Prevalence and factors associated with adolescent pregnancies in Zambia: a systematic review from 2000–2022

This systematic review was conducted following the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [17]. We combined text words and medical subject headings such as pregnancy OR birth OR delivery with adolescent OR teen to search Medline, EMBASE, CINAHL, and African Journals Online from the database inception to 10 October 2021. The reference lists of eligible full texts were hand searched to retrieve articles missed during the search.

Primary studies (both qualitative and quantitative) that reported, or contained relevant data, on the prevalence and factors associated with adolescent pregnancy in Zambia were considered for inclusion in this review. The study excluded reviews, editorials, case reports, and case series with fewer than 30 participants. Detailed inclusion and exclusion criteria are presented in Table 1. We exported citations retrieved from database searches to EndNote X9 for removal of duplicates, and the (re)-duplicated records were uploaded to Rayyan QCRI for screening based on title and abstract.

Adolescent pregnancy was defined as pregnancy or childbirth occurring in women less than 20 years old. This included self-report or evidence (record) of childbirth at a health facility by women aged less than 20 years. The prevalence of adolescent pregnancy was calculated as the number of women who become pregnant or had a child while they were less than 20 years old.

The second outcome of this study focused on factors influencing adolescent pregnancy. Factors associated with adolescent pregnancy, in the quantitative study conducted, were determined from the p-value less than 0.05 for the statistic from the final inferential analysis carried out. Themes related to factors with a bearing to adolescent pregnancy identified from the findings of the included qualitative studies were collated.

Quality assessment of the six (6) papers was carried out using Sirriyeh et al.’s tool for quality assessment of studies with diverse designs (QATSDD) [18]. The use of this tool also allows for comparisons of the qualities of quantitative, qualitative, or mixed-methods papers within the same domain field of study. Two authors (FIS and LEB) independently assessed the included studies and where disagreements were encountered, a third author resolved this through a third assessment informed by discussions held. The tool awards quality score on a four-point scale from 0 to 3 with accompanying guidance notes for using the scoring criteria (qualitative study (14), quantitative study (14), and mixed methods (16)) [18]. The criteria included a theoretical framework, a statement of aims and objectives, a description of the research setting, sample size in relation to analysis, representativeness, sample size, data collection procedures, rationale for choice of data collection tool, recruitment data, statistical assessment of reliability and validity of tool (quantitative only), fitness between stated research question and method of data collection (quantitative only), stated research question and format and content of data collection tool (qualitative only), stated research question and method of analysis, justification for selected analytical method, reliability of analytical process assessment (qualitative only), and user involvement in design, and discussion of strengths and limitations [18].

Each research paper was awarded a score per criterion and index scores generated. The total quality scores of the included studies were computed in percentage using the expected total scores (42 for quantitative and qualitative studies and 48 for mixed methods studies). Percentage scores greater than 50% were designated as high while those less than or equal to 50% were graded low-quality.

Two authors independently extracted data on the surname of the first author, year of publication, and determinants of adolescent pregnancy. The data extraction tool contained information on the author and year of publication, study aim, study design, study area and province, sample size, study population, and factors associated with adolescent pregnancy. A thematic synthesis was adopted in this review. The analysis of the factors influencing adolescent pregnancy was based on themes from the SEM. As such the themes for determinants of influence on adolescent pregnancy were categorized into individual (for example, exposure to media, and knowledge on use of contraceptives), interpersonal/socio-economic (for example, educational status, household wealth), and policy /legal framework (for example, knowledge of and access to sexual reproductive health services) factors. Bronfenbrenner’s model is used because it allows for an assessment of all possible factors that contribute to a phenomenon, addressing the issue as resulting from influences of multifaceted and multilevel factors [19].

A total of 3,452 papers were initially accessed from various electronic database searches and digital library catalogues, however 2,665 research had to be eliminated since they were duplicates. The remaining articles’ titles and abstracts were then evaluated, and 504 studies were reviewed and removed because they were not germane to the study. On the basis of the established inclusion and exclusion criteria, the eligibility of the remaining five (5) full-text publications was assessed. From snowballing through references of eligible articles, one (1) study was further assessed and included. Thus, in the final analysis, only the six papers that met the eligibility requirements were included (Fig. 1).

Of the 6 studies retained in the final review, two (2) were qualitative, two (2) were mixed methods and two (2) were cross-sectional. The studies were carried out in the following Provinces of Zambia: Southern (2), Central (2), Eastern (1), North-Western, as well as in Lusaka (2) and Copperbelt (1) while two used national-level data from the Zambia Demographic and Health Surveys (ZDHS). All the six studies in the provinces showed high quality ranging from lowest 54.7% [20] and 76.2% [21] (Table 2). Guided by Wado et al.’s [21] study, all the five did not have explicit theoretical frameworks. Additionally, all the six studies did not provide any justification for the choice of data collection tools and all the quantitative and mixed methods studies did not assess statistical reliability and validity of the tools, which could be due to the use of secondary data for some of the studies [21, 22]. Such studies did not select the data collection tools and conducted their own data collection. With the exception of Svanemyr’s [23] study, all the other studies did not have user involvement in the study’s design.

The prevalence of adolescent pregnancy ranged from 29% in the study using the 2014 ZDHS [21] to 48% in a study in Petauke, Chadiza, and Katete districts in Eastern province [24]. Additionally, prevalence of adolescent motherhood ranged from 20.81% in the Zambezi district, North-Western province [26] to 23.1% in the study using 2014 ZDHS data [22].

The prevalence of adolescent pregnancy in Zambia continues on an upward trajectory according to Zambia’s recent Demographic and Health Surveys. For instance, the total adolescent pregnancy rate declined from 33.8% in 1992 to 30.7% in 1996, then increased to 31.6% in 2002. Table 3 shows an increase from 28.5% in 2014 to 29.2% in 2018 [10]. The Southern (42.5%), Western (41.2%), Eastern (39.5%), North Western (35.7%), and Central (30.6%) provinces recorded higher prevalence than the national average [10].

Prevalence was also observed to be positively correlated with increasing age, occasioning a distinct possibility of increasing vulnerability to risky sexual behaviours on the part of adolescents [1]. We observed a high prevalence among those in rural areas compared to those in urban areas which is indicative of rural–urban disparities from 1992 to 2018. Adolescents in urban areas may have access to improve economic opportunities and social services including SRH services and information compared to those in rural settings.

Also, prevalence of adolescent pregnancy was observed to have decreased with increasing levels of education with the trend showing increasing pregnancy for adolescents with primary and secondary education from 1992 to 2018 and decreasing for those with no formal education (Table 3). The level of education is a key determinant of adolescent pregnancy, with prevalence decreasing with increasing level of education. It is, however, a matter of concern that between 1992 and 2018, the teenage pregnancy rates among students in secondary school remained fairly constant (21.2% in 1992 and 22.2% in 2018). It might be of interest to understand reasons for this stagnation, as well as the comprehensive sexuality education packages in schools. This means that increasing access to education at all levels will help limit adolescent pregnancy [23].

Of the six papers selected, varying associated factors were identified. These factors are consistent with the individual, socio-economic, environmental, and policy levels of the SEM (Table 4).

Bronfenbrenner’s [19] socio-ecological model/framework posits a multiplicity of factors associated with adolescent pregnancy operationalize at four levels: the individual (microsystem); interpersonal (mesosystem); environmental (ecosystem) and policy level (macrosystem).This model offers explanations regarding the influence of social environments on human development and provision of health services to guide the research effort focusing on key actors/variables which include: age, sex, gender issues, employment status, residence, religion, early marriage, poverty, education level, peer pressure, coercive sexual relations, comprehensive sexuality education, access to sexual and reproductive health services including use contraceptives.

At the Microsystem (individual) level of influence, the adolescent is the main centre of focus in terms of factors affecting responses advocating refraining from indulging in risky sexual behaviours, including having multiple sexual partners. Menon et al. and Wood et al. [24, 27] attest, through various studies carried out, that risky sexual behaviour can be construed as a function of age at sexual debut [21, 24], exposure to media [21], knowledge of sexuality and reproduction [23], knowledge of contraception [22, 24], contraceptive use [22, 24], risky sexual behaviours (transactional sex [25], multiple sexual partners [25], child marriage [24], low levels of education, attitude/ perceptions by the adolescent as well as poverty in the household to which the individual belongs.

At the Mesosystem (interpersonal/socio-economic) level of influence, the individual’s disposition towards sexual behaviour is mostly influenced by close family members without ruling out other factors/variables such as educational attainment [20‐23], marital status [22], household wealth/poverty [20, 22‐24, 26], peer pressure [23], and relationship to household head [21]. Low household wealth was a major factor identified in many studies [21, 22, 24]. The studies reported that low household economic status increases the risk of adolescent pregnancy with adolescents from the wealthiest households being protected from adolescent pregnancy. Menon et al. note that nationally, household poverty increases the risk of pregnancy among adolescents in Zambia using the 2014 ZDHS data [24].

It was also reported that major factors to adolescent pregnancy at the socio-economic factors were peer pressure [23] and relationship to household head (spouse) [21]. The researchers noted that adolescents who were spouses to the household head were more likely to get pregnant (OR = 14.79, 95%CI = 11.55–8.63) [21]. This however begs the question pertaining to the period the adolescent became pregnant as it is not abundantly clear whether conception occurred before or after marriage to the household head (presumably male-headed household).

Ecosystem (Community/Societal) level factors for adolescent pregnancy were also reported. Key at this level of influence, are neighbours, family and friends who potentially can influence an adolescent to be favourably disposed towards a certain trajectory. It is also at this level, that differences in prevalence rates become glaring as reported by Menon et al. [24]. These were socio-cultural, gender and sexual norms [21, 22] which encouraged early or child marriage, early sexual initiation, unprotected sex, and barriers to accessing sexual and reproductive health information and services. For instance, Svanemyr observes that socio-cultural norms about the use of contraceptive increases the risk of adolescent pregnancy in Zambia [23].

The Macrosystem (Policy/Legal) level of influence can be considered to be the most influential factor as enshrined in legal statutes which have force and effect of law, as attested for instance, by regulatory and policy framework on early marriages. Policy level factors comprised limited accessibility to SRH information and SRH services, including laws which prohibit adolescents from accessing SRH services including contraceptives [24]. The authors observed that due to limited access to sexual and reproductive health information and services many adolescents in Zambia are at risk of unplanned pregnancies.

The review revealed a significantly high prevalence of adolescent pregnancy ranging from 29% nationally [21] to 48% in the Eastern and Southern provinces [24]. Some of the studies [21, 22] also assessed adolescent motherhood, which presents as a limitation to capturing the magnitude of the problem, since this approach does not cover adolescents who may have lost pregnancy due to induced abortion or miscarriage. Nevertheless, this prevalence is in tandem with the high prevalence of the national adolescent pregnancy reported in the 2018 ZDHS [10]. It is important to understand the context-specific drivers of adolescent pregnancy in the country. One way could be through carrying out robust country-wide studies to allow for comparisons. This would address information gaps relating to possible ineffective or weak implementation of existing policies and interventions instituted to address the problem [28‐30].

This review showed that low level exposure to media is a factor significantly impacting adolescent pregnancy in Zambia. Exposure to various media platforms provides SRH information to young people [31, 32], and in some cases, a useful source of information on where to obtain SRH services including contraceptives and abortion services, hence contributing to the removal of some existing barriers to accessing SRH care. Further, exposure to the media has been found to trigger discussions about ASRH issues in addition to encouraging adolescents to utilize SRH youth reproductive and family planning services [33]. Exposure to mass media, can greatly influence adolescents’ attitudes, social expectations and avoidance of risky sexual behaviours thereby reducing risk of pregnancy [31].

Insufficient knowledge about sexual and reproductive health has been noted to increase the risk of adolescent pregnancy in Zambia. This is in line with the findings of many studies carried out in low- and middle-income countries (LMICs) [34‐38]. Lack of knowledge about SRH issues indicates prevalence of poor safe sexual practices including contraceptive use and STIs prevention. As such, age-appropriate sexual health education programmes for adolescents are key to developing safe SRH as well as preventing adolescent pregnancy [39].

Early or child marriage was noted to be another a factor impacting Zambia’s adolescent pregnancy as attested by similar findings in previous studies across Africa pointing out the influential role of socio-cultural norms and practices in most developing countries [3, 4, 39]. Child marriage is still prevalent across Africa despite most countries having laws prohibiting and criminalizing marriage before the age of 18 years which might be due to sporadic or total lack of enforcement of these laws with no sanctions when they are routinely violated [40].

It is common cause that risky sexual behaviours among most adolescents have been reported as factors that increase the risk of adolescent pregnancy [27, 41]. Frequent sexual activity at an early age, multiple sexual partners, transactional and unprotected sex are common among adolescents and young people. Such risky sexual behaviours can be attributed, amongst other factors, to household poverty levels as noted in many studies [27, 41, 42]. In fact, a number of studies have reported a tendency among some adolescent girls to use sexual relationships with older men to cater for some of their financial needs [41, 42]. Early sexual initiation and lack of contraceptives as factors impacting adolescent pregnancy, as reported elsewhere in this review, is corroborated by findings from previous studies [3, 43]. In contrast, Worku et al.’s study indicated that contraceptive use among adolescents tends to increase the risk of adolescent pregnancy [44]. This could be explained as result of failed contraceptive failure due to poor knowledge on its correct use.

This review also accentuates the importance of two main interpersonal level factors, namely, peer pressure and relationship to household head. At the interpersonal level, supportive relationships that reinforce positive sexual and reproductive health behaviours among adolescents such as family influence, have an impact on the sexual and reproductive health vulnerabilities and experiences of adolescents [45]. Peer pressure as a major factor contributing to adolescent pregnancy has been reported in many other studies in Africa [46‐48]. It has been established that peer groups play a significant role as agents of socialization which often influences lifestyle choices including sexual behaviours of teenagers. Pressure from this group may be so overwhelming to an extent that teenagers tend to conform to the sexual behaviour norms considered acceptable to their peer group [49].

Regarding the finding that relationship with household head is associated with adolescent pregnancy, this is not inconsistent with the findings by Worku et al. where relation to household head was significantly associated with teenage pregnancy in East Africa [44]. This could be attributed to the fact that some of the adolescents are married to the household head which increases their likelihood of becoming pregnant as wedded adolescents, a fact also identified by Wado et al. [21]. It is worth noting that teenage pregnancy mostly precedes the marriage. This is specially the case in most developing countries where sometimes entering into marriage is occasioned by unplanned pregnancy [50, 51].

Additionally, low level educational attainments were associated with increased risk of adolescent pregnancy in Zambia. This finding resonates well with those of other similar studies which report that low educational expectations and school dropouts are common factors for adolescent pregnancy [8, 43, 52‐54]. Studies have found a link between the number of years spent in school and modern contraceptive use with more years associated with higher chances of using modern contraceptives [55‐59]. In addition, educational attainment and adolescent pregnancy are interconnected such that while low educational attainment increases risk of adolescent pregnancy, unplanned pregnancy often truncates education for most adolescent girls. This could be attributed to the fact that education empowers adolescent girls regarding their sexual relationships and safe sex practices [60].

Low socio-economic status in the form of poor household wealth, poverty, and lack of resources tend to have a direct and high impact on the risk of adolescent pregnancy. Several studies have indicated that low family incomes increase the risk of adolescent pregnancy [38, 42]. These studies further observed that adolescent girls engaged in sexual relations with men in exchange for gifts and money or in return for financial support for their needs including but not limited to basic necessities such as sanitary pads and school fees [61, 62]. It can be noted that household poverty potentially increases the vulnerability of adolescent girls due to the need to meet some basic necessities. This may be construed as explaining the use of transactional sex as an economic survival strategy by young girls from households with high poverty levels risking pregnancy at a younger age [63].

The finding that existing socio-cultural, gender and sexual norms in Zambia are linked to increased risk of adolescent pregnancy is supported by many other studies [35, 37, 64, 65]. For instance, on cultural beliefs, encouraging larger families and early childbearing, Ayele et al. and Kaphagawani and Kalipeni reported that the prevailing cultural norms of marrying off young girls once they reach puberty leads to adolescent pregnancy [3, 4]. Studies have also found that being a spouse was a significant predictor of adolescent pregnancy [4, 44, 54]. These norms may contribute to adolescent girls seeing unplanned pregnancies as normal and socially accepted.

Similarly, norms governing contraceptive use were found to be a factor in adolescent pregnancy. Community-wide negative perceptions and beliefs regarding modern contraceptives was noted to be a common predictor of adolescent pregnancy in previous studies [66‐69]. These norms lead to the unacceptability and non-use of modern contraceptives among adolescents. In some cases, parents and guardians do not support the provision of contraceptive services to adolescents hence recommendations have been made to engage and persuade key community gatekeepers such as religious leaders and chiefs to generate wider community support [66, 70]. In this regard, Svanemyr et al. contend that as part of efforts to reduce environmental level factors, positive social norms and community support need to be enabled for adolescents to practice safer sexual behaviours and have increased access to SRH information and services [45].

In addition to the above-mentioned factors, limited access to SRH information and services was identified as a policy/legal level factor for adolescent pregnancy in Zambia. Other studies in LMICs have also reported that lack of SRH service increases the risk of adolescent pregnancy [71‐73]. Such a situation hinders adolescents from receiving much-needed SRH information, counselling and contraceptive services [71], against a background where the unmet need for contraceptives alone is considerably high in LMICs [74]. At the policy level, embedded provisions in some laws and policies related to the health, socio-economic, educational sectors and broad societal norms may hinder access to SRH and adolescents’ enjoyment of their human rights and freedoms [45]. In Zambia, adolescents below 16 years of age require parental/guardian consent to access SRH services, especially contraceptives.

The factors identified in this review suggest that the factors associated with adolescent pregnancy are varied but multifaceted and interconnected. There was more evidence about determinants of adolescent pregnancy in the individual and socio-economic levels compared to the environmental and policy levels. This indicates paucity of empirical evidence at these levels to support policy and intervention development and likely cause of the failure of some efforts to reduce the problem of teenage pregnancy. It thus shows that in designing interventions, little evidence from environmental and policy levels have been considered and incorporated. We did not observe any differences in the factors/levels identified in the literature based on the indicators of adolescent pregnancy measured by researchers (record of live birth or self-report). However, studies using primary data captured factors at more levels than those using secondary data.

This systematic review presents findings of critical importance to the prevention of adolescent pregnancy in Zambia. However, the pulled studies do not cover all the regions, hence limiting the representativeness of the review findings to generalize on the whole country. In some of the included studies, adolescent motherhood was instead assessed as proxy to adolescent pregnancy. This presents as a limitation in capturing the magnitude of the problem since this approach does not cover adolescents who may have lost pregnancy due to induced abortion or miscarriage. In addition to this, the low number of studies included in the final review and the inclusion of various designs of studies that led to the heterogeneity which hindered the ability to perform a meta-analysis. In addition, there was no control over the data collection, cleaning and analysis of the studies considered in this study. There is however, merit on the integrity of the data and findings of the selected studies which are very germane and relevant to the prevention of adolescent pregnancy in Zambia and other SSA countries.

The prevalence of adolescent pregnancy in Zambia remains high and could be a major hindrance for the country to attain SDGs on health and wellbeing (SDG3) gender equality and women’s empowerment (SDG5) and decent work (SDG8) [7]. Numerous health interventions such as health education, skill-building, and increasing accessibility to contraceptives [75‐78] including policies and legal frameworks have been implemented over the years to improve the health of young people and decrease adolescent pregnancy in the country [79]. Adolescent pregnancy however remains high. These interventions and policies are somewhat fragmented and not designed with multi-level and multi-sectoral connections. Thus, they may tend to overlook other aspects of the multifaceted factors being delivered or may serve to create other avenues which may mitigate against these efforts. Robust and nationally representative studies are needed to fill this lacuna.