Untangling the Evidence: A review of teenage pregnancy research and statistics in South Africa

Ibis Reproductive Health 12 October 2014

This work was generously supported by the Ford Foundation’s Office for Southern Africa.

Acknowledgements This work was generously support by the Ford Foundation’s Office for Southern Africa. Naomi Lince, Alicia Flanagan and Tshego Bessenaar contributed to the writing and content. Introduction Teen pregnancy is a global issue. It disproportionately affects young women around the globe on many levels and may limit future opportunities for teens. Globally, the teen birth rate is declining; however significant regional disparities persist. In 1990, the global fertility rate (births per 1,000 adolescents) was 60 per 1,000. In 2007 the rate decreased to 48 births per 1,000 adolescents. In 2008, 95% of teen births worldwide occurred in middle- and low- income countries. Global adolescent fertility rates range from 121 per 1,000 in sub-Saharan Africa to 5 per 1,000 in Eastern Asia (WHO, 2011). South Africa has the lowest overall fertility rate of all sub-Saharan African countries at a rate of 19, 21 or 22 births per 1,000 women age 15 to 49, depending on the source providing the data (Kaiser Family Foundation, 2012; Panday et al., 2009; PRB, 2012; UNFPA, 2010). Young people under age 18 make up 50% of the population in South Africa (Reddy et al., 2010). Teen pregnancy undermines future opportunities for many women in the population (Panday et al., 2009). All young women who experience pregnancy may be subject to social disadvantages due to gender-based discrimination and violence, as well as limited growth opportunities in education and occupation. Teen pregnancy may result in psychological harm because so many young women who become pregnant experience sexual coercion. Irrespective of the context in which women become pregnant, there are psychological, social, and physical risks associated with teenage pregnancy and childbirth, including morbidity and mortality. Maternal mortality is five times higher for girls under 15 years of age, and it is doubled for girls age 15/16 to 19, compared to women over age 20 (Nour, 2006; Shaw, 2009). Environmental factors limit the ability for teens to access health care, thus contributing to the health complications associated with early childbearing among teens. Adolescents are less likely than older women to have prenatal care (Panday et al., 2009), and a 2012 study by Holt et al. reported that South African health care workers have judgmental attitudes about sexually active adolescents. Some health care workers reported that adolescents should not have sex before marriage. One fifth of the health care workers reported that abstinence is the preferred method of contraception. The study found that all health care workers need additional training on modern forms of contraception to be able to provide comprehensive family planning counseling (Holt et al., 2012). Of additional concern, adolescents in Sub-Saharan Africa are at high risk for choosing unsafe abortion, provided by clandestine means (Varga, 2003). The high prevalence of unsafe abortion can be attributed to lack of information about the right to access

2

safe and legal abortion services, barriers to access, and social and cultural beliefs regarding the provision of legal abortion services. Varga (2003) reported forty-five percent of rural teens, and 64% of urban teens believed that pregnant girls their age would choose to terminate their pregnancies, and most teens who wanted to end a pregnancy were likely to do so using clandestine (or “backstreet”) means. Teen pregnancy and early childbirth are concerning public health problems that deserve unbiased attention from health professionals, policy makers, educators, and community members (P.T. Mngadi et al., 2002). Reddy et al., (2010) state, “health-risk behaviors once established in adolescence, often persist into adulthood and may have serious consequences such as… psycho-social problems, unwanted pregnancies and infectious diseases, such as sexually transmitted infections, including HIV and AIDS,” (Reddy et al., 2010). To better understand the problems associated with teen pregnancy in South Africa, it is imperative to have good data with which to study the prevalence and detriments. There are limited statistics on teenage pregnancy in South Africa (Panday, Makiwane, Ranchod, & Letsoala, 2009), and among the available data, there are inconsistencies concerning the teenage fertility rate. Fertility rates refer to the number of live births per 1,000 women and do not include pregnancies that result in stillbirth, miscarriage or termination. The inconsistencies in available data may be attributed to variations in the studies that have been conducted, such as different study aims, methodologies, and populations. Given that reducing teenage pregnancy in the country is a priority, examining the variations in the data is critical for understanding the current situation, formulating recommendations for interventions and future attempts to measure progress. The purpose of this review is to compare and contrast the design and objective(s) of recent studies that present data on teenage pregnancy to highlight the strengths and limitations of each report in terms of its data on teenage pregnancy, and to provide guidance on how the studies may or may not be compared to determine trends over time. Methods Between March-July 2011 and December 2012 and February 2013, several research databases [Academic OneFile (Gale), Access World News, Africa Bibliography (Cambridge), Gender Studies Database (EBSCO), Health Reference Center (GALE), PsycINFO, and PubMed (Medline)] were used to locate articles for possible inclusion in a literature review that aimed to describe the prevalence, contributing factors and consequences of teenage pregnancy in South Africa. Data from national and international studies, as well as governmental and nongovernmental reports were included. Books and dissertations/theses, and well as news media were excluded. There was not a rigorous systematic quantitative process for interpreting research

3

relevant to teenage pregnancy, and thus all relevant publications may not have been reviewed. A convenience sample of s studies that aimed to contain nationally representative data on adolescent pregnancy in South Africa that were frequently cited in the literature review were included in this paper. Results In the end, six large studies presenting data on teen pregnancy were selected for inclusion (see Table 1.0). All were conducted in the last 15 years (and were published since 2000), and all are often cited in South Africa as indicative of current trends or the prevalence of teenage pregnancy. All aim to be nationally representative, and while each contains valuable information regarding the prevalence and determinants of teenage pregnancy and/or fertility, only one paper focused exclusively on adolescent fertility. Panday et al. (2009) completed a desktop literature review on adolescent fertility, with a focus on school-going learners, age 15-19. The other papers provide original empirical data on adolescent fertility and pregnancy from studies that aimed to research overall health, HIV infection, or risk behaviors (including risky sexual behaviors).



4

Table 1.0: Listing of studies Study authors

Title

Purpose

Department of Health, Medical Research Council (2007)

2003 South Africa Demographic and Health Survey

- Provide reliable, nationally representative basic health data∞ - Assist policymakers in evaluating and designing programs / strategies for improving national health services

Reddy et al (2003)

The First South African National Youth Risk Behaviour Survey 2002

- Establish the prevalence of risk behaviors⌘ and project how behaviors change - Provide early warning signs for future health epidemics

Reddy et al (2010)

The Second South African Youth Risk Behaviour Survey 2008

- Provide data on priority risk behaviors‡ that impact adolescent health and future morbidity and mortality

Panday, Makiwane, Ranchod, & Letsoala (2009)

Teenage Pregnancy in South Africa: with a specific focus on school-going learners

- Review existing literature† - Establish the prevalence and determinants of teenage pregnancy; - Assess the individual, familial and educative impact of teenage pregnancy; - Identify and assess the impact of interventions for teenage pregnancy; and - Propose a conceptual framework for research and interventions to prevent and mitigate the impact of teenage pregnancy.

Pettifor et al (2005)

Young people’s sexual health in South Africa: HIV prevalence and sexual behaviors from a nationally representative household survey

-- To determine the prevalence of HIV infection,HIV risk factors,and exposure to national HIV prevention programs - To identify factors associated with HIV infection among South African youth, aged 15–24 years

Department of Health, Medical Research Council (2002)

1998 South Africa Demographic and Health Survey

- Provide reliable,nationally representative basic health data* - Assist policymakers in evaluating and designing programs / strategies for improving national health services

Data collection method

Nationally representative two-stage sample from 1996 Census of 690 urban and 282 non-urban sampling units. Ten households were selected in each urban area with 20 households per non-urban area, for a total of 12,540 households. All women age 15-49 completed a Household Questionnaire and Women’s Questionnaire. Male and female adults over age 15 were selected to complete the Adult Health Questionnaire at every second household. Interviewers captured approximately 12,000 women and 13,500 adults for an n of approximately 17,500. Nationally representative two-stage sample from 2001 Census with a total of 10,214 households selected from 368 urban and 262 non-urban sampling units. With a response rate of 84.5, 7,756 households interviewed. 7,041 women and 3,118 men completed the questionnaire. Male and female adults over age 15 were selected to complete the Adult Health Questionnaire at every second household. Stratified,two-state cluster sample of public school learners grades 8,9,10,11 in nine provinces. 207 schools [27/ province] selected in first stage sampling, 2 classes per school selected in the second stage. All classes in Eastern Cape public schools were sampled due to lower enrolment. Stratified,two-stage cluster sample crosssectional national prevalence study of public school learners grades 8,9,10,11 in nine provinces. 207 schools [27/ province] selected in first stage sampling, 2 classes per school selected in the second stage, with oversampling of Indian students from 21 schools with Indian learner enrolment greater than 50%. 1200 learners per province identified for a self-administered 135-question instrument. Desktop review of literature published from 2000-2008 focused on teenage pregnancy and fertility. Secondary analysis of 2003 Human Sciences Research Council Status of Youth Survey of 3500 young people age 18-35. Purpose of secondary analysis was to identify factors associated with early pregnancy. The 2003 HSRC Youth Survey is a self-weighting national sample drawn from the 2001 Census, and interviewers administered a face-to-face questionnaire to one person age 18 to 35 per household. Disproportionate stratified design from 2001 National Census sampling frame weighted to represent a nationally representative distribution of young people age 15-24. Inclusion criteria required participants to provide an oral specimen for HIV testing and complete face-to-face interview.

*Information on: fertility and childhood mortality levels; fertility preferences; awareness and use of contraceptive methods; breastfeeding practices; maternal and child health; awareness of HIV/AIDS; chronic health conditions among adults; dental health; lifestyle habits; adult height, weight, blood pressure ∞ In addition to data collected in 1998 SADHS, information on: reproductive health and sexual behavior of men; malaria; pensions/grants received by members of the household; infant and child feeding; adolescent health; mortality and morbidity in adults; utilization of health services; hypertension, chronic pulmonary disease and Asthma; risk factors for chronic diseases; oral health; child height, weight, blood pressure **Published peer-reviewed journal articles, conference presentations, reports, book chapters, abstracts on prevalence, determinants, and interventions for teenage pregnancy and fertility ⌘ Risk behaviors: injuries, substance abuse, sexual behaviors, nutrition, physical activity, hygiene † This included a summary of data drawn from previously published studies (i.e. Human Sciences Research Council (HSRC) 2003 Status of Youth Survey, DHS data from 1998, and South African Census data) as well as secondary analysis of the HSRC 2003 study data and primary analysis of data drawn from the Education Management Information System (EMIS). ‡ Behaviors related to: infectious diseases; chronic disease; injury and trauma; mental health

In addition to the national studies, which are the main focus of this review, we also briefly describe two additional studies as they are helpful in providing extra information on teenage pregnancy in the country. They are: Human Capital Consequences of Teenage Childbearing in South Africa (Karra & Lee, 2012) (which is based on data collected in the Cape Town Area Study (CAPS) (University of Michigan & University of Cape Town, 2012) and data from the Africa Centre Demographic Information System (Africa Centre, 2013) Study aims Each of the studies had different objectives. Both of the South Africa Demographic and Health Surveys (SADHS) (1998, 2003) aimed to provide reliable, nationally representative basic health information on maternal, child, and adult health indicators. The 1998 SADHS was

5

initiated to ensure that the health policy and legislation of the 1994 democratic elections in South Africa were meeting the objectives of addressing the inequalities in health care distribution. Its aim was to assess whether all South Africans were achieving greater access and quality of health benefits, and to provide indicators of achievements as well as deficiencies of the new programs (Department of Health et al., 2002). Like the 1998 SADHS, the purpose of the 2003 SADHS was to provide reliable, nationally representative basic health data. Its objective was to track the changes in the health status of the South African population in the five-year period since the 1998 survey and to provide more information for policymakers and program managers. A secondary aim of the 2003 SADHS survey was to expand the study with additional surveys to monitor achievements and deficiencies of new governmental programs, as well as to collect more demographic data than the 1998 SADHS. The 2003 survey collected data on men and children, whereas the 1998 SADHS did not(Department of Health et al., 2007). The 2003 SADHS experienced fieldwork problems in KwaZulu-Natal that significantly impacted the reliability of national fertility data. The survey found the fertility in the province to be about 25 percent lower than that observed anywhere in the world (DHS, RSA Medical Research Council, 2007). The errors found in the 2003 Demographic and Health Survey contribute to the significant decline in fertility since 1998 SADHS (74 per 1,000). The DHS states, “for all practical purposes, the data on children ever born, living or dead from the 2003 SADHS... cannot be used to form a reliable opinion on childbearing, fertility, child survival or any of their correlates in the five years before the survey” (DHS, RSA Medical Research Council, 2007). Despite this, data from the 2003 SADHS is included in this review. The National Department of Health commissioned the Medical Research Council to conduct the first national survey of public secondary school learner risk behavior since South Africa had a democratic school system. Reddy et al. (2003) aimed to provide provincially and nationally representative health data on learner risk behaviors to inform intervention development and health policy development. The behaviors studied included intentional and unintentional injury, mental health, substance abuse, sexual behavior, nutrition, physical activity, and personal hygiene. The Medical Research Council (MRC) conducted a second study on learner risk behavior, also authored by Reddy et al., published in 2010. The second Youth Risk Behavior Survey had the same aim as the first: to provide provincially and nationally representative crosssectional data regarding risk behaviors among learners in grades 8-11 that impact adolescent health and future morbidity and mortality (Reddy et al., 2010). The Second 2008 YRBS was designed to gather evidence-based data on the population of young people to inform public

6

policy and to inform interventions that aim to reduce public health “epidemics” (Reddy et al., 2010). The areas of study included youth behaviors related to: infectious disease, chronic disease, injury and trauma, and mental health. Data on teenage sexual behaviors – including the prevalence of pregnancy and fertility – were included in the study of infectious disease. The report led by the Department of Basic Education (DBE) and the Human Sciences Research Council (HSRC) included a comprehensive literature review derived from local and international studies in order to provide an overview of research on the prevalence, contributing factors, and interventions for teenage pregnancy. The aim of the DBE report was to assess the individual, familial and educational impacts of teenage pregnancy among school-going adolescents by reviewing existing literature and assessing current interventions, again using existing data. (Panday et al., 2009). The Reproductive Health Research Unit (RHRU) conducted a nationally representative survey of young peoples sexual health in South Africa, in 2003. The objective of this survey was to “determine the prevalence of HIV infection, HIV risk factors, and exposure to national HIV prevention programs, and to identify factors associated with HIV infection among South African youth, aged 15–24 years” (Pettifor et al., 2005, p. 1525). Therefore, while this study contains valuable nationally representative data on teenage pregnancy and fertility, the study aim was focused on HIV prevalence. Many of the statistics in the paper pertain to the study objectives, thus, there is less detail regarding teen pregnancy. It is included herein because young women are at the greatest risk for HIV infection; the combination of high HIV prevalence and teenage pregnancy highlights the importance of dual protection and highlights an unmet need for contraception and comprehensive sexuality education in South Africa (Pettifor et al., 2005, p. 1531). Lastly, as noted above, although not nationally representative, we include in this review additional comments regarding teenage childbearing data drawn from data in the Africa Centre Demographic Information System and the Cape Town Area Study, which are highlighted in the 2012 Population Reference Bureau (PRB) Brief on Human Capital Consequences of Teenage Childbearing in South Africa. This research brief describes findings between teen fertility, educational attainment, and health outcomes in urban and rural regions in South Africa (Karra & Lee, 2012). The Cape Area Panel Study aimed to follow a representative sample of adolescents in Cape Town during transition into adulthood and focus on schooling, employment, sexual and reproductive health, and familial supports. The Africa Centre Study aimed to gather household demographic surveillance data in KwaZulu-Natal to track the impact of the HIV epidemic (Africa Centre, 2013). The Cape Town Area Study and the Africa Centre Study both offer longitudinal, regional data.

7

Methods, study population(s) and sampling The 1998 SADHS conducted a national survey of 12,540 households in all 9 provinces of South Africa during the study year. The SADHS selected units in each province from the 1996 census data, and then stratified them by rural and urban areas. The SADHS identified 690 urban units and 282 rural units. Then, 10 households were identified per urban unit, and 20 households were identified per rural unit for the household survey. The design was non-self weighted to: (1) increase the sampling rate for smaller provinces (Northern Cape, Free State, Eastern Cape); (2) increase the sampling rate in Gauteng and KwaZulu-Natal to capture more data on Indian/Asian households. Every second household within the selected unit was chosen to participate in the adult health survey. Therefore, in addition to selecting all women age 15-49 per household, all adults at every other household were selected to participate. Approximately 17,500 people were interviewed, with a response rate above 95% (Department of Health et al., 2002). The 2003 SADHS targeted 10,214 households in all 9 provinces of South Africa with comprehensive surveys for women age 15 to 49, and children age 5 and younger. Every second household was selected to survey men age 15 to 59, as well as to survey adults age 15 and older. Interviews were conducted from 2003 to 2004. The 2003 SADHS had a response rate of 85% and interviewed 7,756 people, which is lower than the 1998 SADHS response rate. The 2003 SADHS response rate varied by region and gender; more women responded to the survey (75% response rate for females vs. 67% response rate for males) (Department of Health et al., 2007). As noted, the 2003 SADHS had problems with data collection, reducing the reliability of results pertaining to adolescent pregnancy and fertility. The DBE report documented, reviewed, and analyzed literature on teenage pregnancy, and focused on teens attending school. Literature dated between 2000 and 2008 and focused on teenagers between ages 15 to 19. Panday et al. (2009) used six public research databases (Ebscohost, ProQuest, ScienceDirect, SABINET, African Health link and African Journals online) to obtain the literature (Panday et al., 2009). After collecting literature regarding teenage pregnancy, Panday et al. (2009) also focused on four seminal papers regarding the prevalence of HIV among South African youth, and reviewed policy documentation regarding adolescent health in South Africa. Lastly, the authors used Cox regression analysis with the EMIS and HRSC South African National HIV Prevalence, HIV Incidence, Behavior and Communication Survey data to examine the links between early pregnancy (the dependent variable) and family structure, type of childhood residence, childhood poverty, and school dropout (independent variables).



8

The DBE report also reviewed seminal works pertaining to the prevalence of HIV among young South Africans: 

The 2003 Reproductive Health Research Unit survey on HIV and sexual behavior of youth aged 15-24 years (Pettifor et al., 2005)1



The 2006 SABC and Kaiser Family Foundation survey that focused on HIV awareness, sexual behavior and broadcast media among youth aged 15-24 years (Kaiser Family Foundation & SABC, 2006)



The 2002 and 2005 HSRC South African National HIV Prevalence, HIV Incidence, Behavior and Communication Surveys (Shisana & Simbayi, 2002; Shisana et al., 2005) Therefore, the DBE does not provide data from an original empirical study. The DBE conducted secondary analysis on the pregnancies among school-going

adolescents reported through the Education Management Information System (EMIS). The EMIS is a unit of the DBE, and it gathers national census data on education information, including data on the health of young people attending public and independent schools during the academic year (EMIS, 2012). The DBE report on teenage pregnancy also conducted secondary analysis of factors associated with early pregnancy among young people in the HSRC 2003 Status of Youth Survey. The 2003 HSRC is nationally representative study of more than 3,500 young people aged between 18 and 35 years (Panday et al., 2009, p. 6). The Reproductive Health Research Unit Survey was a national survey of HIV prevalence and sexual behavior among 11,904 15-to-24-year-olds (men and women) in 2003. Pettifor et al. (2005) used 2001 South African national census data as a sampling frame. One participant per household was eligible to participate, provided that the individual agreed to provide an oral fluid specimen for HIV testing and agreed to participate in a comprehensive face-to-face interview regarding his/her demographics, sexual behaviors, symptoms of sexually transmitted infections, and exposure to HIV prevention programs. The authors used regression models (excluding sexually inexperienced participants) to determine risk factors for HIV infection (Pettifor et al., 2005, p. 1526). The 2002 First and 2008 Second Youth Risk Behavior Survey are nationally representative cross-sectional surveys of youth risk behavior among learners attending public school in grades 8 to 11. Using the DBE database of school information, Reddy et al. used stratified cluster sampling from the respective 1999 and 2007 databases to select 23 schools per province for participation. With cluster sampling methods, an average of two classes per school were selected to sample learners. The overall response rate of the 2008 YRBS was 71.6% with 1 Included in this review.

9

10,270 learner participants, which is a higher response rate than the First YRBS of 68.3% (Reddy et al., 2003, 2010). Data from respondents were weighted to reflect nationally representative demographic characteristics of learners by gender, grade, and provincial population (Reddy et al., 2003, 2010). The 2002 and 2008 YRBS survey administrators (local health professionals and volunteers trained in standardized survey administration) collected the height and weight of participants, and learners completed an anonymous self-administered 135-question survey (Reddy et al., 2010, p. 19). The Medical Research Council adopted the close-ended questionnaire from the Center for Disease Control Youth Risk Behavior Surveillance System survey, used biannually to assess adolescent behavior among secondary school students in the United States. 2008 YRBS questions covered youth risk behavior in four categories: (1) infectious disease; (2) chronic disease; (3) injury and trauma; and (4) mental health. Data regarding sexual health were obtained in the infectious disease category. Learners that participated in the 2003 and 2008 YRBS reported the following information regarding their sexual health: age of sexual debut; sexual activity in the past 3 months; use of alcohol and drugs before sex; contraceptive and condom use; pregnancy experience (including experience with abortion and parenting); experience with sexually transmitted infections; and knowledge about protection against HIV (Reddy et al., 2003, 2010). Secondary analysis conducted for the PRB Brief (2012) examined the effects of childbearing on education, specifically whether women who have a child before age 20 have higher dropout rates and less years of schooling than women who have children after age 20 (Karra & Lee, 2012). This analysis uses two longitudinal studies to examine the effects of childbearing: the Africa Centre Study and the Cape Area Panel Study (CAPS). Yet, the Africa Centre Study and the CAPS used different methodologies to obtain their results. The Africa Centre Study is a longitudinal annual demographic household survey that collects socio-demographic and health-related information of 11,000 households in KwaZuluNatal, with approximately 90,000 household members. The primary objective of the study is to monitor the impact of the HIV epidemic. The initial survey started in 2000 and it is ongoing. The sampling frame includes all resident and non-resident household members. The head of household reports data for all members. Questionnaires collect data on births, deaths, marriages and moves every six months. Household Socio-economic data are collected once a year (Africa Centre, 2013). The CAPS randomly selected 4,800 youth age 14 to 22 to participate in the first wave of the survey in 2002 in a metropolitan area of Cape Town. The original sample has been surveyed 4 subsequent times in-person (2003/2004, 2005, 2006, 2009), with an additional follow-up

10

telephone interview in 2010. CAPS questionnaires focus on schooling, employment, health, intergenerational supports, and family growth (University of Michigan & University of Cape Town, 2012). Definition of teenage pregnancy All of the studies described herein define adolescence/teenage years differently, depending on the sampling frame and study methodology. The SADHS collected household surveys with a diverse age population; however, the DHS defined “adolescents” as 15 to 19 year olds (Department of Health et al., 2007). The DBE also focused on teens age 15 to 19, even though some studies in the DBE literature review defined the beginning of adolescence as early as age 10 and extended “youth” to age 24 (Panday et al., 2009). The RHRU survey on HIV and sexual behavior collected data on youth age 15 to 24. Some of the data on teenage pregnancy and teen fertility prevalence in the RHRU are isolated for the study participants age 15 to 19 (Pettifor et al., 2005). The sample of youth studied in the YRBS included young people, ages

13 to 20.

Seventy-eight percent of study participants in the First YRBS were between 14 and 18 years old (Reddy et al., 2003). Nearly 70% of the population studied in the Second YRBS was between the ages of 15 to 18; 12% were over age 20, and 15% were age 14 or younger (Reddy et al., 2010). The PRB Brief describes a “teen mother” as a woman who has had her first child “before age 20” (Karra & Lee, 2012, p. 2, p. 3). The CAPS data highlighted in the PRB Brief studied a population age 14 to 22 in metropolitan Cape Town (University of Michigan & University of Cape Town, 2012). Whereas, the Africa Centre Study highlighted in the PRB Brief captured household data in rural KwaZulu-Natal, and the head of household provided the demographic data (including fertility rates) for the rest of the household members (Africa Centre, 2013). Teenage pregnancy data Sexual activity The DBE reported that 10% of 15-year-olds have had sex, compared to 61% of 19-yearolds (Panday et al., 2009). Findings from the SADHS 1998 and 2003 reported that the median age of sexual debut for young women is 18 years, with seven percent of teens initiating sex before age 15. Pettifor et. al (2005) found that, on average, adolescent males initiate sexual activity earlier (age 16) than females (age 17). In the 2002 First YRBS, 50% of male learners reported that they ever had sex, whereas 34% of female learners reported ever having sex (Reddy et al., 2003). In the 2008 YRBS, Reddy et al. (2010) reported that prevalence of male learners who had ever had sex (45.2%) was fifteen percent greater than the prevalence of female learners

11

who had ever had sex. Adolescents who initiate sex before age 15 are significantly more likely to be male. Pettifor et al. (2005) reported that males are more than twice as likely (17.5%) to initiate sex before age 14 than females (7.8%). Reddy et al. (2010) reported that early sexual activity was three times more likely among male learners (12.6%) than female learners (4.3%). With the exception of age-specific data for later teenage years (18 to 19 years), the literature showed that less than half (38 to 48 percent) of all adolescents have ever had sex (Department of Health et al., 2002, 2007; Pettifor et al., 2005; Reddy et al., 2010). The prevalence of sexual activity decreased in the five years between the first and second Youth Risk Behavior Survey (YRBS). Both studies surveyed a nationally representative sample of learners, grades 8 – 11, using the same methodology. Even though rates of sexual activity declined between the 2002 and 2008 YRBS, the proportion of self-reported sexually active adolescents who became pregnant or made someone pregnant increased from 16.4% to 19% of total learners by the 2008 YRBS [see yellow cell in Table 3]. There was no gender variance in the percent of learners who reported having been pregnant or having made someone pregnant in the First 2002 YRBS. The 2008 YRBS gender variance is given in Table 2.0. (Reddy et al., 2003, 2010). Table 2.0: Learner Sexual Behavior; Results of 2002 and 2008 YRBS Study 2002 YRBS 2008 YRBS Study 2002 YRBS 2008 YRBS

n (%) Male Female 46.0 54.0 48.5 51.5 Always used condom Total 29.0 30.7

Ever had sex (% learners) Total Male Female 41.0 50.1 34.1 37.5 45.2 30.2

Sexual initiation <14 Total Male Female 14.0 25.4 5.6 12.6 21.2 4.3

No contraceptive Ever pregnant (% method learners) Parent (% learners) Total Male Female Total Male Female Total Male Female 28.1 16.4 17.9 19.9 15.0 19.0 15.2 24.4 17.7 21.9 14.7

Teenage fertility The studies reported a range in fertility rate from 78 births per 1,000 adolescents in 1998 to 63 per 1,000 in 2008 (see Table 3). Only three studies reported teenage fertility in births per 1,000: the SADHS 1998 and 2003, and the DBE literature review by Panday et al. (2009). The 1998 SADHS and census data from the DBE is nationally representative, whereas the 2003 SADHS data may have underreported fertility rates due to data collection errors. While data from multiple sources shows that in teenage pregnancy is on the decline (Panday et al., 2009; WHO, 2011), the decrease in teenage fertility prevalence reported in the 2003 DHS is implausible, due to reporting errors in KwaZulu-Natal. The 1998 SADHS reported that 17% of

12

teens had begun childbearing by age 19 in KwaZulu-Natal; whereas the 2003 SADHS found that only 2% of teens had begun childbearing by age 19 in KwaZulu-Natal. (Department of Health et al., 2007). The only data that reports an increase in teenage fertility in recent years (between 2004 and 2008) is derived from the Education Management Information System (EMIS). The variant result may be from improved reporting or data capturing abilities by surveying young people at school (Panday et al., 2009). Table 3.0: Adolescent Fertility Rate by Year of Study Fertility rate Year

(per 1,000 adolescents)

Study

1996

78

Census data (Panday et al., 2009)

1998

76

SADHS (Department of Health et al., 2002)

2001

65

Census data (Panday et al., 2009)

2003

54

SADHS (Department of Health et al., 2007)

2004

51

EMIS data (Panday et al., 2009)

2005

56

EMIS data (Panday et al., 2009)

2006

56

EMIS data (Panday et al., 2009)

s2007

60

EMIS data (Panday et al., 2009)

2007

54

Census data (Panday et al., 2009)

2008

63

EMIS data (Panday et al., 2009)

Teenage pregnancy Rates of pregnancy and parenthood vary by study [see Table 4]. Karra and Lee (2012) and Reddy et al., (2010) report a higher rate of birth (22%) among adolescents than all other studies. The 1998 and 2003 SADHS reported the lowest birth rate among the studies in this paper.



13

Table 4.0: Percentage of Ever Pregnant/Parenting Female Adolescents by Study

The 1998 SADHS reported that 13% of young women were childbearing between the ages of 15 and 19, and just over 16% had ever been pregnant by age 19. Pettifor et al. (2005) also reported the pregnancy rate among adolescents age 15 to 19 was just under 16% in 2003. Karra and Lee (2012) found much higher rates of motherhood before age 20 in 2008, with a prevalence of 22% nationally, 31% in KwaZulu-Natal, and 19% in the Western Cape Town. According to the DBE, more sexually active young women become pregnant women as age increases. The ration of sexually active adolescents to pregnant adolescents is 13:1 at age 15, 7:1 at age 16, and 3:1 by age 17 and onward (Panday et al., 2009). Contraceptive use Pettifor et al., (2005) reported that condom use during last act of sexual intercourse was between 48% (females) and 57% (males) (Pettifor et al., 2005). Panday et al. (2009) also reported that female adolescents used condoms less frequently than males during their last sexual intercourse with a range of 47 to 62%) (Panday et al., 2009). The First YRBS (2002) reported that 29% of learners (male and female) reported consistent condom use, with 28% who did not use any method of contraception during “most” sexual intercourse (Reddy et al., 2003). By the Second YRBS (2008), more adolescents reported consistent condom use (31%), with no variation by gender (Reddy et al., 2010).



14

Discussion Study aims It is noteworthy that none of the studies described here were designed to study the prevalence of teenage pregnancy in the country. The 1998 and 2003 SADHS aimed to capture nationally representative data on basic health information; therefore, the results contain more detail regarding teen fertility than the other nationally representative studies. The RHRU study addresses HIV prevalence and sexual behaviors, and the YRBS addresses poor health outcomes associated with youth engagement in risky behaviors. The PRB Brief addresses the impact teen pregnancy has on education with data from two longitudinal studies that collect a broad array of demographic and health information. Study locations/populations Four out of the six studies highlighted in this paper contain nationally representative data on adolescent sexual behavior, including pregnancy and parenthood: the 1998 SADHS; 2002 and 2008 YRBS; and the RHRU survey of HIV prevalence and sexual behavior among South African Youth. The 2003 SADHS aimed to provide nationally representative data on sexual and reproductive health among South African women, but there were errors in data collection and reporting in the Kwa-Zulu Natal Province that contributed to unreliable results. While all studies focused on South African adolescents/youth, the DBE literature review included both national and international reports to depict the widespread teenage pregnancy in South Africa. The DBE focused on South African learners, and reviewed studies of youth who may or may not have been learners. The 2002 and 2008 YRBS focused exclusively on public school learners in grades 8 to 11. The 2002 and 2008 YRBS, and the RHRU survey included data on male and female participants, and sometimes data was not separated by gender if there were no significant statistical variances. All studies noted that teenage pregnancy and parenthood disproportionally affects young women with potential negative health, education, career, and economic outcomes. Teenage pregnancy results Because of considerable variations in methods and study populations, comparisons across the studies should be made with caution. If there were not data collection errors, one could more easily compare results from the 1998 and 2003 SADHS. Results from the 2002 and 2008 YRBS show trends, as they had similar study populations and methodologies. Results, though varied, all indicate that the prevalence of teenage pregnancy in South Africa is high. The negative health consequences of early pregnancy are 600 times higher in sub-Saharan Africa than in developed countries (Pettifor et al., 2005). South Africa’s rate is not as high as other countries

15

in the region, but it is a major concern, particularly because many pregnancies are unintended (Pettifor et al., 2005). More nationally and provincially representative studies on adolescent fertility and pregnancy are needed to determine trends and assess health policy and programs. Moving forward, it would be helpful to have a nationally representative longitudinal study that aimed to measure the rate of teenage pregnancy and fertility in South Africa. Limitations There is a limitation among all sexual behavior studies due to self-reported data. In many social contexts, there is a desirability bias for young people to underreport sexual activity. Results from the RHRU survey on HIV prevalence found that 3.8% of young women who reported that they never had sex were found to be HIV positive on their oral fluid specimen test (Pettifor et al., 2005). The context of data collection also contributes to social desirability bias. Some of the studies herein collected data from adolescents while they were at school, others collected data while the adolescents were at home, and sometimes a relative may have reported data on behalf of a family member or housemate, as is the case with the SADHS. This paper used a convenience sample of frequently cited literature on South African teenage pregnancy from another literature review, thus there are many limitations to the rigor, completeness, and repeatability of this methodological review. Recommendations Teen pregnancy is a complex social problem with many contributing factors that vary by region and require special, targeted interventions in South Africa. All young women who experience pregnancy may be subject to social disadvantages due to gender-based discrimination and violence as well as limited growth opportunities in education and occupation. Teenage pregnancy is an issue that deserves attention, and if progress is to be measured over time, consensus should be sought regarding how to measure teenage pregnancy nationally. The findings of this review highlight the need to develop a well-defined standardized means of measuring teen pregnancy to determine prevalence and measure trends. Designing effective targeted interventions to reduce teen pregnancy is dependent on obtaining reliable nationally representative data and studying changes in prevalence over time. For meaningful change to occur and improve health of teens and reduce teen pregnancy in South Africa, policy and programmatic initiatives depend on utilization of rigorous and comprehensive data on teen pregnancy. It is recommended that a national study be implemented and repeated on a triennial to quinquennial basis to monitor and track the prevalence of teen pregnancy, identify groups that need specially targeted interventions, and monitor health service distribution and utilization. Ibis

16

Reproductive Health is currently launching a working group to address this issue with the aim of developing recommendations for future research.



17

References Africa Centre. (2013). ACDIS. University of KwaZulu-Natal. Retrieved from http://www.africacentre.ac.za/Default.aspx?tabid=333 Department of Health, Medical Research Council, & Measure DHS+. (2002). South Africa Demographic and Health Survey 1998 (pp. i–40). Pretoria: Department of Health. Retrieved from http://www.mrc.ac.za//bod/demographicsurvey.htm Department of Health, Medical Research Council, & ORCMacro. (2007). South Africa Demographic and Health Survey 2003 (pp. i–411). Pretoria: Department of Health. Retrieved from http://www.mrc.ac.za//bod/sadhs.htm EMIS. (2012). Education Statistics in South Africa 2010. Department of Basic Education. Retrieved from http://www.education.gov.za/EMIS/StatisticalPublications/tabid/462/Default.aspx Holt, K., Lince, N., Hargey, A., Struthers, H., Nkala, B., Mclntyre, J., ... Blanchard, K. (2012). Assessment of service availability and health care workers’ opinions about young women’s sexual and reproductive health in Soweto, South Africa. African journal of reproductive health, 16(2), 283–293. Kaiser Family Foundation. (2012). Birth Rate (Births Per 1,000 Population) GlobalHealthFacts.org. Retrieved February 11, 2013, from http://www.globalhealthfacts.org/data/topic/map.aspx?ind=86#notes Karra, M., & Lee, M. (2012). Human capital consequences of teenage childbearing in South Africa (pp. 1– 5). Washington DC: Population Reference Bureau. Macleod, C. I., & Tracey, T. (2010). A decade later: follow-up review of South African research on the consequences of and contributory factors in teen-aged pregnancy. South African Journal of Psychology, 40(1), 18–31. Mngadi, P.T., Thembi, I. T., Ransjo-Arvidson, A. B., & Ahlberg, B. M. (2002). Quality of maternity care for adolescent mothers in Mbabane, Swaziland. International Nursing Review, 49, 38–46.

18

Nour, N. M. (2006). Health consequences of child marriage in Africa. Emerging infectious diseases, 12(11), 1644–1649. doi:10.3201/eid1211.060510 Panday, S., Makiwane, M., Ranchod, C., & Letsoala, T. (2009). Teenage pregnancy in South Africa: with a specific focus on school-going learners (Child, Youth, Family and Social Development, Human Sciences Research Council) (pp. 1–96). Pretoria: Department of Education. Pettifor, A. E., Rees, H. V., Kleinschmidt, I., Steffenson, A. E., MacPhail, C., HlongwaMadikizela, L., … Padian, N. S. (2005). Young people’s sexual health in South Africa: HIV prevalence and sexual behaviors from a nationally representative household survey. AIDS (London, England), 19(14), 1525–1534. PRB. (2012). Birth Rate (annual number of births per 1,000 total population) by The Population Reference Bureau. Retrieved February 11, 2013, from http://www.prb.org/DataFinder/Topic/Rankings.aspx?ind=3 Reddy, S., Panday, S., Swart, D., Jinabhai, C., Amosun, S., & James, S. (2010). Umthente Uhlaba Usamila – The South African Youth Risk Behaviour Survey 2008 (pp. 1–176). Cape Town: South African Medical Research Council. Reddy, S., Panday, S., Swart, D., Jinabhai, C., Amosun, S., James, S., … Van den Borne, H. (2003). Umthenthe Uhlaba Usamila – The South African Youth Risk Behaviour Survey 2002. Cape Town: South African Medical Research Council. Retrieved from http://www.mrc.ac.za/healthpromotion/reports.htm Shaw, D. (2009). Access to sexual and reproductive health for young people: bridging the disconnect between rights and reality. International journal of gynaecology and obstetrics: the official organ of the International Federation of Gynaecology and Obstetrics, 106(2), 132–136. doi:10.1016/j.ijgo.2009.03.025 UNFPA. (2010). Country profiles for population and reproductive health; Policy developments in indicators 2009/2010. United Nations Population Fund and Population Reference Bureau. Retrieved from www.unfpa.org/webdav/site/global/.../countryprofiles_2010_en.pdf



19

University of Michigan, & University of Cape Town. (2012). South Africa - Cape Area Panel Study 2002-2006, Waves 1-5 - Overview. University of Cape Town. Retrieved from http://www.datafirst.uct.ac.za/catalogue3/index.php/catalog/266/overview Varga, C. A. (2003). How gender roles influence sexual and reproductive health among South African adolescents. Studies in Family Planning, 34(3), 160+. WHO. (2011). Early marriages, adolescent and young pregnancies; Report by the Secretariat (No. EB130/12). World Health Organization. Retrieved from http://apps.who.int/iris/bitstream/10665/23744/1/B130_12-en.pdf



20