Abstract
Cell phones have increased communication and connection across the globe and particularly in sub-Saharan Africa (SSA) — with potential consequences for the HIV epidemic. We examined the association among ownership of cell phones, sexual behaviors (number of sexual partners, alcohol use before sex, inconsistent condom use), and HIV prevalence. Data were from four rounds (2010–2016) of the Rakai Community Cohort Study (N = 58,275). Sexual behaviors and HIV prevalence were compared between people who owned a cell phone to people who did not own a cell phone. We stratified analysis by younger (15–24 years) and older (25+ years) age groups and by gender. Using logistic regression and after adjusting for sociodemographic characteristics, we found cell phone ownership was independently associated with increased odds of having two or more sexual partners in the past twelve months across age and gender groups (young men AOR 1.67, 95% CI 1.47–1.90; young women AOR 1.28 95% CI 1.08–1.53; older men AOR 1.54 95% CI 1.41–1.69; older women AOR 1.44 95% CI 1.26–1.65). Interestingly, young men who owned cell phones had decreased odds of using condoms inconsistently (AOR 0.66, 95% CI 0.57–0.75). For young women, cell phone ownership was associated with increased odds of using alcohol before sex (AOR 1.38 95% CI 1.17–1.63) and increased odds of inconsistent condom use (AOR 1.40, 95% 1.17–1.67). After adjusting for sociodemographic characteristics, only young women who owned cell phones had increased odds of being HIV positive (AOR 1.27 95% CI 1.07–1.50). This association was not mediated by sexual behaviors (Adjusted for sociodemographic characteristics and sexual behaviors AOR 1.24, 95% CI 1.05–1.46). While cell phone ownership appears to be associated with increased HIV risk for young women, we also see a potential opportunity for future cell phone-based health interventions.
Keywords: cell phones, sexual behaviors, HIV prevalence, sub-Saharan Africa
Introduction
Cell phones have changed the way people communicate, socialize and engage in intimate relationships, with major implications for youth development[(1,2)]. In sub-Saharan Africa, a confluence of technological, epidemiological and developmental factors contributes to a potentially distinct relationship between cell phones, health, and youth behaviors. Technologically, people in sub-Saharan Africa rely heavily on cell phones, which serve as a primary means for communicating as well as accessing goods, services, and information [3]. In Uganda, where landlines are used by 1% of the population [4], and only 15% of the population has access to the internet [5], owning a cell phone is rapidly becoming essential for acquiring information and resources [3].
Epidemiologically, sub-Saharan Africa has the highest rates of HIV infection in the world, and young women (aged 15–24 years) have disproportionally higher HIV incidence [6]. Developmentally, adolescence and young adulthood is a life period associated with initiation of new health behaviors and a sense of invulnerability, both of which affect sexual behaviors [7,8]. Given the high prevalence of HIV in the region, the transition to adulthood inevitably involves the added risk of HIV infection, particularly for young women [6,9–13]. These technological, epidemiological, and developmental factors motivate our examination of the association between cell phone ownership, sexual behaviors and HIV prevalence in Uganda.
People in sub-Saharan Africa who own cell phones appear to maintain wider social networks [14,15]. The current paper examines if owning a phone may also be associated with a greater number of sexual partners as compared to people who do not own phones. Given the rates of HIV in sub-Saharan Africa, the consequences of unprotected sex or having multiple concurrent partners are substantial [16,17]. For adolescents and young adults negotiating multiple physical, emotional, and social transitions [18,19], the particular ways that cell phone ownership increases the possibility for social and sexual connections may compound risk of HIV infection. For example, a young person with a cell phone would be able to communicate with potential sexual partners who are outside of their immediate community and geographic location. This could result in planning and maintaining social and sexual relationships that would not be possible without a cell phone.
Further, in part because of their ubiquity, cell phones are a technology that researchers and health workers use to access information [20], and conduct sexual health interventions. These interventions can range from short message service (SMS) targeted to patients to improve adherence to HIV antiretroviral therapy (ART) [21], to mobile applications that health workers can use when conducting behavioral health interventions [22–25]. People who own phones may benefit from the increased exposure to health information as well as the overall increased access to goods and information.
Thus, cell phones provide important ways for people in Uganda - and across sub-Saharan Africa - to communicate and access information and resources. Simultaneously, in the context of the HIV epidemic this connection may also be linked with sexual behaviors and thus with risk of HIV infection. Using survey data from the Rakai Community Cohort Study (RCCS), we explored associations between cell phone ownership and sexual behaviors and HIV prevalence in Rakai, Uganda. Our findings have relevance for behavioral interventions targeted at adolescents and young adults in Uganda and across sub-Saharan Africa.
Methods
Ethical Considerations
Approvals for the current analysis were obtained from the Uganda National Council for Science and Technology (UNCST) and Institutional Review Boards (IRBs) at the Uganda Virus Research Institute, Columbia University, Johns Hopkins University, and Western Institutional Review Board (WIRB)1 in the United States.
Data Source
The data were drawn from the RCCS. The RCCS, which began in 1994 and continues to-date, gathers demographic information, and sexual and reproductive health responses from an open cohort of residents 15–49 years of age in 40 communities in the Rakai district of southcentral Uganda. The district borders Lake Victoria on the East and Tanzania in the south. For most people agricultural work is the main source of income and the population is approximately 516,000 [26]. Most of the population (57%) is between 0 and 17 years old, 19% are 18 to 30 years old. Estimates of HIV prevalence range from 10–25% depending on the community [27]. The RCCS research design and procedures have been described in detail elsewhere [28–30]. In brief, during survey rounds, which occur every 12–18 months, people aged 15–49 years in the RCCS communities are consented. Minors (ages 15–17) are assented, and parent/guardian consent for participation is obtained. Participants (n ~ 14,000 per year) then complete a private face-to-face interview conducted by same-sex interviewers fluent in the local languages (usually Luganda), and are asked to provide biological specimens for HIV and sexually transmitted infection (STI) testing. Among all residents in the communities, RCCS achieves over 66% coverage of the de jure population (i.e., all eligible censused persons) and over 90% coverage of the de facto population (all eligible persons present in the community at the time of the survey), with 99% of consenting participants responding to the complete questionnaire and over 90% agreeing to specimen collection. In addition, all members of the study communities receive voluntary HIV prevention education, counseling and testing.
The current analyses focus on RCCS participants residing in trading and agrarian communities, interviewed between March 2010 and April 2016 (survey rounds 14–17). Prior to 2010, questions on cell phone ownership were not included.
Description of Measures
Cell phone ownership was considered as both the outcome of sociodemographic characteristics and the predictor of sexual behaviors and HIV status.
Cell phone ownership.
The binary cell phone measure was generated from a contact form where participants indicated that they owned a cell phone. Participants who reported only having access to a shared phone were included in the no phone category
Sexual behavior and HIV status.
Sexual behaviors were measured using the following indicators: number of sex partners in the past 12 months (categorized as 0–1, or 2+), alcohol consumption prior to sex, and inconsistent / non-use versus consistent use of condoms. These measures were constructed based on detailed questions regarding sexual partnerships in the last year. The RCCS questionnaire assessed an individual’s four most recent sexual partners after inquiring if they had ever had “sexual intercourse with any person”. Condom use was classified as “consistent” if the participant indicated they always used condoms with all recorded partners. Alcohol use was classified as “no” if the participant indicated the absence of alcohol before sex in the most recent sexual encounter with each of the reported partners. Sexual behavior measures considered in this study were consistent with previous RCCS analyses [31,32]. HIV status was assessed using biological testing.
Sociodemographic measures.
Information on sociodemographic measures including gender, age, place of residence (rural and trading center), marital status (never married, currently married and formerly married), and educational attainment (no formal education, some primary education, some post-primary education), were assessed using single item questions.
Based on the most common responses we collapsed occupation into the following categories: agricultural and housework, administrative and teaching, business, current student, and other. Socioeconomic status (SES) was categorized as low, middle, and high, and based on a series of questions about household possessions such as radios and the use of modern building materials to construct their dwelling.
Statistical Methods
Prevalence of sociodemographic characteristics, including round, place of residence, SES, religion, marital status, educational attainment and occupation, were evaluated over the four survey rounds using basic descriptive statistics and stratified by gender. These sociodemographic variables have shown to be associated with either cell phone ownership or HIV status in previous research [3, 12, 13, 20, 14, 15, 28–32]. Their bivariate associations with cellphone ownership in the RCCS cohort were evaluated using crude odds ratios and their 95% confidence intervals. To further adjust their correlations and confounding effects, we also investigated their conditional associations using logistic regression with repeated outcomes, in which individual cellphone ownership at each round was included as the outcome, and all the sociodemographic variables are included as covariates. The resulting adjusted odds ratios (AOR) and their 95% confidence interval were reported. In the RCCS cohort, we do have multiple observations for some participants, though the average participant only contributed to 2.4 rounds of data, with the average youth only contributing 1.9 rounds and older adults contributing an average of 2.7 rounds. To account for repeated measures, we used generalized estimating equations (GEE) with exchangeable correlation structure to estimate the logistic models. Given the short longitudinal duration, we selected the exchangeable correlation structure as we did not expect a lot of decay in the correlation of measurements with increasing time. Next, we evaluated the potential impact of cell phone ownership on sexual behaviors using logistic models with GEE and adjusting for survey round and the sociodemographic variables noted above. Since it was reported that young people use cell phones in distinct ways as compared to older generations [14], and young women are widely recognized as an especially high risk group for HIV infection [6,9–13], we stratified our analyses by age groups (15–24 years or 25+ years) and gender, and used a regression model with an interaction term to formally test gender-specific and age-specific interactions.
Finally, we investigated whether owning cell phones was associated with higher HIV prevalence and if such associations were mediated by sexual behaviors. To do so, we constructed two sets of models. First, we built an HIV prevalence model that included cellphone ownership as a key covariate, and sociodemographic factors as controlling variables; logistic regression with repeated measures was used as the HIV prevalence model. Second, we expanded the HIV prevalence model by further including the sexual behaviors as covariates, and compared the coefficients associated with cell phone ownership with and without the inclusion of sexual behaviors. To robustify the results, in these analyses, we bootstrapped subjects to determine the standard errors of the model estimates, which account for the within subject correlations without a pre-specified correlation structure [33]. We performed all these analyses using STATA 14.0 and R 3.4.3.
Results
Sociodemographic Characteristics
Sociodemographic characteristics of the sample are described in Table 1. Overall, we analyzed 58,275 observations from 32,571 individuals. Sixty-seven percent of men owned a cell phone as compared to 49% of women. A greater percentage of observations were drawn from rural (56% men, 53% women) than trading center (44% men, 47% women) communities. The most common occupations for men and women were agriculture and housework (31% men, 45% women) and business (17% men, 21% women). Prevalence of HIV was 10% for men and 16% for women. In terms of sexual behaviors 15% of men and 10% of women reported never having sex, 32% of men and 6% of women reported having two or more partners in the last 12 months, 36% of men and 19% of women reported consuming alcohol before sex, and 13% of men and 8% of women reported consistent condom use with all sexual partners.
Table 1:
Characteristics of Rakai participants (non-fishing villages), Rakai Community Cohort Study (RCCS), 2010–2016
| Observations (%) | ||
|---|---|---|
| Characteristics of participants | Men | Women |
| Number of Unique Persons | 14520 | 18051 |
| Number of Observations | 25868 (100%) | 32407 (100%) |
| Cellphone Status | ||
| Owning a cellphone | 17290 (67%) | 16673 (49%) |
| Do not own a cellphone | 8578 (33%) | 15734 (51%) |
| Age (years) | ||
| 15–24 | 10041 (39%) | 11675 (36%) |
| 25+ | 15827 (61%) | 20732 (64%) |
| Round Number | ||
| 14 | 7039 (27%) | 8882 (27%) |
| 15 | 5885 (23%) | 7357 (23%) |
| 16 | 6244 (24%) | 7804 (24%) |
| 17 | 6700 (26%) | 8364 (26%) |
| Place of Residence | ||
| Rural | 14499 (56%) | 17294 (53%) |
| Trading Center | 11369 (44%) | 15113 (47%) |
| Socioeconomic Status | ||
| Low | 3317 (13%) | 3532 (11%) |
| Middle | 7476 (29%) | 8538 (26%) |
| High | 15075 (58%) | 20337 (63%) |
| Religion | ||
| Catholic | 16429 (64%) | 20260 (63%) |
| Protestant | 4464 (18%) | 5756 (18%) |
| Muslim | 3531 (17%) | 4641 (14%) |
| Other | 1244 (5%) | 1750 (5%) |
| Marital Status | ||
| Never Married | 10076 (39%) | 7369 (23%) |
| Currently Married | 13721 (53%) | 19303 (60%) |
| Separated/Widowed | 2071 (8%) | 5735 (18%) |
| Highest Level of Schooling | ||
| None | 700 (3%) | 1462 (5%) |
| Primary | 16142 (62%) | 18689 (58%) |
| Post-Primary | 9026 (35%) | 12256 (38%) |
| Primary Occupation | ||
| Agriculture/housework | 8040 (31%) | 14548 (45%) |
| Admin /Teaching | 1576 (6%) | 2102 (6%) |
| Student | 5325 (20%) | 4174 (13%) |
| Business | 4305 (17%) | 6836 (21%) |
| Other | 6622 (26%) | 4747 (15%) |
| HIV Status | ||
| Negative | 23310 (90%) | 27183 (84%) |
| Positive | 2558 (10%) | 5224 (16%) |
| Ever had sex | ||
| Yes | 21861 (85%) | 29103 (90%) |
| No | 4007 (15%) | 3304 (10%) |
| Number of Sex Partners in the Past 12 months† | ||
| 0–1 | 13551 (62%) | 27076 (93%) |
| 2+ | 8310 (38%) | 2027 (7%) |
| Alcohol used before sex† | ||
| Yes | 7589 (36%) | 5127 (19%) |
| No | 13347 (64%) | 22231 (81%) |
| Condoms used† | ||
| Inconsistent | 19027 (87%) | 26709 (92%) |
| Consistent with all partners | 2834 (13%) | 2394 (8%) |
Included observations who reported ever having sex in number of sexual partners, alcohol use and inconsistent condom use analyses.
Cell Phone Ownership
Table 2 describes the association between cell phone ownership and sociodemographic variables. A greater proportion of men (67%) than women (49%) owned a cell phone.
Table 2:
Crude and adjusted associations of cell phone ownership with demographic factors, RCCS, 2010–2016
| Own Personal Cell Phone vs Not Owning Cell Phone | ||||||||
|---|---|---|---|---|---|---|---|---|
| Men | Women | |||||||
| Demographics | N | %† | Crude OR (95% CI) |
Adjusted OR § (95% CI) |
N | %† | Crude OR (95% CI) |
Adjusted OR § (95% CI) |
| Number of observations | 17290 | 67% | 16673 | 49% | ||||
| Age (years) | ||||||||
| 15–19 | 1784 | 31% | 0.13 (0.11–0.14)*** | 0.22 (0.19–0.25)*** | 909 | 15% | 0.13 (0.12–0.15)*** | 0.13 (0.11–0.15)*** |
| 20–24 | 3232 | 75% | 0.78 (0.71–0.86)*** | 0.98 (0.88–1.10) | 2812 | 49% | 0.66 (0.61–0.70)*** | 0.58 (0.53–0.63)*** |
| 25–29 (reference) | 3279 | 79% | 1 | 1 | 3493 | 59% | 1 | 1 |
| 30–35 | 3032 | 79% | 1.04 (0.93–1.13) | 1.04 (0.93–1.16) | 3467 | 62% | 1.15 (1.07–1.23)*** | 1.36(1.25–1.79)*** |
| 35–39 | 2702 | 78% | 0.94 (0.84–1.04) | 0.91 (0.80–1.03) | 3000 | 66% | 1.32 (1.22–1.43)*** | 1.61(1.47–2.15)*** |
| 40+ | 3261 | 75% | 0.84 (0.75–0.94)** | 0.80 (0.70–0.90)*** | 2992 | 63% | 1.27 (1.16–1.38)*** | 1.64(1.48–2.71)*** |
| Round Number | ||||||||
| 14 (reference) | 4164 | 59% | 1 | 1 | 3457 | 39% | 1 | 1 |
| 15 | 4031 | 68% | 1.47 (1.39–1.56)*** | 1.50 (1.40–1.62)*** | 3924 | 53% | 1.68 (1.60–1.77)*** | 1.68 (1.58– 1.79)*** |
| 16 | 4258 | 68% | 1.62 (1.53–1.72)*** | 1.74 (1.61–1.87)*** | 4268 | 55% | 2.03 (1.93–2.13)*** | 2.02 (1.89– 2.15)*** |
| 17 | 4837 | 72% | 2.07 (1.95–2.20)*** | 2.29 (2.12–2.47)*** | 5024 | 60% | 2.56 (2.44–2.69)*** | 2.54 (2.38– 2.71)*** |
| Place of Residence | ||||||||
| Rural (reference) | 9280 | 64% | 1 | 1 | 7927 | 46% | 1 | 1 |
| Trading Centers | 8010 | 79% | 1.41 (1.30–1.50)*** | 1.13 (1.05–1.22)** | 8746 | 58% | 1.68 (1.59–1.77)*** | 1.28 (1.20– 1.36)*** |
| Socioeconomic Status | ||||||||
| Low (reference) | 1587 | 48% | 1 | 1 | 1024 | 29% | 1 | 1 |
| Middle | 4406 | 59% | 1.45 (1.34–1.57)*** | 1.29 (1.17–1.42)*** | 3229 | 38% | 1.53 (1.41–1.67)*** | 1.30 (1.18– 1.43)*** |
| High | 11297 | 75% | 2.69 (2.48–2.92)*** | 2.08 (1.88–2.29)*** | 12420 | 61% | 3.27 (3.01–3.55)*** | 2.25 (2.05– 2.48)*** |
| Educational Attainment | ||||||||
| None (reference) | 298 | 43% | 1 | 1 | 421 | 29% | 1 | 1 |
| Primary | 9980 | 62% | 2.12 (1.76–2.55)*** | 3.27 (2.69–3.98)*** | 8506 | 46% | 2.05 (1.77–2.36)*** | 3.08 (2.64– 3.58)*** |
| Post-Primary | 7012 | 78% | 4.51 (3.73–5.45)*** | 8.51(6.91–10.49)*** | 7746 | 63% | 4.19 (3.63–4.85)*** | 7.86 (6.69– 9.24)*** |
| Current Marital Status | ||||||||
| Never Married (reference) | 4929 | 49% | 1 | 1 | 2535 | 34% | 1 | 1 |
| Currently Married | 11225 | 82% | 4.14 (3.88–4.41)*** | 1.86 (1.67–2.07)*** | 10332 | 54% | 2.11 (1.97–2.25)*** | 0.78 (0.71– 0.86)*** |
| Separated/Widowed | 1136 | 55% | 1.57 (1.42–1.73)*** | 0.72 (0.63–0.83)*** | 3806 | 66% | 3.48 (3.21–3.76)*** | 1.18 (1.06– 1.33)** |
| Primary Occupation | ||||||||
| Agriculture | 5131 | 64% | 1 | 1 | 6382 | 44% | 1 | 1 |
| Admin/Teaching | 1529 | 97% | 10.8 (8.34–13.98)*** | 4.34 (3.20–5.88)*** | 1883 | 90% | 7.99 (6.92–9.23)*** | 3.76 (3.18– 4.44)*** |
| Student | 2189 | 41% | 0.42 (0.39–0.45)*** | 0.59 (0.53–0.66)*** | 821 | 20% | 0.34 (0.31–0.37)*** | 0.60 (0.53– 0.68)*** |
| Business | 3645 | 85% | 2.40 (2.19–2.62)*** | 2.04 (1.84–2.26)*** | 4776 | 70% | 2.40 (2.26–2.56)*** | 2.02 (1.87– 2.17)*** |
| Other | 4796 | 72% | 1.41 (1.31–1.52)*** | 1.47 (1.35–1.60)*** | 2811 | 59% | 1.66 (1.56–1.78)*** | 1.49 (1.37– 1.62)*** |
Abbreviations: OR: odds ratio; CI: confidence interval.
P-value <0.05,
<0.01, and
<0.001.
The percentages are given as row percent representing the proportions of owning a personal cell phone among all the observations within each characteristic category.
Odds ratio (OR) greater than 1 indicates a greater likelihood of owning a phone. OR less than 1 indicates a lower likelihood of owning a phone.
The OR ratios were adjusted for age group, visit rounds, place of residence, SES, religion, education, current marital status, and occupation.
Cell phone ownership increased over time, such that fewer people owned phones at round 14 (59% of men, 39% of women) as compared to round 17 (72% of men, 60% of women). Cell phone ownership was higher in trading centers than rural areas (Men AOR 1.13, 95% CI 1.30–1.50; Women: AOR 1.28 95% CI 1.20–1.36).
Cell phone ownership was highest in the highest SES group compared to the lowest SES group (High SES Men 75% AOR 2.08, 95% CI 1.88–2.29; High SES Women AOR 2.25 95% CI 2.05–2.48). Similarly, cell phone ownership increased with educational attainment, and those with post-primary education had the highest proportion of phone ownership (Men 78% AOR 8.51 95% CI 6.91–10.49; Women 63% AOR 7.86 95% CI 6.69–9.24) as compared to those who had no formal schooling.
Cell phone ownership also differed by occupation and administrators and teachers had the highest proportions of cell phone ownership (Men 97%, AOR 4.34, 95% CI 3.20–5.88; Women 90% 3.76 95% CI 3.18–4.44) as compared to agricultural workers (Men 64%; Women 44%). There were significant interactive effects between age and place of residence, SES, marital status and occupation for cell phone ownership, and we present these data in appendix table 2a.
Sexual Behaviors
The following results include the odds of each sexual behavior comparing those who owned a cell phone and those who did not own a cell phone included in Table 3. The associations between cell phone ownership and sexual behaviors differed by gender and age group. When we examined these associations for narrower 10-year age groups, we found the results were similar to the broader age group categories and provide this information in appendix 3a.
Table 3:
The association between cell phone ownership and sexual behaviors for sexually active RCCS participants stratified by gender, and age group, from 2010–2016
| Adjusted Odds Ratio § (95% CI) |
||||
|---|---|---|---|---|
| N† | Own Personal Cell Phone vs. No cell phone (reference) |
|||
| Outcomes | Men | Women | For men | For women |
| Number of sex partners in the past 12 months (0–1 vs 2+) | ||||
| Ages 15–24 years | 6222 | 8491 | 1.67(1.47–1.90)* | 1.28(1.08–1.53)* |
| Ages 25+ years | 15639 | 20612 | 1.54(1.41–1.69)* | 1.44(1.26–1.65)* |
| Alcohol used before sex (yes vs. no) | ||||
| Ages 15–24 years | 5750 | 8287 | 0.85(0.71–1.03) | 1.38(1.17–1.63)*** |
| Ages 25+ years | 15186 | 19071 | 0.68(0.62–0.75)*** | 1.28(1.18–1.39)*** |
| Inconsistent condom use | ||||
| Ages 15–24 years | 6222 | 8491 | 0.66(0.57–0.75)*** | 1.40(1.17–1.67)*** |
| Ages 25+ years | 15639 | 20612 | 0.87(0.74–1.03) | 0.65(0.57–0.74)*** |
P-value <0.05,
<0.01, and
<0.001.
N are the person-rounds of each stratified group that were used to fit the models for odds ratios.
Odds ratio (OR) greater than 1 indicates that a participant who owns a phone has a greater likelihood of engaging in this behavior as compared to someone who does not own a phone. OR less than 1 indicates that a participant who owns a phone has a lower likelihood of engaging in this behavior as compared to someone who does not own a phone.
The OR were adjusted for visit rounds, location, SES, religion, education, current marital status, and occupation.
Number of sexual partners.
Cell phone ownership was associated with increased odds of having two or more sexual partners in the past year for all age and gender groups (young men AOR 1.67, 95% CI 1.47–1.90; young women AOR 1.28 95% CI 1.08–1.53; older men AOR 1.54 95% CI 1.41–1.69; older women AOR 1.44 95% CI 1.26–1.65).
Alcohol consumption before sex.
Cell phone ownership was associated with increased odds of using alcohol before sex for young women (AOR 1.38, 95% CI 1.17–1.63), and for older women (AOR 1.28, 95% CI 1.18–1.39). However, in men aged 25–49, cell phone ownership was associated with decreased odds of using alcohol before sex (AOR 0.68, 95% CI 0.62–0.75) and the association between phone ownership and alcohol use was not significant for men under 25 years of age.
Consistency of condom usage.
For condom use we compared respondents who always used a condom to respondents who reported inconsistently or never using a condom. Young women who owned a cell phone had increased odds of using condoms inconsistently (AOR 1.40, 95% CI 1.17–1.67). In older women, cell phone ownership was associated with decreased odds of using condoms inconsistently (AOR 0.65, 95% CI 0.57–0.74). Young men who owned cell phones had decreased odds of using condoms inconsistently (AOR 0.66, 95% CI 0.57–0.75) and the association between phone ownership and condom use was not significant for men over 25 years of age.
HIV Prevalence
Table 4 shows that the associations between cell phone ownership and HIV prevalence were moderated by gender. For young women owning a cell phone was significantly associated with increased odds of being HIV positive (Adjusted for sociodemographic characteristics OR 1.27 95% CI 1.07–1.50). Sexual behaviors did not appear to mediate this association (Adjusted for sociodemographic characteristics and sexual behaviors OR 1.24 95% CI 1.05–1.46). The association between cell phone ownership and HIV status was not significant for any other group.
Table 4:
HIV prevalence among those owning a cell phone vs. not owning a cell phone for sexually active RCCS participants between stratified by round, gender, and age group from 2010–2016
| Odds Ratio § (95% CI) Own Personal Cell Phone vs. No cell phone (reference) |
||||||||
|---|---|---|---|---|---|---|---|---|
| N† | Crude | Adjusted for demographics‡ | Adjusted for both demographics and sexual behaviors § | |||||
| Outcomes | Men | Women | For men | For women | For men | For women | For men | For women |
| HIV prevalence overall N=58275 | ||||||||
| Ages 15–24 years | 6222 | 8491 | 1.09(0.78–1.52) | 1.19(1.03–1.39)* | 0.89(0.63–1.27) | 1.27(1.07–1.50)** | 0.89(0.62–1.27) | 1.24(1.05–1.46)* |
| Ages 25+ years | 15639 | 20612 | 0.72(0.65–0.80)*** | 1.09(1.02–1.17)* | 0.98(0.87–1.09) | 1.07(0.99–1.15) | 0.98(0.87–1.10) | 1.02(0.94–1.10) |
Odds ratio (OR) greater than 1 indicates that a participant who owns a phone has a greater likelihood of engaging in this behavior as compared to someone who does not own a phone. OR less than 1 indicates that a participant who owns a phone has a lower likelihood of engaging in this behavior as compared to someone who does not own a phone.
N are the person-rounds of each stratified group that were used to fit the models for odds ratios.
The OR were adjusted for visit rounds, location, SES, religion, education, current marital status, and occupation.
The OR were adjusted for number of sexual partners, alcohol use before sex, and inconsistent condom use.
Discussion
To our knowledge, the current study is the first to examine the association between cell phones, sexual behaviors and HIV prevalence using data from a large open community cohort. It is vital to consider the implications of increased connectivity as cell phone use rises globally. In other sub-Saharan contexts researchers have identified how cell phones may be changing the ways that young people form social connections [3, 4, 14, 15]. Our findings suggest that cell phone ownership may also be associated with sexual behaviors in distinct ways for different age groups and genders. Young women in sub-Saharan Africa are at great risk of HIV infection [9,11], and for young women in the Rakai District cell phone ownership was associated with an increased risk of being HIV positive. Further, our analyses suggest that the magnitude of these associations are sufficient to be clinically relevant.
Across age and gender, we found that cell phone ownership was associated with having multiple (2+) sexual partners. It is possible that owning a cell phone facilitates increased social [3, 4, 14, 15] and sexual connections. Beyond number of sexual partners, cell phone ownership had distinct associations within age group and gender in terms of other sexual behaviors. Older men who owned cell phones were less likely to use alcohol before sex, and younger men who owned phones were more likely to use condoms consistently. In contrast, younger and older women who owned cell phones were more likely to use alcohol before sex. Younger women were less likely to use condoms consistently and older women were more likely to use condoms consistently.
Although, younger women who owned cell phones had a greater number of partners, were more likely to use alcohol before sex, and were less likely to use condoms consistently, these sexual behaviors did not appear to mediate the association between cell phone ownership and HIV prevalence. Additional quantitative and qualitative research is needed to establish whether cell phone ownership influences HIV risk and to determine what came first, the sexual behavior or cell phone ownership. In Rakai further research is needed to examine how cell phone technology may be related to changing social and sexual dynamics.
Existing ethnographic evidence suggests that inexpensive phones are one common gift from older men to younger women, and this gifting practice is so common in South Africa that the cheap phones have been given their own colloquial term mobile foza orana named after an invasive species of crayfish [34]. Therefore, owning a cell phone for young women may be the result of engaging in increased sexual activity. While ethnographic reports, as well as the popular media, have emphasized the relationship between cell phones and transactional sex in southern Africa [35,36], the RCCS has not included a question on cell phones and transactional sex and therefore we could not examine this association in our current analyses.
Limitations and Future Directions
There may be alternative explanations for the directionality of the association between cell phone ownership and sexual behaviors and HIV prevalence. Further longitudinal research could determine if the association between cell phones and HIV status is specific to a particular age range or if there is a generational difference between the way younger and older women use phones. Additionally, young women who are already engaging in certain sexual behaviors (e.g. sex with multiple partners, alcohol before sex, condomless sex) could be more likely to acquire their own phones. Thus, the behaviors may predate the phone ownership and instead, phone ownership may in fact be a proxy for youth who engage in specific sexual behaviors. Given the complexity and depth of the RCCS data, and the rapidly increasing rates of cell phone ownership, especially among adolescents and young adults, future research could bolster our theory of directionality by identifying participants and measuring sexual behaviors before and after cell phone ownership. Marginal structural modeling and ethnographic research are two approaches that could be used to examine the direction of this association.
Another limitation is that the current work only considers prevalence, and future research should examine how cell phones and sexual behaviors may be related to HIV incidence. In addition, the sexual behavior estimates were based on self-report measures and thus may be subject to social desirability and recall bias. In addition, it is possible that the sexual behaviors are correlated and thus there may be confounding. Results from this study may be generalizable to other rural and semi-urban settings in East Africa with similar cell phone ownership rates and HIV prevalence.
Conclusion
Technological, epidemiological, and developmental factors in sub-Saharan Africa are changing the dynamics of the HIV epidemic with implications for intervention. Cell phones are critical for connecting people and services in sub-Saharan Africa. Previous research has found that cell-phone based interventions can improve outcomes along the continuum of HIV Care (CoC) for persons living with HIV [37], while another review found that cell phone-based interventions were feasible and improved both prevention and care of HIV/STIs for hard to reach populations [38]. To determine the best approach for cell phone-based interventions, future research is needed to examine the types of phones that people own in the region as there is wide variety in the functionality and capabilities of smartphones versus flip phones.
We have shown that in the Rakai District, there is an association between cell phone ownership and sexual behaviors across age and gender and that cell phone ownership is associated with HIV prevalence for young women. Therefore, we see an important need, and a potential opportunity, for cell phone-based interventions to disseminate HIV information and provide health resources especially for highly mobile youth in Rakai and other hard to reach areas. Further, exploratory and qualitative work may illuminate the relationship between cell phone ownership and HIV risk and assist in the design of such technology-based health interventions. Although we found an association between cell phones and HIV prevalence for young women, cell phones also present a potential platform for disseminating health-based interventions in Uganda and elsewhere.
Acknowledgments
We would like to thank the Rakai community cohort participants and the Rakai Health Sciences researchers who conducted the surveys. We would also like to thank Dr. Susie Hoffman and the members of the HIV Center Manuscript workshop for their contributions to the development of this manuscript. We would like to thank Margaret Berrigan and Alyssa Basmajian for their assistance formatting the references, and Esther Spindler for her assistance with copyedits and the manuscript submission.
Appendix
Table 2a:
Crude and adjusted associations of cell phone ownership with demographic factors for younger and older RCCS participants, 2010–2016
| Own Personal Cell Phone vs Not Owning Cell Phone | ||||||||
|---|---|---|---|---|---|---|---|---|
| Age 15–24 years | Age 25+ years | |||||||
| Demographics | N | %† | Crude OR (95% CI) |
Adjusted OR § (95% CI) |
N | %† | Crude OR (95% CI) |
Adjusted OR § (95% CI) |
| Number of observations | 8737 | 40% | 25226 | 69% | ||||
| Gender (years) | ||||||||
| Men (reference) | 5016 | 50% | 1 | 1 | 12274 | 78% | 1 | 1 |
| Women | 3721 | 32% | 0.48 (0.45–0.51)*** | 0.23 (0.21–0.25)*** | 12952 | 62% | 0.48 (0.45–0.51)*** | 0.43(0.41–0.46)*** |
| Round Number | ||||||||
| 14 (reference) | 1699 | 30% | 1 | 1 | 5922 | 58% | 1 | 1 |
| 15 | 2066 | 42% | 1.73 (1.60–1.86)*** | 1.56 (1.43–1.70)*** | 5889 | 71% | 1.62 (1.55–1.70)*** | 1.58 (1.49– 1.67)*** |
| 16 | 2264 | 42% | 2.04 (1.89–2.20)*** | 2.09 (1.92–2.28)*** | 6262 | 72% | 1.85 (1.76–1.94)*** | 1.80 (1.70– 1.90)*** |
| 17 | 2708 | 48% | 2.71 (2.51–2.92)*** | 3.01 (2.76–3.28)*** | 7153 | 76% | 2.24 (2.14–2.35)*** | 2.14 (2.02– 2.27)*** |
| Place of Residence | ||||||||
| Rural (reference) | 4288 | 36% | 1 | 1 | 12919 | 65% | 1 | 1 |
| Trading center | 4449 | 45% | 1.48 (1.40–1.58)*** | 1.20 (1.12–1.29)*** | 12307 | 74% | 1.61 (1.52–1.70)*** | 1.24 (1.16– 1.32)*** |
| Socioeconomic Status | ||||||||
| Low (reference) | 657 | 27% | 1 | 1 | 1954 | 44% | 1 | 1 |
| Middle | 1842 | 31% | 1.25 (1.12–2.60)*** | 1.13 (1.01–1.28)* | 5793 | 57% | 1.63 (1.52–1.75)*** | 1.38 (1.28– 1.49)*** |
| High | 6238 | 46% | 2.35 (2.13–1.39)*** | 1.69 (1.51–1.90)*** | 17479 | 80% | 3.70 (3.45–3.98)*** | 2.47 (2.28– 2.67)*** |
| Educational Attainment | ||||||||
| None (reference) | 43 | 19% | 1 | 1 | 676 | 35% | 1 | 1 |
| Primary | 4264 | 35% | 2.09 (1.47–2.96)*** | 3.00 (2.05–4.40)*** | 14222 | 63% | 3.37 (2.99–3.80)*** | 2.91 (2.56– 3.30)*** |
| Post-Primary | 4430 | 48% | 3.83 (2.70–5.44)*** | 8.53(5.80–12.55)*** | 10328 | 85% | 11.79(10.34–13.44)*** | 7.12 (6.19– 8.18)*** |
| Current Marital Status | ||||||||
| Never Married (reference) | 5318 | 37% | 1 | 1 | 2146 | 70% | 1 | 1 |
| Currently Married | 2976 | 45% | 1.48 (1.39–1.58)*** | 1.57 (1.44–1.72)*** | 18581 | 70% | 0.97 (0.89–1.06) | 1.28 (1.16– 1.41)*** |
| Separated/Widowed | 443 | 59% | 2.49 (2.15–2.89)*** | 2.65 (2.23–3.16)*** | 4499 | 64% | 0.82 (0.74–0.90)*** | 1.23 (1.10– 1.37)*** |
| Primary Occupation | ||||||||
| Agriculture | 1754 | 34% | 1 | 1 | 9759 | 56% | 1 | 1 |
| Admin/Teaching | 514 | 86% | 11.8 (9.24–15.03)*** | 6.56 (5.07–8.49)*** | 2898 | 94% | 7.90 (6.80–9.18)*** | 3.46 (2.92– 4.10)*** |
| Student | 2305 | 27% | 0.67 (0.62–0.72)*** | 0.34 (0.31–0.38)*** | 705 | 85% | 3.51 (2.94–4.19)*** | 1.61 (1.32– 1.97)*** |
| Business | 1922 | 63% | 3.18 (2.89–3.50)*** | 2.54 (2.28–2.83)*** | 6499 | 80% | 2.29 (2.16–2.43)*** | 1.86 (1.74– 1.98)*** |
| Other | 2242 | 54% | 2.19 (2.00–2.38)*** | 1.59 (1.44–1.75)*** | 5365 | 75% | 1.90 (1.79–2.02)*** | 1.32 (1.23– 1.41)*** |
Abbreviations: OR: odds ratio; CI: confidence interval.
P-value <0.05,
<0.01, and
<0.001.
The percentages are given as row percent representing the proportions of owning a personal cell phone among all the observations within each characteristic category.
Odds ratio (OR) greater than 1 indicates a greater likelihood of owning a phone. OR less than 1 indicates a lower likelihood of owning a phone.
The OR ratios were adjusted for visit rounds, place of residence, SES, religion, education, current marital status, and occupation.
Table 3a:
The association between cell phone ownership and sexual behaviors for sexually active RCCS participants stratified by gender, and 5-year age groups, from 2010–2016
| Adjusted Odds Ratio § (95% CI) |
||||
|---|---|---|---|---|
| N† | Own Personal Cell Phone vs. No cell phone (reference) |
|||
| Outcomes | Men | Women | For men | For women |
| Number of sex partners in the past 12 months (0–1 vs 2+) | ||||
| Ages 15–24 years | 6222 | 8491 | 1.67(1.47–1.90)*** | 1.28(1.08–1.53)** |
| Ages 25–34 years | 7872 | 11372 | 1.42(1.26–1.61)*** | 1.64(1.39–1.93)*** |
| Ages 35+ years | 7767 | 9240 | 1.64(1.44–1.88)*** | 1.27(0.99–1.62) |
| Alcohol used before sex (yes vs. no) | ||||
| Ages 15–24 years | 5750 | 8287 | 0.85(0.71–1.03) | 1.38(1.17–1.63)*** |
| Ages 25–34 years | 7721 | 11134 | 0.68(0.60–0.78)*** | 1.31(1.17–1.46)*** |
| Ages 35+ years | 7465 | 7937 | 0.69(0.60–0.80)*** | 1.23(1.09–1.39)*** |
| Inconsistent condom use | ||||
| Ages 15–24 years | 6222 | 8491 | 0.66(0.57–0.75)*** | 1.40(1.17–1.67)*** |
| Ages 25–34 years | 7872 | 11372 | 0.88(0.69–1.12) | 0.71(0.57–0.88)** |
| Ages 35+ years | 7767 | 9240 | 0.92(0.73–1.15) | 0.67(0.56–0.80)*** |
P-value <0.05,
<0.01, and
<0.001.
N are the person-rounds of each stratified group that were used to fit the models for odds ratios.
Odds ratio (OR) greater than 1 indicates that a participant who owns a phone has a greater likelihood of engaging in this behavior as compared to someone who does not own a phone. OR less than 1 indicates that a participant who owns a phone has a lower likelihood of engaging in this behavior as compared to someone who does not own a phone.
The OR were adjusted for visit rounds, location, SES, religion, education, current marital status, and occupation.
Footnotes
Conflicts of Interest Statement
This project was supported by an Award Number R01HD091003 (Principal Investigator, John Santelli, MD., M.P.H.) from the National Institute of Child Health and Human Development. In addition, Philip Kreniske’s contribution was supported by an Award Number T32 MH019139 (Principal Investigator, Theodorus Sandfort, Ph.D.) from the National Institute of Mental Health and a grant from the National Institute of Mental Health to the HIV Center for Clinical and Behavioral Studies at NY State Psychiatric Institute and Columbia University (P30-MH43520; Principal Investigator: Robert Remien, Ph.D.). The content is solely the responsibility of the authors and does not necessarily represent the official views of National Institute of Mental Health or the National Institutes of Health or Columbia University.
WIRB offers review services for more than 400 institutions around the world (http://wirb.com)
References
- 1.Giedd JN. The Digital Revolution and Adolescent Brain Evolution. J Adolesc Health Off Publ Soc Adolesc Med. 2012;51(2):101–105. doi: 10.1016/j.jadohealth.2012.06.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Yan Z. Child and Adolescent Use of Mobile Phones: An Unparalleled Complex Developmental Phenomenon. Child Dev. 2018;89(1):5–16. doi: 10.1111/cdev.12821 [DOI] [PubMed] [Google Scholar]
- 3.Cell Phones in Africa: Communication Lifeline. Pew Res Cent. April 2015. http://www.pewglobal.org/files/2015/04/Pew-Research-Center-Africa-Cell-Phone-Report-FINAL-April-15-2015.pdf.
- 4.Emerging Nations Embrace Internet, Mobile Technology. Pew Res Cent Glob Attitudes Proj. February 2014. http://www.pewglobal.org/2014/02/13/emerging-nations-embrace-internet-mobile-technology/.
- 5.Communications Technology in Emerging and Developing Nations. Pew Res Cent Glob Attitudes Proj. March 2015. http://www.pewglobal.org/2015/03/19/1-communications-technology-in-emerging-and-developing-nations/.
- 6.Putting young people at the centre of the response to HIV. UNAIDS http://www.unaids.org/en/resources/presscentre/featurestories/2012/august/20120810youthday. Published August 10, 2012.
- 7.Breakwell GM. Risk Estimation and Sexual Behaviour: A Longitudinal Study of 16–21-year olds. J Health Psychol. 1996;1(1):79–91. doi: 10.1177/135910539600100107 [DOI] [PubMed] [Google Scholar]
- 8.Anderson KG, Beutel AM, Maughan-Brown B. HIV risk perceptions and first sexual intercourse among youth in Cape Town South Africa. Int Fam Plan Perspect. 2007;33(3):98–105. doi: 10.1363/ifpp.33.098.07 [DOI] [PubMed] [Google Scholar]
- 9.Dellar RC, Dlamini S, Karim QA. Adolescent girls and young women: key populations for HIV epidemic control. J Int AIDS Soc. 2015;18(2 Suppl 1):19408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Napierala Mavedzenge S, Olson R, Doyle AM, Changalucha J, Ross DA. The epidemiology of HIV among young people in sub-Saharan Africa: know your local epidemic and its implications for prevention. J Adolesc Health Off Publ Soc Adolesc Med. 2011;49(6):559–567. doi: 10.1016/j.jadohealth.2011.02.012 [DOI] [PubMed] [Google Scholar]
- 11.Pettifor AE, Rees HV, Kleinschmidt I, et al. Young people’s sexual health in South Africa: HIV prevalence and sexual behaviors from a nationally representative household survey. AIDS 2005;19(14):1525–1534. [DOI] [PubMed] [Google Scholar]
- 12.Santelli JS, Edelstein ZR, Mathur S, et al. Behavioral, Biological, and Demographic Risk and Protective Factors for New HIV Infections among Youth, Rakai, Uganda. J Acquir Immune Defic Syndr 1999. 2013;63(3):393–400. doi: 10.1097/QAI.0b013e3182926795 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Santelli J, Mathur S, Song X, et al. Rising School Enrollment and Declining HIV and Pregnancy Risk Among Adolescents in Rakai District, Uganda, 1994–2013. Glob Soc Welf Res Policy Pract. 2015;2(2):87–103. doi: 10.1007/s40609-015-0029-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Porter G, Hampshire K, Abane A, et al. Intergenerational relations and the power of the cell phone: Perspectives on young people’s phone usage in sub-Saharan Africa. Geoforum. 2015;64:37–46. doi: 10.1016/j.geoforum.2015.06.002 [DOI] [Google Scholar]
- 15.Porter G, Hampshire K, Abane A, et al. Connecting with home, keeping in touch: physical and virtual mobility across stretched families in sub-Saharan Africa. Africa. 2018;88(2):404–424. doi: 10.1017/S0001972017000973 [DOI] [Google Scholar]
- 16.Mmari K, Blum RW. Risk and protective factors that affect adolescent reproductive health in developing countries: a structured literature review. Glob Public Health. 2009;4(4):350–366. doi: 10.1080/17441690701664418 [DOI] [PubMed] [Google Scholar]
- 17.Toska E, Pantelic M, Meinck F, Keck K, Haghighat R, Cluver L. Sex in the shadow of HIV: A systematic review of prevalence, risk factors, and interventions to reduce sexual risk-taking among HIV-positive adolescents and youth in sub-Saharan Africa. PLOS ONE. 2017;12(6):e0178106. doi: 10.1371/journal.pone.0178106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Delany-Moretlwe S, Cowan FM, Busza J, Bolton-Moore C, Kelley K, Fairlie L. Providing comprehensive health services for young key populations: needs, barriers and gaps. J Int AIDS Soc. 2015;18(2Suppl 1). doi: 10.7448/IAS.18.2.19833 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Patton GC, Sawyer SM, Santelli JS, et al. Our future: a Lancet commission on adolescent health and wellbeing. The Lancet. 2016;387(10036):2423–2478. doi: 10.1016/S0140-6736(16)00579-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chang LW, Njie-Carr V, Kalenge S, Kelly JF, Bollinger RC, Alamo-Talisuna S. Perceptions and acceptability of mHealth interventions for improving patient care at a community-based HIV/AIDS clinic in Uganda: a mixed methods study. AIDS Care. 2013;25(7):874–880. doi: 10.1080/09540121.2013.774315 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rana Y, Haberer J, Huang H, et al. Short Message Service (SMS)-Based Intervention to Improve Treatment Adherence among HIV-Positive Youth in Uganda: Focus Group Findings. PLOS ONE. 2015;10(4):e0125187. doi: 10.1371/journal.pone.0125187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Chang LW, Mbabali I, Kong X, et al. Impact of a community health worker HIV treatment and prevention intervention in an HIV hotspot fishing community in Rakai, Uganda (mLAKE): study protocol for a randomized controlled trial. Trials. 2017;18. doi: 10.1186/s13063-017-2243-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Moore L, Chersich MF, Steen R, et al. Community empowerment and involvement of female sex workers in targeted sexual and reproductive health interventions in Africa: a systematic review. Glob Health. 2014;10:47. doi: 10.1186/1744-8603-10-47 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kirby T, Thornber-Dunwell M. Phone apps could help promote sexual health in MSM. Lancet. 2014;384(9952):1415. doi: 10.1016/S0140-6736(14)61849-3 [DOI] [PubMed] [Google Scholar]
- 25.Swendeman D, Rotheram-Borus MJ. Innovation in sexually transmitted disease and HIV prevention: Internet and mobile phone delivery vehicles for global diffusion. Curr Opin Psychiatry. 2010;23(2):139–144. doi: 10.1097/YCO.0b013e328336656a [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Uganda Bureau of Statistics. (2016). National population and housing census 2014: Main Report. Kampala, Uganda. [Google Scholar]
- 27.Chang LW, Grabowski MK, Ssekubugu R, et al. Heterogeneity of the HIV epidemic in agrarian, trading, and fishing communities in Rakai, Uganda: an observational epidemiological study. Lancet HIV 2016; 3(8):e388–e396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Grabowski MK, Serwadda DM, Gray RH, et al. HIV Prevention Efforts and Incidence of HIV in Uganda. N Engl J Med. 2017;377(22):2154–2166. doi: 10.1056/NEJMoa1702150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wawer MJ, Gray RH, Sewankambo NK, et al. A randomized, community trial of intensive sexually transmitted disease control for AIDS prevention, Rakai, Uganda. AIDS Lond Engl. 1998;12(10):1211–1225. [DOI] [PubMed] [Google Scholar]
- 30.Wawer MJ, Sewankambo NK, Serwadda D, et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Rakai Project Study Group. Lancet. 1999;353(9152):525–535. [DOI] [PubMed] [Google Scholar]
- 31.Kankaka EN, Ssekasanvu J, Prodger J, et al. Sexual risk behaviors following circumcision among HIV-positive men in Rakai, Uganda. AIDS Care. 2018;30(8):990–996. doi: 10.1080/09540121.2018.1437253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kagaayi J, Gray RH, Whalen C, et al. Indices to measure risk of HIV Acquisition in Rakai, Uganda. PLOS ONE. 2014;9(4):e92015. doi: 10.1371/journal.pone.0092015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sherman M, le Cassie S. A comparison between bootstrap methods and generalized estimating equations for correlated outcomes in generalized linear models. Communications in Statistics - Simulation and Computation, 26:3, 901–925, DOI: 10.1080/03610919708813417 [DOI] [Google Scholar]
- 34.Stoebenau K, Nixon SA, Rubincam C, et al. More than just talk: the framing of transactional sex and its implications for vulnerability to HIV in Lesotho, Madagascar and South Africa. Glob Health. 2011;7(1):34. doi: 10.1186/1744-8603-7-34 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Archambault JS. Cruising through uncertainty: Cell phones and the politics of display and disguise in Inhambane, Mozambique. Am Ethnol. 40(1):88–101. doi: 10.1111/amet.12007 [DOI] [Google Scholar]
- 36.Stark L Transactional sex and mobile phones in a Tanzanian slum. Suom Antropol J Finn Anthropol Soc. 2013;38(1):12–36. [Google Scholar]
- 37.Henny KD, Wilkes AL, McDonald CM, Denson DJ, Neumann MS. A Rapid Review of eHealth interventions addressing the Continuum of HIV Care (2007–2017). AIDS Behav. 2018;22(1):43–63. doi: 10.1007/s10461-017-1923-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Daher J, Vijh R, Linthwaite B, et al. Do digital innovations for HIV and sexually transmitted infections work? Results from a systematic review (1996–2017). BMJ Open. 2017;7(11). doi: 10.1136/bmjopen-2017-017604 [DOI] [PMC free article] [PubMed] [Google Scholar]