

Maboko

Grosskurth

Hoelscher

(2006). Decline in sexually transmitted infection prevalence and HIV incidence in female barworkers attending prevention and care services in Mbeya Region, Tanzania. AIDS, 20, 609-615. doi:10.1097/01.aids.0000210616.90954.47

32.

Ross

D. A.

Changalucha

Obasi

A. I.

Todd

Plummer

M. L.

Cleophas-Mazige

Hayes

R. J.

(2007). Biological and behavioural impact of an adolescent sexual health intervention in Tanzania: A community-randomized trial. AIDS, 21, 1943-1955. doi:10.1097/QAD.0b013e3282ed3cf5

33.

Saitoti

T. O.

(1986). The worlds of a Maasai warrior: An autobiography. New York, NY: Random House.

34.

Siegler

A. J.

(2006). HIV prevalence and behavioral measures among Maasai of Kenya: Analysis of DHS data. Unpublished raw data. Emory University, Atlanta, GA.

35.

Sikar

N. K.

Hodgson

D. L.

(2006). In the shadow of the MDGs: Pastoralist women and children in Tanzania. Indigenous Affairs, 2006(1), 30-37.

36.

Sindiga

(1987). Fertility control and population growth among the Maasai. Human Ecology, 15, 53-66.

37.

Spencer

(1988). The Maasai of Matapato: A study of rituals of rebellion. Manchester, England: Manchester University Press.

38.

Spencer

(1993). Becoming Maasi: Being in time. In Spear

Waller

(Eds.), Being Maasai: Ethnicity and identity in East Africa (pp. 140-156). Athens, OH: Ohio University Press.

39.

StataCorp. (2011). Stata Statistical Software: Release 11. College Station, TX: Author.

40.

Sterk

C. E.

Theall

K. P.

Elifson

K. W.

Kidder

(2003). HIV risk reduction among African-American women who inject drugs: A randomized controlled trial. AIDS and Behavior, 7, 73-86.

41.

Sweat

Morin

Celentano

Mulawa

Singh

Mbwambo

Coates

(2011). Community-based intervention to increase HIV testing and case detection in people aged 16-32 years in Tanzania, Zimbabwe, and Thailand (NIMH Project Accept, HPTN 043): A randomised study. Lancet Infect Diseases, 11, 525-532. doi:10.1016/s1473-3099(11)70060-3

42.

Talle

(1994). The making of female fertility. Anthropological perspectives on a bodily issue. Acta Obstetrica et Gynecologica Scandinavica, 73, 280-283.

43.

Talle

(2007). ‘Serious games’: Licenses and prohibitions in Maasai sexual life. Africa, 77, 351-370.

44.

Tanzania Commission for AIDS, Zanzibar AIDS Commission, National Bureau of Statistics, Office of the Chief Government Statistician, & Macro International Inc. (2008). Tanzania HIV/AIDS and Malaria Indicator Survey 2007-08. Dar es Salaam, Tanzania: Author.

45.

UNAIDS. (2005). Working report measuring HIV stigma: Results of a field test in Tanzania. Washington, DC: USAID. Retrieved from http://www.popcouncil.org/Horizons/ORToolkit/AIDSQuest/instruments/stigindicators.pdf

46.

UNAIDS, UNDP, UNFPA, UNHCR, UNICEF, UNIFEM, . . . WMA. (2010). Global strategy to stop health-care providers from performing female genital mutilation. Geneva, Switzerland: World Health Organization. Retrieved from http://whqlibdoc.who.int/hq/2010/WHO_RHR_10.9_eng.pdf

47.

United Republic of Tanzania President’s Office. (2003). The Association of Local Authorities of Tanzania: RALG-C/L.50. (RALG-C/L.50). Dodoma, Tanzania.

48.

Vaughan

P. W.

Rogers

E. M.

Singhal

Swalehe

R. M.

(2000). Entertainment-education and HIV/AIDS prevention: A field experiment in Tanzania. Journal of Health Communication, 5(Suppl.), 81-100. doi:10.1080/10810730050019573

49.

The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. (2000). Efficacy of voluntary HIV-1 counselling and testing in individuals and couples in Kenya, Tanzania, and Trinidad: A randomised trial. Lancet, 356(9224), 103-112.

50.

Weir

(2000). Raves: A review of the culture, the drugs and the prevention of harm. CMAJ: Canadian Medical Association Journal, 162, 1843-1848.