Abstract
Background
The impact of pregnancy on the health and livelihood of adolescents aged 15–19 is substantial. This study explored sociodemographic, behavioral and environmental-level factors associated with adolescent pregnancy across 5 urban disadvantaged settings.
Methods
The Well Being of Adolescents in Vulnerable Environments study used Respondent Driven Sampling (RDS) to recruit males and females from Baltimore (456), Johannesburg (496), Ibadan (449), Delhi (500) and Shanghai(438). RDS-II and post-stratification age weights were used to explore the odds associated with “ever had sex” and “ever pregnant”; adjusted odds of pregnancy and 95% CI were developed by site and gender.
Results
Among the sexually experienced, pregnancy was most common in Baltimore (females 53%, males 25%) and Johannesburg (females 29%, males 22%). Heterosexual experience and therefore pregnancy were rare in Ibadan, Delhi and Shanghai. Current schooling and condom use at first sex decreased the odds of pregnancy among females in Baltimore and Johannesburg participants. Factors associated with higher odds of pregnancy were: early sexual debut (Johannesburg participants, Baltimore females) being raised by someone other than 2 parents (Johannesburg females); alcohol use and binge drinking in the past month (Baltimore participants); greater community violence and poor physical environment (Baltimore males, Johannesburg participants).
Conclusions
The reported prevalence of adolescent pregnancy varies substantially across similarly economically disadvantaged urban settings. These differences are related to large differences in sexual experience, which may be underreported, as well as differences in environmental contexts. Pregnancy risk needs to be understood within the specific context that adolescents reside, with particular attention to neighborhood-level factors.
Keywords: adolescent pregnancy, adolescent sexual activity, sexual risk behaviour, urban neighbourhood disadvantage, sexual debut, family structure, community violence, physical environment
Introduction
Annually, approximately 11% of all births globally occur to adolescents aged 15 to 19 years and 95% of these births occur in developing countries.1 Adolescent pregnancies have a long lasting impact on the physical and mental health, education and livelihood of young women, men and their families.2,3 The health impact of teen pregnancies is significant with increased risks of maternal death, illness and disability, including obstetric fistula, preterm delivery, complications of unsafe abortion, sexually transmitted infections, including HIV, and health risks to infants.2 Early pregnancy has been shown to result in poor social, health and economic outcomes not only for young mothers and fathers2 but for children of young mothers who typically have poorer educational achievement scores, worse socioemotional outcomes,4,5 and adverse birth outcomes1,6 compared to children of older mothers.
An ecological approach to adolescent pregnancy has been proposed to address the individual, environmental and structural correlates of adolescent pregnancy and birth.7 Access to sexual and reproductive health services can be facilitated at the national, environmental and individual levels. Although studies have shown that poor run-down neighborhoods, and housing instability were associated with higher rates of STIs,8–10 few have focused on how urban poverty specifically affects adolescent sexual and reproductive health outcomes. Many adolescents today are growing up in a context of rapid urbanization and migration in search of better opportunities, which when combined with underlying poverty and unstable housing, can exacerbate outcomes such as crime, alcohol, drug use and HIV and STIs. Recognizing the unique challenges of urban poverty is critical in order to contextualize the correlates of adolescent pregnancy in this environment and to combat the high sexual and reproductive health risks in this age group.
We have a unique opportunity to examine risk factors for pregnancy among adolescents in five different impoverished city settings where the same methodology was used to explore these domains. We are also able to examine differences by gender. Examination of factors that are similar as well as unique due to national and environmental differences will improve our understanding of the factors driving pregnancy among adolescents in disadvantaged, urban settings in different parts of the world.
Methods
The Well Being of Adolescents in Vulnerable Environments (WAVE) is a global study of adolescents aged 15 to 19 years living in disadvantaged, urban settings in Baltimore, (USA), Johannesburg (South Africa), Ibadan (Nigeria), Delhi (India) and Shanghai, China. All sites recruited approximately 500 adolescents from economically distressed urban settings and in addition, in Shanghai, the participants were migrant adolescents.
All 5 sites used Respondent Driven Sampling (RDS) to recruit participants due to the financial and logistical challenges of conducting population-based surveys in inner-city environments where high levels of migration and low housing stability make the sampling frame unknown (see Decker et. al in this volume for details of methodology). Seed respondents were recruited from diverse venues where adolescents congregate such as youth centers, theaters, parks and churches.. The recruited seed participants were then encouraged to recruit up to three additional eligible individuals from their peer networks. Coupons were distributed to seeds and successive waves of respondents to link participants back to their respective recruiters as well as identify their place in the recruitment chain. This continued until a sample size of 500 was reached in each site. Average network size (degree) was similar for males and females and there was a high gender homophily, indicating a preference for participants to recruit individuals of the same sex. Interviews were conducted using audio-computer assisted self interview (ACASI) with a standardized instrument using validated measures for items including sexual behavior, health seeking behavior, substance use, violence, gender power relations, and family structure. All instruments were administered in English and the local languages at each site (Delhi: Hindi, Ibadan: Yoruba, South Africa: IsiZulu and Sesotho, Shanghai: Mandarin).
Measures
The main outcome of the current study was “ever pregnant”, conceptualized as ever being pregnant (females) or gotten a partner pregnant (males). Pregnancy experience questions were restricted to those reporting that they ever had heterosexual intercourse. Questions on pregnancy outcomes were: ever abortion, whether a birth ever occurred, and number of children ever born.
Social and demographic measures: The age of participants was dichotomized into 15–16 years vs. 17–19 and treated as a continuous measure in the multivariate analysis. Participants were asked if they were currently in school and educational attainment was collapsed into four categories (<8th grade/primary school, some high school, completed high school, some tertiary education but no degree). Other factors assessed were: relative wealth (better than most, same as most, worse than most); family of origin (raised by two parents, including biological, step- or adoptive parents, one parent or by another person including grandmother, sister, other relative, other non-relative); housing stability assessed by whether participants had a regular place to stay, or if they stayed somewhere other than their regular place for more than three nights per week during the last 30 days.
Sexual and other risk measures: age at sexual debut was dichotomized as age 14 or younger vs. age 15 or above.2 A categorical variable was used to assess number of lifetime sexual partners (1, 2–4, 5 or more). Other sexual measures assessed were; ever sex with someone of the same gender; ever gave or received sex in exchange for money, shelter, food, drugs or other goods; and unwanted sex during (combined measure of ever being coerced or physically forced to have sex during last 12 months). Contraceptive use at first sex included any form of contraception and excluded condom use which was an additional binary measure to capture consistency of condom use during the last 12 months. Finally, alcohol use was measured both as binary variable (ever vs. never finished any alcoholic bevarage) and further conceptualized as alcohol use during the last 30 days (no drink, less than 5 drinks in a row, binge drinking which was 5 or more drinks in a row).
Environmental Factors: Three scales were used to capture the characteristics of the respondents’ environment; physical environment scale with scores ranging from 0–24; perceived fear scale which ranged from 0 to 18 and observation of violence in the past year in one’s neighborhood which ranged from 0 to 18 (see Mmari et. al in this volume for details of environment measures).
Statistical analyses
Data were imported into Stata v12.1 (StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX: StataCorp LP.). RDSII estimators were derived using code developed by Schonlau and Liebau.11. All results were adjusted for cluster, RDS weights and post stratification weights for age. Both weighted and un-weighted estimates were calculated; however, only weighted proportions are reported. For more details on RDS and the effects of weighting, see Decker et al in the present volume.
Descriptive statistics were summarized for adolescent sexual and pregnancy experience among males and females by site. Bivariate associations were explored between the proportion ever pregnant and socio-demographic, behavioral, and environmental variables using Pearson’s corrected chi-square. Means along with standard errors were calculated for each environmental-level scale. Because of the low prevalence of pregnancy and sexual activity in Ibadan, Delhi and Shanghai, an additional bivariate analysis was conducted to better understand the determinants of ever having had sex by site.
Multivariate logistic regression models were used to assess factors associated with pregnancy in Baltimore and Johannesburg. Model development was a multi-stage process in which variables were entered as three different clusters; model I included socio-demographic factors; model II added environmental factors; and in model III behavioral factors were included. The contribution of different factors within each cluster was explored using Akaike’s Information Criterion (AIC), with the lowest AIC value guiding model fit. In this process, factors commonly associated in the literature with adolescent pregnancy (such as current schooling)12,13 were retained despite not being significant in the bivariate analysis. The goal of model development was to find the best fit by gender and site; thus, four different models were developed adjusting for factors contributing specifically to each model. For all analyses the variable lifetime sexual partners was dropped because of large numbers of extreme outliers. All multivariate models among females were restricted to those not currently pregnant as currently pregnant women were less likely to drink alcohol confounding the relationship with alcohol use. Adjusted Odds Ratios (aOR) and associated 95% Confidence Intervals (CI) were then calculated separately for males and females by site.
Results
The final cleaned data set included 2,339 adolescents (Baltimore, N=456, Johannesburg, N=496, Ibadan, N=449, Delhi, N=500 and Shanghai, N=438). Marital status was not included in analyses as all males and 99.9% of females in our study were unmarried. Proportions reporting being sexually experienced ranged from 0.3–17% among female and male adolescents in Delhi to over 75–86% in Baltimore. The proportion ever pregnant ranged from 16% in Shanghai to 53% in Baltimore among sexually experienced females; among sexually experienced males reports of ever impregnating a partner ranged from 11% in Shanghai to 25% in Baltimore. Of females who were ever pregnant, one third reported ever having an induced abortion in Baltimore, 15% in Johannesburg, and 54% in Ibadan (albeit among very small numbers). Male reports of partners experiencing an abortion varied by site to 66% in Baltimore, 62% in Ibadan, 51% in Shanghai, and 45% in Johannesburg. Only three males in Delhi reported a partner having an abortion. The proportion of ever pregnant females reporting a birth ranged from 32% in Johannesburg to 56% in Baltimore; and among similar males, 4% in Delhi to 72% in Ibadan. There were no births reported by females in Shanghai and Delhi nor by males in Shanghai. Of females who reported a birth, most had one child and this was also true for males in Baltimore and Johannesburg as well (Tables 1a and 1b).
TABLE 1.
| A. PREVALENCE OF PREGNANCY AND REPRODUCTIVE HISTORY: FEMALES* | |||||
|---|---|---|---|---|---|
| FEMALES | |||||
| BALTIMORE N=193 |
JOHANNESBURG N=224 |
IBADAN N=229 |
SHANGHAI N=216 |
NEW DELHI N=250 |
|
| W% (U%, n) | W% (U%, n) | W% (U%, n) | W% (U%, n) | W% (U%, n) | |
| EVER SEXUAL | 75.3 (67.4, 130/193) | 55.5 (58.0, 130/224) | 16.0 (13.5, 31/229) | 8.4 (8.8, 19/216) | 0.3 (1.2, 3/250) |
| INTERCOURSE± | |||||
| EVER PREGNANT | 52.9 (54.8, 53/130) | 28.8 (19.2, 25/130) | 24.1 (22.6, 7/31) | 16.1 (15.8, 3/19) | 52.2 (33.3, 1/3) |
| CURRENTLY | 14.4 (17.0, 9/53) | 31.6 (16.0, 4/25) | 36.0 (42.9, 3/7) | 84.0 (66.7, 2/3) | 100.0 (100.0, 1/1) |
| PREGNANT | |||||
| EVER ABORTION | 31.9 (43.4, 23/53) | 14.6 (28.0, 7/25) | 54.5 (57.1, 4/7) | 16.0 (33.3, 1/3) | 100.0 (100.0, 1/1) |
| EVER GAVE BIRTH | 56.1 (43.4, 23/53) | 32.5 (48.0, 12/25) | 38.4 (28.6, 2/7) | -- | -- |
| # CHILDREN GAVE BIRTH | |||||
| TO | |||||
| 1 | 94.0 (82.6 19/23) | 99.4 (91.7, 11/12) | 52.7 (50.0, 1/2) | -- | -- |
| 2 or more | 6.0 (17.4, 4/23) | 0.6 (8.3, 1/12) | 47.3 (50.0, 1/2) | -- | -- |
| B. PREVALENCE OF PREGNANCY AND REPRODUCTIVE HISTORY: MALES* | |||||
|---|---|---|---|---|---|
| MALES | |||||
| BALTIMORE N=263 |
JOHANNESBURG N=272 |
IBADAN N=220 |
SHANGHAI N=222 |
NEW DELHI N=250 |
|
| W% (UW%, n) | W% (UW%, n) | W% (UW%, n) | W% (UW%, n) | W% (UW%, n) | |
| EVER SEXUAL | 86.4 (84.4, 222/263) | 69.2 (79.4, 216/272) | 44.0 (39.1, 86/220) | 25.5 (35.1, 78/222) | 16.7 (15.6, 39/250) |
| INTERCOURSE± | |||||
| PARTNER EVER PREGNANT | 24.9 (28.3, 63/222) | 22.2 (26.4, 57/216) | 17.6 (12.8, 11/86) | 10.7 (19.2, 15/78) | 15.4 (15.4, 6/39) |
| PARTNER EVER ABORTION | 65.7 (55.6, 35/63) | 45.2 (50.9, 29/57) | 62.3 (72.7, 8/11) | 51.4 (60.0, 9/15) | 9.3 (50.0, 3/6) |
| PARTNER EVER GAVE BIRTH | 46.0 (36.5, 23/63) | 32.8 (35.1, 20/57) | 72.2 (63.6, 7/11) | -- | 3.6 (16.7, 1/6) |
| # OF CHILDREN | |||||
| 1 | 77.2 (69.6, 16/23) | 91.3 (80.0, 16/20) | 5.6 (14.3, 1/7) | -- | 100.0 (100.0, 1/1) |
| 2 or more | 22.8 (30.4, 7/23) | 8.7 (20.0, 4/20) | 94.5 (85.7, 6/7) | -- | -- |
Displaying weighted (W) proportions using cluster and combined RDS and age post-stratification weights, followed by unweighted (U) proportions and observations.
Ever/never had heterosexual (vaginal) intercourse
99.9% of females were unmarried at time of survey.
All males were unmarried at time of survey.
Among females in Ibadan and Shanghai, no single factor was associated with ever having (vaginal) sex in both sites, whereas alcohol use was the only factor associated with sexual experience among males in these two cities. Individual-level factors associated with sexual experience included older age (Ibadan participants, females in Shanghai), not being in school (Shanghai participants), and ever-use of alcohol (females in Ibadan, Delhi and all males). Being raised by a single parent or someone other than parents was associated with higher proportion reporting ever sex among males in Ibadan and Shanghai. Unstable housing was associated with a higher proportion of adolescents reporting ever sex among males in Ibadan and participants in Shanghai. Some environmental level factors were associated with a report of ever sex; observed violence in the community (males in Shanghai), perceived fear of being robbed or attacked (males in Shanghai) and poorer physical environment (females in Shanghai) (Tables 2a and b). Bivariate analysis with ever-sex was not possible in Delhi because of the low proportion reporting being sexually experienced.
TABLE 2.
| A. CHARACTERISTICS OF FEMALES WHO REPORT HAVING HAD SEXUAL INTERCOURSE: IBADAN AND SHANGHAI | ||||
|---|---|---|---|---|
| EVER HAD SEX (FEMALES) | ||||
| IBADAN N=229 |
SHANGHAI N=216 |
|||
| W% | UW n | W% | UW n | |
| AGE | ||||
| 15–16 years | 8.5*** | 12/146 | 2.8*** | 4/65 |
| 17–19 years | 25.0 | 19/83 | 11.7 | 15/151 |
| EDUCATION LEVEL | ||||
| Less than 8th grade | 9.3 | 2/36 | 11.4 | 3/34 |
| Some high school | 17.8 | 18/127 | 8.2 | 10/115 |
| High school degree | 16.2 | 10/53 | 4.9 | 4/59 |
| Some college, no degree | 20.3 | 1/13 | 11.2 | 2/8 |
| CURRENTLY IN SCHOOL | ||||
| No | 22.3 | 10/45 | 10.3* | 13/149 |
| Yes | 15.0 | 21/184 | 5.2 | 6/67 |
| RELATIVE WEALTH | ||||
| Better than most | 14.1 | 12/114 | 2.3 | 1/10 |
| Same as most | 17.6 | 15/91 | 8.0 | 12/183 |
| Worse than most | 22.9 | 3/19 | 13.4 | 6/22 |
| PERSON(S) RAISED BY$ | ||||
| Two parents | 15.6 | 21/161 | 8.9 | 17/181 |
| One parent | 31.4 | 4/19 | -- | -- |
| Other | 9.7 | 3/37 | 7.5 | 2/26 |
| UNSTABLE HOUSING | ||||
| No | 16.4 | 23/179 | 5.2* | 13/194 |
| Yes | 17.2 | 8/47 | 33.4 | 6/22 |
| EVER DRANK ALCOHOL | ||||
| No | 13.4* | 21/188 | 3.6 | 3/71 |
| Yes | 30.6 | 10/41 | 10.8 | 16/145 |
| ALCOHOL USE PAST 30 DAYS£ | ||||
| Did not drink last month | 16.6*** | 3/19 | 4.3 | 2/50 |
| < 5 drinks in a row last month | 44.1 | 6/16 | 7.5 | 6/57 |
| ≥ 5 drinks in a row last month | 77.0 | 1/2 | 25.9 | 8/37 |
| PERCEIVED SAFETY IN ENVIRONMENT | ||||
| Safe | 16.6 | 26/195 | 7.9 | 15/179 |
| Unsafe | 17.7 | 15/30 | 10.9 | 4/35 |
| Eversex – W mean (SE) | Ever sex – W mean (SE) | |||
| Yes | No | Yes | No | |
| PERCEIVED FEAR (scale)§ | 4.5 (1.29) | 4.9 (0.31) | 4.5 (0.04) | 2.9 (0.38) |
| COMMUNITY VIOLENCE (scale)§ | 3.1 (0.35) | 3.2 (0.61) | 4.3 (0.39)* | 2.6 (0.18) |
| PHYSICAL ENVIRONMENT (scale)§ | 13.9 (0.63) | 15.1 (0.22) | 7.9 (1.07) | 9.5 (0.15) |
| B. CHARACTERISTICS OF MALES WHO REPORT HAVING HAD SEXUAL INTERCOURSE: IBADAN, NEW DELHI, SHANGHAI | ||||
|---|---|---|---|---|
| EVER HAD SEX (MALES) | ||||
| IBADAN N=220 |
SHANGHAI N=222 |
|||
| W% | UW n | W% | UW n | |
| AGE | ||||
| 15–16 years | 29.8** | 34/121 | 21.9 | 13/62 |
| 17–19 years | 53.2 | 52/99 | 28.9 | 65/160 |
| EDUCATION LEVEL | ||||
| Less than 8th grade | 72.1 | 11/17 | 45.0 | 28/61 |
| Some high school | 40.2 | 43/118 | 15.4 | 28/103 |
| High school degree | 38.8 | 27/72 | 21.8 | 19/51 |
| Some college, no degree | 36.3 | 5/13 | 39.7 | 3/7 |
| CURRENTLY IN SCHOOL | ||||
| No | 51.3 | 18/39 | 33.0* | 68/164 |
| Yes | 40.1 | 67/180 | 1.6 | 10/58 |
| RELATIVE WEALTH | ||||
| Better than most | 48.1** | 56/123 | 19.4 | 9/22 |
| Same as most | 33.7 | 27/83 | 23.0 | 55/172 |
| Worse than most | 19.4 | 1/9 | 48.5 | 14/27 |
| PERSON(S) RAISED BY$ | ||||
| Two parents | 34.9** | 52/159 | 20.3** | 57/182 |
| One parent | 71.5 | 9/14 | 33.6 | 2/4 |
| Other | 54.2 | 20/40 | 46.8 | 13/25 |
| UNSTABLE HOUSING | ||||
| No | 37.6* | 63/171 | 23.4* | 56/166 |
| Yes | 47.0 | 23/48 | 34.6 | 22/56 |
| EVER DRANK ALCOHOL | ||||
| No | 33.8*** | 45/145 | 7.1** | 9/46 |
| Yes | 58.7 | 41/75 | 31.9 | 69/179 |
| ALCOHOL USE PAST 30 DAYS£ | ||||
| Did not drink last month | 51.6 | 18/39 | 19.0 | 12/39 |
| < 5 drinks in a row last month | 34.8 | 10/20 | 35.6 | 23/74 |
| >5 drinks in a row last month | 86.4 | 13/16 | 37.8 | 34/64 |
| PERCEIVED SAFETY IN COMMUNITY | ||||
| Safe | 41.9 | 74/192 | 24.5 | 62/182 |
| Unsafe | 45.8 | 12/27 | 30.9 | 16/40 |
| Ever sex – W mean (SE) | Ever sex – W mean (SE) | |||
| Yes | No | Yes | No | |
| PERCEIVED FEAR (scale)§ | 4.3 (0.55) | 3.9 (0.43) | 5.0 (0.83) | 4.44 (0.19) |
| COMMUNITY VIOLENCE (scale)§ | 4.0 (0.25) | 3.4 (0.51) | 5.3 (0.50) | 3.6 (0.12) |
| PHYSICAL ENVIRONMENT (scale)§ | 12.6 (0.48) | 13.5 (0.72) | 11.4 (1.08) | 9.7 (0.50) |
Displaying weighted (W) row proportions using cluster and combined RDS and age post-stratification weights, followed by unweighted (U) observations by variable.
p<0.001,
p<0.01,
p<0.05.
Comparing ever vs. never had sex by characteristics (row percent).
Note: 99.9% of females were unmarried at time of survey.
Displaying weighted (W) row proportions using cluster and combined RDS and age post-stratification weights, followed by unweighted (U) observations by variable. The varaible unstably housed was excluded for Delhi.
Note: all males were unmarried at time of survey.
Many of the characteristics associated with ever having sex among females in Baltimore and Johannesburg were also related to pregnancy and are described subsequently (See Tables 3a and 3b). The missing sections in these tables are for variables that are not possible to compare with ever-sex (age at first sex, number of sexual partners etc) since those participants have never had sex.
TABLE 3.
| A. SEXUAL EXPERIENCE AND PREGNANCY BY KEY SOCIODEMOGRAPHIC, NEIGHBORHOOD AND BEHAVIORAL CHARACTERISTICS: BALTIMORE AND JOHANNESBURG FEMALES | ||||||||
|---|---|---|---|---|---|---|---|---|
| EVER HAD SEXUAL INTERCOURSE | EVER PREGNANT | |||||||
| BALTIMORE N=193 |
JOHANNESBURG N=224 |
BALTIMORE N=130 |
JOHANNESBURG N=130 |
|||||
| W% | UW n | W% | UW n | W% | UWn | W% | UW n | |
| AGE | ||||||||
| 15–16 years | 62.0*** | 47/91 | 41.2*** | 21/53 | 34.8*** | 10/47 | 47.4* | 6/21 |
| 17–19 years | 82.1 | 83/102 | 69.7 | 109/171 | 62.0 | 43/83 | 18.4 | 19/109 |
| EDUCATION LEVEL | ||||||||
| Less than 8th grade | 51.2* | 15/26 | 39.7* | 6/15 | 61.7 | 9/15 | 11.6 | 1/6 |
| Some high school | 72.8 | 76/118 | 46.3 | 73/140 | 39.0 | 18/76 | 27.8 | 9/73 |
| High school degree | 82.8 | 37/46 | 85.9 | 29/36 | 84.6 | 26/37 | 40.2 | 9/29 |
| Some college, no degree | 94.6 | 2/3 | 70.6 | 22/33 | 0 | 0/0 | 20.5 | 6/22 |
| CURRENTLY IN SCHOOL | ||||||||
| No | 78.2 | 28/32 | 80.9* | 9/12 | 72.3 | 19/28 | 35.2 | 4/9 |
| Yes | 71.4 | 101/160 | 54.4 | 121/211 | 45.7 | 33/101 | 28.2 | 21/121 |
| RELATIVE WEALTH | ||||||||
| Better than most | 78.4 | 35/54 | 72.9** | 34/50 | 57.2 | 14/35 | 30.0 | 4/34 |
| Same as most | 70.4 | 83/124 | 51.8 | 84/154 | 50.6 | 34/83 | 28.5 | 17/84 |
| Worse than most | 64.5 | 11/13 | 57.6 | 12/20 | 27.1 | 5/11 | 28.1 | 4/12 |
| PERSON(S) RAISED BY | ||||||||
| Two parents | 72.0** | 60/93 | 58.8 | 81/131 | 47.2 | 23/60 | 17.0* | 11/81 |
| One parent | 84.3 | 31/45 | 61.6 | 13/22 | 59.8 | 12/31 | 51.7 | 5/13 |
| Other | 59.6 | 29/43 | 60.3 | 32/58 | 56.8 | 15/29 | 44.2 | 8/32 |
| UNSTABLE HOUSING | ||||||||
| No | 72.7 | 125/187 | 53.9 | 106/189 | 51.8*** | 51/125 | 20.5* | 16/106 |
| Yes | 94.1 | 5/6 | 65.1 | 24/35 | 14.2 | 2/5 | 61.9 | 9/24 |
| EVER USED ALCOHOL | ||||||||
| No | 58.6*** | 50/97 | 37.9*** | 37/83 | 42.8* | 22/50 | 12.0** | 4/37 |
| Yes | 85.5 | 80/96 | 66.1 | 93/141 | 56.6 | 31/80 | 34.3 | 21/93 |
| ALCOHOL USE PAST 30 DAYS£ | ||||||||
| Did not drink | 87.6 | 34/42 | 59.5 | 31/54 | 54.0± | 10/32 | 19.1 | 4/30 |
| < 5 drinks in a row | 77.7 | 27/32 | 74.6 | 32/49 | 35.9 | 12/27 | 31.1 | 4/29 |
| >5 drinks in a row | 91.6 | 18/20 | 62.4 | 30/38 | 73.7 | 18/16 | 26.6 | 9/30 |
| AGE FIRST SEX | ||||||||
| 15 years or older | 45.0** | 29/77 | 22.6* | 17/114 | ||||
| 14 years or younger | 61.3 | 24/53 | 57.0 | 8/16 | ||||
| LIFETIME SEXUAL PARTNERS | ||||||||
| 1 | 20.8** | 9/30 | 30.5* | 6/59 | ||||
| 2–4 | 42.6 | 11/47 | 21.3 | 10/55 | ||||
| ≥ 5 | 72.4 | 29/45 | 60.2 | 9/16 | ||||
| EVER SAME SEX PARTNER | ||||||||
| No | 45.3 | 32/90 | 29.4 | 21/114 | ||||
| Yes | 66.4 | 21/39 | 22.6 | 4/16 | ||||
| EVER TRANSACTIONAL SEX | ||||||||
| No | 48.3± | 46/121 | 27.4* | 20/121 | ||||
| Yes | 82.9 | 7/9 | 50.2 | 5/9 | ||||
| UNWANTED SEX PAST YEAR | ||||||||
| No | 49.2 | 43/114 | 29.0 | 20/109 | ||||
| Yes | 75.5 | 10/15 | 28.7 | 5/21 | ||||
| ANY CONTRACEPTION AT FIRST SEX± | ||||||||
| No | 72.4*** | 16/30 | 52.8* | 7/19 | ||||
| Yes | 50.0 | 37/99 | 20.9 | 18/111 | ||||
| CONDOM AT FIRST SEX | ||||||||
| No | 75.3*** | 16/29 | 57.1* | 8/23 | ||||
| Yes | 48.9 | 36/95 | 20.1 | 16/105 | ||||
| OTHER CONTRACEPTION AT FIRST SEX±± | ||||||||
| No | 49.8 | 37/91 | 31.5 | 18/88 | ||||
| Yes | 57.9 | 16/37 | 23.2 | 6/35 | ||||
| ALWAYS USE CONDOMS | ||||||||
| No | 55.6* | 35/81 | 43.4** | 19/57 | ||||
| Yes | 27.1 | 9/28 | 5.2 | 3/46 | ||||
| PERCEIVED SAFETY IN ENVIRONMENT | ||||||||
| Safe | 67.9** | 68/118 | 57.6 | 62/113 | 51.2 | 25/68 | 28.8 | 11/62 |
| Unsafe | 82.7 | 62/75 | 53.8 | 68/111 | 51.5 | 28/62 | 28.7 | 14/68 |
| Ever had sexual intercourse - W Mean (SE) | Ever pregnant - W Mean (SE) | |||||||
| Yes | No | Yes | No | Yes | No | Yes | No | |
| PERCEIVED FEAR (scale)§ | 5.5 (0.6) | 5.0 (0.4) | 6.4 (0.33) | 6.6 (0.18) | 5.7 (0.73) | 5.3 (1.0) | 7.9 (0.86)** | 6.1 (0.11) |
| COMMUNITY VIOLENCE (scale)§ | 6.56 (0.60) | 5.4 (0.30) | 7.2 (0.21) | 6.8 (0.39) | 6.5 (0.69) | 6.6 (0.66) | 8.9 (0.74)*** | 6.5 (0.28) |
| PHYSICAL ENVIRONMENT (scale)§ | 11.3 (0.44)** | 15.0 (0.43) | 12.6 (0.40) | 13.1 (0.51) | 10.3 (0.50)** | 12.4 (0.76) | 8.6 (1.05)** | 14.2 (0.42) |
| B. SEXUAL EXPERIENCE AND PREGNANCY BY KEY SOCIODEMOGRAPHIC, NEIGHBORHOOD AND BEHAVIORAL CHARACTERISTICS: BALTIMORE AND JOHANNESBURG MALES | ||||||||
|---|---|---|---|---|---|---|---|---|
| EVER HETEROSEXUAL INTERCOURSE | PARTNER EVER PREGNANT | |||||||
| BALTIMORE N=263 |
JOHANNESBURG N=272 |
BALTIMORE N=222 |
JOHANNESBURG N=216 |
|||||
| W% | UW n | W% | UW n | W% | UW n | W% | UW n | |
| AGE | ||||||||
| 15–16 years | 77.8* | 127/161 | 54.0** | 47/70 | 12.5* | 24/127 | 15.8 | 8/47 |
| 17–19 years | 94.3 | 95/102 | 82.3 | 169/202 | 36.0 | 39/95 | 27.0 | 49/169 |
| EDUCATION LEVEL | ||||||||
| Less than 8th grade | 68.2 | 38/55 | 26.9* | 10/22) | 25.9 | 8/38 | 18.9 | 3/10 |
| Some high school | 92.0 | 141/160 | 74.8 | 147/177 | 27.1 | 39/141 | 25.8 | 43/147 |
| High school degree | 85.6 | 36/41 | 84.9 | 37/42 | 27.1 | 14/36 | 17.1 | 7/37 |
| Some college, no degree | 100.0 | 7/7 | 57.6 | 22/31 | 28.9 | 2/7 | 15.2 | 4/22 |
| CURRENTLY IN SCHOOL | ||||||||
| No | 96.8 | 34/35 | 88.8*** | 69/76 | 41.9*** | 47/188 | 28.9 | 32/146 |
| Yes | 84.9 | 188/228 | 61.2 | 146/195 | 23.1 | 16/34 | 19.5 | 25/69 |
| RELATIVE WEALTH | ||||||||
| Better than most | 84.9 | 84/100 | 54.1 | 45/60 | 18.3* | 17/84 | 11.6 | 7/45 |
| Same as most | 89.6 | 118/138 | 70.0 | 150/188 | 30.7 | 39/118 | 23.6 | 42/150 |
| Worse than most | 86.5 | 13/16 | 86.0 | 21/24 | 47.6 | 5/13 | 32.4 | 8/21 |
| PERSON(S) RAISED BY$ | ||||||||
| Two parents | 86.8 | 92/113 | 68.4 | 108/139 | 22.6 | 23/92 | 17.8 | 23/108 |
| One parent | 87.0 | 50/60 | 63.1 | 21/26 | 32.7 | 14/50 | 29.5 | 6/21 |
| Other | 88.7 | 60/66 | 68.1 | 77/94 | 28.4 | 19/60 | 24.1 | 24/77 |
| UNSTABLE HOUSING | ||||||||
| No | 88.7 | 201/234 | 68.5 | 168/212 | 25.3 | 53/201 | 20.7 | 40/168 |
| Yes | 94.6 | 21/24 | 64.7 | 48/60 | 44.3 | 10/21 | 27.3 | 17/48 |
| EVER USED ALCOHOL | ||||||||
| No | 84.8 | 110/136 | 46.7* | 28/56 | 21.8 | 23/110 | 23.2 | 6/28 |
| Yes | 89.2 | 112/127 | 75.7 | 188/216 | 31.3 | 40/112 | 18.7 | 51/188 |
| ALCOHOL USE PAST 30 DAYS£ | ||||||||
| Did not drink | 95.4 | 49/54 | 48.6* | 33/47 | 20.6* | 10/49 | 13.5** | 6/33 |
| < 5 drinks in a row | 83.0 | 30/34 | 82.6 | 71/78 | 37.6 | 12/30 | 14.8 | 14/71 |
| >5 drinks in a row | 88.2 | 30/33 | 89.7 | 84/91 | 48.3 | 18/30 | 33.3 | 31/84 |
| AGE FIRST SEX | ||||||||
| 15 years or older | 16.9*** | 9/41 | 12.1*** | 23/123 | ||||
| 14 years or younger | 30.9 | 56/181 | 34.7 | 34/93 | ||||
| LIFETIME SEXUAL PARTNERS | ||||||||
| 1 | 10.2* | 3/13 | 16.4* | 2/16 | ||||
| 2–4 | 10.3 | 5/24 | 3.6 | 7/65 | ||||
| ≥ 5 | 29.2 | 41/147 | 30.4 | 38/116 | ||||
| EVER SAME SEX PARTNER | ||||||||
| No | 27.3 | 56/203 | 21.5 | 53/202 | ||||
| Yes | 37.1 | 7/13 | 31.0 | 4/14 | ||||
| EVER TRANSACTIONAL SEX | ||||||||
| No | 27.3 | 54/196 | 17.7* | 38/169 | ||||
| Yes | 28.9 | 9/23 | 42.1 | 19/47 | ||||
| UNWANTED SEX PAST YEAR | ||||||||
| Yes | 23.7 | 11/26 | 31.4 | 17/50 | ||||
| No | 28.5 | 52/192 | 18.9 | 40/166 | ||||
| ANY CONTRACEPTION AT FIRST SEX± | ||||||||
| No | 29.6 | 16/60 | 22.7 | 23/76 | ||||
| Yes | 23.8 | 41/151 | 22.6 | 33/136 | ||||
| CONDOM AT FIRST SEX | ||||||||
| No | 40.7± | 21/63 | 19.0 | 26/81 | ||||
| Yes | 21.5 | 36/145 | 23.8 | 29/127 | ||||
| OTHER CONTRACEPTION AT FIRST SEX±± | ||||||||
| No | 27.5 | 37/142 | 19.1 | 34/147 | ||||
| Yes | 30.8 | 18/47 | 33.7 | 18/41 | ||||
| ALWAYS USE CONDOMS | ||||||||
| No | 36.7± | 49/133 | 31.9** | 35/107 | ||||
| Yes | 21.5 | 12/58 | 22.8 | 16/49 | ||||
| PERCEIVED SAFETY IN ENVIRONMENT | ||||||||
| Safe | 84.4*** | 117/146 | 71.9 | 105/138 | 23.4± | 31/117 | 16.7 | 25/105 |
| Unsafe | 90.9 | 103/113 | 64.1 | 110/133 | 31.8 | 31/103 | 27.3 | 32/110 |
| Ever had sexual intercourse - W Mean (SE) | Partner ever pregnant - W Mean (SE) | |||||||
| Yes | No | Yes | No | Yes | No | Yes | No | |
| PERCEIVED FEAR (scale)§ | 3.1 (0.14)* | 3.8 (0.25) | 4.9 (0.30) | 6.2 (0.75) | 4.9 (0.22)* | 2.4 (0.34) | 3.2 (0.05)*** | 5.4 (0.32) |
| COMMUNITY VIOLENCE (scale)§ | 7.3 (0.32)*** | 5.2 (0.18) | 9.1 (0.24) | 8.3 (1.13) | 9.8 (0.43)** | 6.3 (0.39) | 9.4 (0.64) | 9.0 (0.30) |
| PHYSICAL ENVIRONMENT (scale)§ | 11.4 (0.55)* | 14.4 (0.99) | 10.4 (0.31) | 10.6 (1.43) | 9.4 (1.26) | 12.1(0.53) | 7.7 (0.27)*** | 11.1 (0.30) |
Displaying weighted (W) row proportions using cluster and combined RDS and age post-stratification weights, followed by unweighted (U) observations by variable.
p<0.001,
p<0.01,
p<0.05 ± p<0.1.
Using Pearson’s corrected Chi Square test to compare the proportion ever-sex vs. never-sex (columns 1–2) and ever-pregnant vs. never-pregnant (columns 3–4) by characteristics (row percent).
Note: majority of sample unmarried (>98%).
Excluding currently pregnant females because the prevalence of drinking could be affected by being pregnant.
Age at first sex calculcated through survival analysis; differences determined by log-rank test
Before you were 15, who was the primary woman/man who raised you? Biological vs. none or other female/male figure
Neighborhood fear (0–18): higher score = more fear of moving around; Physical environment (0–24): higher score = better physical environment; Neighborhood violence (0–24): higher score = more violence observed
Condom or any other method for pregnancy prevention (Did you use something other than a condom to prevent pregnancy that last time, like birth control pills or something else (i.e. morning-after-pill, emergency contraceptives, patches, withdrawal, rhythm, injection, creams/suppository/jelly)?
Any other method than condom for pregnancy prevention
Table 4 summarizes the adjusted odds ratios of ever being pregnant among sexually experienced females and males in Baltimore and Johannesburg. Only variables that contributed to each site were included in each model by gender; thus, not all models included the same variables. The grey areas therefore mean that these variables were not included in the model for that site and sex. Among females in Baltimore, the odds of ever being pregnant were higher with binge drinking (eg. 5 or more drinks in a row) compared to no drinks in the past month (aOR=4.4, CI: 3.52, 5.46). Factors decreasing the odds of pregnancy were currently being in school (aOR=0.1, CI: 0.01, 0.6) and using condoms at first sex (aOR=0.2, CI: 0.05, 0.98). Among females in Johannesburg, factors associated with pregnancy were being raised by a single parent (aOR=18.1, CI: 10.4, 31.5) or by others (aOR=5.7, CI: 1.4, 23.6) compared to being raised by two parents, if the female reported unstable housing (aOR=4.7, CI: 2.5, 8.92) or had a sexual debut at 14 years or younger (aOR=7.4, CI: 3.4, 15.9). The odds of ever being pregnant were also found to increase with more observed neighborhood violence (aOR=1.2, CI: 1.1, 1.3) and, although marginally significant, with greater perceived fear of being robbed or attacked (aOR=1.1, CI: 0.99, 1.1). Factors associated with diminished odds of a pregnancy among females in Johannesburg were currently being in school (aOR=0.1, CI: 0.01, 0.8), using condom at first sex (OR=0.2, CI: 0.08,0.44), and better perception of their physical environment (OR=0.9, CI: 0.8, 0.9).
TABLE 4.
ADJUSTED PREVALENCE ODDS RATIO OF PREGNANCY FOR SELECTED FACTORS: BY GENDER AND SITE
| BALTIMORE FEMALES N=71 |
JOHANNESBURG FEMALES N=96 |
BALTIMORE MALES N=96 |
JOHANNESBURG MALES N=137 |
|||||
|---|---|---|---|---|---|---|---|---|
| aOR± | 95% CI | aOR | 95% CI | aOR | 95% CI | aOR | 95% CI | |
| AGE (cont.) | 1.2 | 0.87, 1.62 | 1.1 | 0.59, 1.73 | 1.2 | 0.77, 1.77 | 1.5** | 1.25, 1.77 |
| CURRENTLY IN SCHOOL | 0.1* | 0.01, 0.59 | 0.1* | 0.01, 0.79 | 0.8 | 0.22, 2.92 | 0.4*** | 0.37, 0.44 |
| Ref: not in school | ||||||||
| RAISED BY ONE PARENT | 18.1*** | 10.37, 31.53 | ||||||
| RAISED BY OTHER ADULTS | 5.7* | 1.36, 23.67 | ||||||
| Ref: raised by two parents | ||||||||
| UNSTABLE HOUSING | 4.7** | 2.47, 8.92 | ||||||
| Ref: Stable housing | ||||||||
| < 5 DRINKS IN A ROW PAST 30 DAYS | 0.2 | 0.03, 1.00 | 2.7** | 1.68, 4.26 | 0.8 | 0.49, 1.16 | ||
| ≥ 5 DRINKS IN A ROW PAST 30 DAYS | 4.4*** | 3.52, 5.46 | 4.6** | 2.38, 9.09 | 1.0 | 0.77, 1.18 | ||
| Ref: no drink past 30 days | ||||||||
| EARLY SEXUAL DEBUT (<15) | 1.3 | 0.46, 3.51 | 7.4** | 3.39, 15.93 | 1.4** | 1,12, 1.65 | 5.4*** | 4.34, 6.69 |
| Ref: first sex age 15 or above | ||||||||
| EVER TRANSACTIONAL SEX | 1.7* | 1.08, 2.74 | ||||||
| Ref: never transactional sex | ||||||||
| CONDOM AT FIRST SEX | 0.2* | 0.05, 0.98 | 0.2** | 0.08, 0.44 | 0.7 | 0.11 4.27 | ||
| Ref: No condom at first sex | ||||||||
| ALWAYS USE CONDOMS | 0.6 | 0.15, 2.09 | 1.3 | 0.90, 1.88 | ||||
| Ref: do not always use condoms | ||||||||
| PERCEIVED FEAR (scale)§ | 1.1 | 0.99, 1.14 | 1.1 | 0.97, 1.2 | 0.9*** | 0.83, 0.89 | ||
| COMMUNITY VIOLENCE (scale)§ | 1.2** | 1.12, 1.26 | 1.1* | 1.02, 1.12 | ||||
| PHYSICAL ENVIRONMENT (scale)§ | 1.0 | 0.86, 1.06 | 0.9* | 0.78, 0.95 | 0.9*** | 0.88, 0.94 | ||
p<0.001,
p<0.01,
p<0.05.
Note: ever pregnant does not include females currently pregnant. aOR=adjusted Odds Ratio.
Neighborhood fear: higher score = more fear of moving around Neighborhood violence (0–24): higher score = more violence observed; Physical environment (0–24): higher score = better physical environment;
aOR=Adjusted Odds Ratio
Among sexually experienced males in Baltimore, the odds of a partner ever being pregnant were higher if they reported binge drinking in the past month (aOR 4.6, CI: 2.4, 9.1) or drinking any alcohol (aOR=2.7, CI: 1.7, 4.3) compared with males who reported no drinks in past month. The odds of impregnating a partner were also higher with more observed violence in the community (aOR=1.1, CI: 1.0, 1.1); and, although marginally significant, also with greater perceived fear of being robbed or attacked (aOR=1.1, CI: 1.0, 1.2). Finally, among sexually experienced males in Johannesburg, the odds of a partner ever being pregnant were higher for incremental increases in his age (aOR=1.5, CI: 1.2, 1.8), an early sexual debut (aOR=5.4, CI: 4.3, 6.7), and ever engaging in transactional sex (aOR=1.7, CI: 1.1, 2.7). Factors associated with decreased odds of a partner ever being pregnant were currently in school (aOR=0.4, CI: 0.4, 0.4), greater perceived fear of being robbed or attacked aOR=0.9, CI: 0.8, 0.9), and better perceived physical environment (aOR=0.9, CI: 0.9, 0.9).
Discussion
The goal of our study was to assess the prevalence of pregnancy and explore associated factors among adolescents across five, resource poor urban settings. Our results show that sexual and pregnancy experiences were highly prevalent among adolescents in Baltimore and Johannesburg – a finding that is consistent with the qualitative phase of the WAVE study.14 Among Baltimore and Johannesburg participants, school was found to decrease the odds of sexual activity and pregnancy for both males and females, as was condom use at first sex. Some covariates of experiencing a pregnancy were early sexual debut (< 15 years), being raised by single parent or someone other than parent, alcohol use and binge drinking in the past month, and environmental factors such as greater violence in the community and lower scores for physical environment which correlated with unhygienic and crowded neighbourhood with no recreational spaces for adolescents.
In contrast, sexual experience and consequently also pregnancy was rare among adolescents in Ibadan, Delhi and Shanghai. This could be a result of the majority of participants being unmarried, and possibly that premarital sex in these settings is relatively less common.2 For example, the legal age at of marriage in China is 20 years for females and 22 years for males, making it unlikely that married adolescents would have been recruited for this study. So too, it is possible that some adolescents underreport premarital sexual activity.15 In contrast, sexual experience and pregnancy more commonly occur outside the context of marriage in Baltimore and Johannesburg,3,16,17 and hence the report of pregnancies in these settings was significantly higher. Furthermore, adolescents in Ibadan, Delhi and Shanghai who reported ever having sex were more likely than peers to report not being in school, being raised by a single parent or other adult, binge drinking, and perceiving higher levels of community violence and lack of safety. More research is needed to understand the true prevalence of sexual experience and pregnancy as well as unique vulnerabilities among unmarried adolescents in these settings.
Interestingly, despite varying sexual and pregnancy experiences, reported abortion proportions were high across settings although it is illegal. Almost 60% of unsafe abortions in Africa are among women younger than 25 years and a quarter are adolescents aged 15–19 which might potentially be at higher risk for morbidity and mortality from unsafe abortion.18,19,19,20
In our Baltimore sample, half of females (53% weighted) reported ever being pregnant, which is substantially higher than elsewhere in the country.21 Similarly, in Johannesburg almost a third of the sexually active adolescent females reported a pregnancy (29% weighted), which is well above the South African adolescent pregnancy prevalence of about 12%.3 Our findings confirm many of the determinants of adolescent pregnancy found in national level studies both in South Africa and the US: school dropout, being raised by a single parent, high levels of substance use, early sexual debut, lack of contraception at first sex and neighborhood crime and violence.3,22–25
Pregnancy experienced by a partner was reported by about a quarter of sexually experienced males in Baltimore and Johannesburg. Research has shown that the risk profile of young fathers is similar to adolescent mothers in that they have higher school dropout rates, lower school performance, come from low income households, are less likely to have the resources to support the child,16 and have higher rates of unemployment and earning potential.26,27 These conditions suggest that addressing early pregnancies among this population will require attention to both individual as well as contextual factors, with a focus on adolescent males and females.
Our finding that being in school was associated with lower odds of pregnancy is consistent with the literature on schooling as an important protective factor for delaying pregnancy, particularly among girls.2,12,28,29 Studies have shown that girls who remain in school are less likely to engage in sexual activity and become pregnant12,30 and if they do engage in sex, they are more likely to use condoms consistently.7 However, in our study it is not possible to determine the direction of effects because of its cross-sectional nature. Those not in school may have dropped out as a consequence of getting pregnant.
Being raised by a single parent or by someone other than your parents appeared to have the most severe negative impact (higher odds of pregnancy) for females in Johannesburg. This is in line with family structure being a key risk factor for adolescent risk behaviours.24 Both in the United States25 and South Africa3, studies have shown that growing up with an absent father is associated with elevated risk of early sexual activity and adolescent pregnancy.23 The link between alcohol use and sexual risk behaviours is well documented,31 and high-consumers of alcohol are seen as a critical target group for HIV as well as pregnancy prevention. In the current study, past month binge drinking was found to increase the odds of pregnancy among males and females in Baltimore, and interestingly the odds were also higher among males reporting any drinking in the past month. Studies in Baltimore have linked high-school drop-out to heavy drinking and highlighted the importance of including education completion as a critical component of alcohol treatment interventions.32 In urban environments characterized by multiple disadvantages, substance use significantly increases the risk profile of adolescents, and often becomes both a cause as well as escape from the hardships imposed by urban poverty.
Early age of sexual debut was a significant determinant of pregnancy among both males and females in Johannesburg and males in Baltimore – a finding consistent with other studies3,22 where sexual initiation at or before age 14 also has been correlated with lack of contraceptive use as well as higher rates of HIV and STIs.33,34 Condom use at first sex significantly decreased the odds of a pregnancy among females in both Baltimore and Johannesburg. Given the high rates of HIV and STIs in both Baltimore22 and Johannesburg35, integration of HIV/STI and reproductive health services and tailoring these services to meet the needs of adolecents is critical.
Environmental factors have a significant impact on the health and well-being of adolescents.36,37 In our study, higher levels of observed violence in the environment increased the odds of a pregnancy among females in Johannesburg and males in Baltimore. This is consistent with how neighbourhoods with higher levels of violence, crime and poverty have been shown to increase adolescent sexual risk behaviours.36,38 Both Johannesburg and Baltimore are characterized by housing instability, densely populated neighbourhoods, abandoned buildings occupied illegally, unclean neighbourhoods, and few green and recreational spaces for adolescents. In the qualitative study14 the piling up of trash, inadequate and crowded housing, lack of basic services such as water and electricity were major health concerns among adolescents. More research is thus needed to understand the impact of physical environment on adolescent pregnancy.
Limitations
Given the cross-sectional nature of our study, it was not possible to tease out the causality of some of the associations from our study. In addition, the length of the instrument precluded in-depth questioning. Since we used RDS to recruit our study participants relying on peer recruitment, most of the individuals in our sample were unmarried adolescents. This approach may have led to underestimations of the prevalence of sexual activity and pregnancy particularly among adolescents in Ibadan, New Delhi and Shanghai. Given the low levels of reported sexual experience in Ibadan, New Delhi and Shanghai our study was underpowered to examine the differences in pregnancy outcomes across all the cities.
Implications and Contribution
Prevalence of sexual experience and adolescent pregnancy varies across settings and by gender. While our study found that some individual and environmental level factors were associated with adolescent pregnancy across sites, the influence of other factors was unique to certain settings. Being in school was strongly related to pregnancy in both Baltimore and Johannesburg, whereas alcohol use was significant only in Baltimore. On the other hand, being raised by a single parent, or living in an unstable housing situation were far more important correlates of pregnancy among females in Johannesburg. Among youth living in very impoverished neighbourhoods across the five cities, the adolescents showed varying levels of sexual and reproductive health needs suggesting that prevention approaches need to tailor their approaches to the unique needs of their targeted communities.
Acknowledgements
This research was supported by Young Health Programme, a partnership between AstraZeneca, Johns Hopkins Bloomberg School of Public Health and Plan International, a leading global children’s charity. In Ibadan, the study was funded by The Bill and Melinda Gates Institute at Johns Hopkins Bloomberg School of Public Health through its funding to The Centre for Population and Reproductive Health, University of Ibadan. We would also like to thank Lawrence Mashimbye and Harry Moultrie for assistance with data collection and management.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Reference List
- 1.WHO. Adolescent Pregnancy. Fact Sheet 364. 2012
- 2.UNFPA. Motherhood in Childhood-Facing the Challenge of Adolescet Pregnancy. The State of World Population 2013. 2013
- 3.Panday S, Makiwane M, Ranchod C, Letsoalo T. Teenage Pregnancy in SouthAfrica – With a Specific Focus on School-Going Learners. Pretoria: Department of Basic Education; 2009. Child, Youth, Family and Social Development. Human Sciences Research Council. [Google Scholar]
- 4.Hofferth SL, Reid L. Early childbearing and children's achievement and behavior over time. Perspect Sex Reprod Health. 2002;34:41–49. [PubMed] [Google Scholar]
- 5.Hofferth SL, Reid L, Mott FL. The effects of early childbearing on schooling over time. Fam Plann Perspect. 2001;33:259–267. [PubMed] [Google Scholar]
- 6.Mathews TJ, MacDorman MF. Infant mortality statistics from the 2006 period linked birth/infant death data set. Natl Vital Stat Rep. 2010;58:1–31. [PubMed] [Google Scholar]
- 7.Blum RW, Nelson-Mmari K. The health of young people in a global context. J Adolesc Health. 2004;35:402–418. doi: 10.1016/j.jadohealth.2003.10.007. [DOI] [PubMed] [Google Scholar]
- 8.Mosher WD, Deang LP, Bramlett MD. Community environment and women's health outcomes: contextual data. Vital Health Stat 23. 2003:1–72. [PubMed] [Google Scholar]
- 9.Ford J, Browning C. Neighborhoods and infectious disease risk: acquisition of chlamydia during the transition to young adulthood. J Urban Health. 2014;91:136–150. doi: 10.1007/s11524-013-9792-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Grieb SM, vey-Rothwell M, Latkin CA. Housing stability, residential transience, and HIV testing among low-income urban African Americans. AIDS Educ Prev. 2013;25:430–444. doi: 10.1521/aeap.2013.25.5.430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Schonlau M, Liebau E. Respondent Driven Sampling. The Stata Journal. 2012;12:72–93. [Google Scholar]
- 12.Biddlecom A. Associations between premarital sex and leaving school in four Sub-Saharan African Countries. Studies in Family Planning. 2008;39:337–350. doi: 10.1111/j.1728-4465.2008.00179.x. [DOI] [PubMed] [Google Scholar]
- 13.Lloyd C. Social Determinants of Sexual and Reproductive Health. Informing future research and programme implications. Geneva: WHO; 2010. The role of schools in promoting sexual and reproductive health among adolescents in developing countries. Shawn Malarcher; pp. 113–132. [Google Scholar]
- 14.Mmari K, Blum R, Sonenstein F, et al. Adolescents' perceptions of health from disadvantaged urban communities: Findings from the WAVE study. Soc Sci Med. 2014;104:124–132. doi: 10.1016/j.socscimed.2013.12.012. [DOI] [PubMed] [Google Scholar]
- 15.Darroch JE, Frost JE, Singh S. Teenage Sexual and Reproductive Health Behavior in Developing Countries. New York & Washington DC: Allan Guttmacher Institute; 2001. [Google Scholar]
- 16.Guttmacher Institute. Facts on American Teens' Sexual and Reproductive Health. New York: 2013. [Google Scholar]
- 17.Marteleto L, Rancchod V. Sexual Behavior, Pregnancy, and Schooling among Young People in Urban South Africa. Studies in Family Planning. 2008;39:351–368. doi: 10.1111/j.1728-4465.2008.00180.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.WHO. Unsafe Abortion: Global and Regional Estimates of the Incidence of Unsafe Abortion and Associated Mortality in 2003. WHO; 2007. [Google Scholar]
- 19.Mesce D. Population Reference Burea. Washington DC: 2011. Abortion: Facts and Figures. [Google Scholar]
- 20.WHO. Issues in Adolescent Health and Development. Geneva: WHO; 2004. Contraception in Adolescence. [Google Scholar]
- 21.Martin JA, Hamilton BE, Osterman MJK, Curtin SC, Mathews TJ. National Vital Statistics Reports. 9. Vol. 62. CDC; 2013. Births: Final Data for 2012. [PubMed] [Google Scholar]
- 22.Urban Health Institute. Reducing Teen Births in Baltimore City. Johns Hopkins Urban Institute; 2012. [Google Scholar]
- 23.Miller BC. Family influences on adolescent sexual and contraceptive behavior. J Sex Res. 2002;39:22–26. doi: 10.1080/00224490209552115. [DOI] [PubMed] [Google Scholar]
- 24.Bonell C, Allen E, Strange V, et al. Influence of family type and parenting behaviours on teenage sexual behaviour and conceptions. J Epidemiol Community Health. 2006;60:502–506. doi: 10.1136/jech.2005.042838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ellis BJ, Bates JE, Dodge KA, et al. Does father absence place daughters at special risk for early sexual activity and teenage pregnancy? Child Dev. 2003;74:801–821. doi: 10.1111/1467-8624.00569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gee CB, Rhodes JE. Adolescent mothers' relationship with their children's biological fathers: social support, social strain, and relationship continuity. J Fam Psychol. 2003;17:370–383. doi: 10.1037/0893-3200.17.3.370. [DOI] [PubMed] [Google Scholar]
- 27.Jaffee SR, Caspi A, Moffitt TE, Taylor A, Dickson N. Predicting early fatherhood and whether young fathers live with their children: prospective findings and policy reconsiderations. J Child Psychol Psychiatry. 2001;42:803–815. doi: 10.1111/1469-7610.00777. [DOI] [PubMed] [Google Scholar]
- 28.Chaaban J, Cunningham W. Measuring the Economic Gain of Investing in Girls: The Girl Effect Dividend. Washington DC: World Bank; 2011. [Google Scholar]
- 29.UNFPA. State of World Population 2012: By Choice, Not by Chance: Family Planning, Human Rights and Development. New York: 2012. [Google Scholar]
- 30.Perper K, Peterson K, Manlove J. Child Trends, Fact Sheet. Washington DC: CDC; 2010. Diploma Attainment Among Teen Mothers. [Google Scholar]
- 31.Woolf-King SE, Maisto SA. Alcohol Use and High-Risk Sexual Behavior in Sub-Saharan Africa: A Narrative Review. Arch Sex Behav. 2011;40:42. doi: 10.1007/s10508-009-9516-4. [DOI] [PubMed] [Google Scholar]
- 32.Lillie-Blanton M, MacKenzie E, Anthony JC. Black-white differences in alcohol use by women: Baltimore survey findings. Public Health Rep. 1991;106:124–133. [PMC free article] [PubMed] [Google Scholar]
- 33.Manzini N. Sexual initiation and childbearing among adolescent girls in KwaZulu Natal, South Africa. Reprod Health Matters. 2001;9:44–52. doi: 10.1016/s0968-8080(01)90007-2. [DOI] [PubMed] [Google Scholar]
- 34.Mmari K, Sabherwal S. A review of risk and protective factors for adolescent sexual and reproductive health in developing countries: an update. J Adolesc Health. 2013;53:562–572. doi: 10.1016/j.jadohealth.2013.07.018. [DOI] [PubMed] [Google Scholar]
- 35.South African National AIDS Council. South Africa: HIV Epidemic, Responses and Policy Synthesis; 2011. [Google Scholar]
- 36.Carlson DL, McNulty TL, Bellair PE, Watts S. Neighborhoods and Racial/Ethnic Disparities in Adolescent Sexual Risk Behavior. J Youth Adolesc. 2013 doi: 10.1007/s10964-013-0052-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Sawyer SM, Afifi RA, Bearinger LH, et al. Adolescence: a foundation for future health. Lancet. 2012;379:1630–1640. doi: 10.1016/S0140-6736(12)60072-5. [DOI] [PubMed] [Google Scholar]
- 38.Cubbin C, Santelli J, Brindis CD, Braveman P. Neighborhood context and sexual behaviors among adolescents: findings from the national longitudinal study of adolescent health. Perspect Sex Reprod Health. 2005;37:125–134. doi: 10.1363/psrh.37.125.05. [DOI] [PubMed] [Google Scholar]