Introduction

Adolescence spans ages 10–19 years, although there are developmental variations for early, middle and late adolescence, defined as ages 10–14, 15–17 and 18–19 years, respectively (Dixon-Mueller, Reference Dixon-Mueller2008). Adolescence is an intricate transitional period from childhood to adulthood characterized by physiological and psychological maturity, various external influences, and exposures and experimentation that can enhance sexual risk (WHO, 2001). According to UNICEF/WHO/UNFPA definitions, the stage overlaps with other youthful stages: childhood and youth (UNDESA, no date).

Adolescent sexuality is discussed as part of the broader sexual and reproductive health and rights entrenched by the 1994 International Conference on Population and Development’s Programme of Action (ICPD-PoA), which provided the historic sexual and reproductive health and rights framework, spelled out the population–health–development synergy and addressed the age-driven health needs of adolescents (UN, 1994). Healthy adolescent sexual development is critical for health and socioeconomic well-being (ibid). It is well established that adolescent sexual and reproductive health, and timing of sexual activity in particular, are intricately linked to several outcomes, namely: pregnancy and childbearing, maternal health and HIV and other sexually transmitted infections (Zaba et al., Reference Zaba, Boerma, Pisani and Baptiste2002; Ghebremichael et al., Reference Ghebremichael, Larsen and Paintsil2009; Louie et al., Reference Louie, de Sanjose, Diaz, Castellsague, Herrero and Meijer2009; Stockl et al., Reference Stockl, Kalra, Jacobi and Watts2013). Early sexual debut is also linked to cervical cancers (Louie et al., Reference Louie, de Sanjose, Diaz, Castellsague, Herrero and Meijer2009). In the context of structural, socio-cultural and financial barriers to adolescents accessing sexual and reproductive health services (Chandra-Mouli et al., Reference Chandra-Mouli, Svanemyr, Amin, Fogstad, Say, Girard and Temmerman2015), sexual deferment is a tenable option.

Adolescents in sub-Saharan Africa are especially affected by adverse sexual and reproductive health outcomes associated with early sexual activity, and this exists against the backdrop of the HIV pandemic, largely transmitted through heterosexual contact (UNAIDS, 2016), child marriages (UNFPA, 2012), unmet need for contraceptives (Chandra-Mouli et al., Reference Chandra-Mouli, McCarraher, Phillips, Williamson and Hainsworth2014) and restrictive abortion laws (Sedgh et al., Reference Sedgh, Singh, Shah, Åhman, Henshaw and Bankole2012). Sub-Saharan Africa also has a substantial and rapidly growing adolescent population (UNICEF, 2012), which currently constitutes almost a quarter of the total population of the region (UNICEF, 2015). Female adolescents in particular are disproportionately affected, largely because being female and young poses high sexual and reproductive health risks (Fathalla, Reference Fathalla1998). Female adolescents experience early childbearing, which creates a substantial morbidity and mortality burden among adolescents (UNFPA, 2012). HIV infections among adolescent girls continue to be markedly higher than in adolescent boys in sub-Saharan Africa (UNAIDS, 2016).

There has been substantial research on the socio-cultural risk factors associated with adverse adolescent sexual and reproductive health, one of which is parental absence (Ngom et al., Reference Ngom, Magadi and Owuor2003; McGrath et al., Reference McGrath, Nyirenda, Hosegood and Newell2009; Mmari, Reference Mmari2011; Pilgrim et al., Reference Pilgrim, Ahmed, Gray, Sekasanvu, Lutalo and Nalugoda2014). Parental presence is known to be protective against early sexual debut, sexual activity and pregnancy (Ngom et al. Reference Ngom, Magadi and Owuor2003; McGrath et al., Reference McGrath, Nyirenda, Hosegood and Newell2009; Pilgrim et al., Reference Pilgrim, Ahmed, Gray, Sekasanvu, Lutalo and Nalugoda2014).

There are myriad pathways through which parental absence may influence children’s sexual and reproductive health. Parents provide primary sexual socialization, i.e. impart to their children social, cultural and religious values regarding sexuality (Shtarkshall et al., Reference Shtarkshall, Santelli and Hirsch2007). Parents are well placed to play this role because of their huge stake in the welfare of their children. Involvement of parents in their adolescent children’s lives is also argued to influence adolescents’ self-esteem, which is linked to health-related behaviours (WHO, 2001). Parental absence during the critical adolescence stage may therefore be detrimental to adolescent sexual and reproductive health.

Research on the association between orphanhood status and sexual and reproductive health outcomes seems to have taken centre-stage with the advent of HIV (Robertson et al., Reference Robertson, Gregson and Garnett2010; Mojola Reference Mojola2011; Chae, Reference Chae2013; Mkandawire et al., Reference Mkandawire, Tenkorang and Luginaah2013). In sub-Saharan Africa, an estimated 52 million children aged 0–17 years are orphans, defined as children aged 0–17 years whose parent or parents are dead, out of the global 140 million (UNAIDS et al., 2004; UNICEF, 2015). Research has consistently shown that the majority of orphans are adolescents, regardless of geographic setting (Fisher, Reference Fisher1950; UNAIDS et al., 2004; Hosegood et al., Reference Hosegood, Floyd, Marston, Hill and McGrath2007; Shoko & Ibisomi, Reference Shoko and Ibisomi2016). This is explained by the cumulative increases in orphans with age, and the higher mortality amongst parents of older children due to the older age of the parents as compared with younger children (Fisher, Reference Fisher1950; Palloni & Lee, Reference Palloni and Lee1992; UNAIDS et al., 2004; Hosegood et al., Reference Hosegood, Floyd, Marston, Hill and McGrath2007). Sub-Saharan Africa has the most severe HIV epidemic in the world, estimated to be 70% of global HIV prevalence and of AIDS-related deaths (UNAIDS, 2016). The effect of HIV/AIDS is severe for adults in their prime years resulting in high adult mortality, producing a parallel rise in the number of orphans (Palloni & Lee, Reference Palloni and Lee1992; UNAIDS et al., 2004; Robertson et al., Reference Robertson, Gregson and Garnett2010). The anti-retroviral therapy (ART) roll-out in the region is the most important HIV intervention to date, and has resulted in a substantial reduction in HIV-related mortality (Reniers et al., Reference Reniers, Blom, Calvert, Martin-Onraet, Herbst and Eaton2016) and orphanhood prevalence (Makumbi et al., Reference Makumbi, Nakigozi, Sekasanvu, Lukabwe, Kagaayi and Lutalo2012; Rathod et al., Reference Rathod, Chi, Kusanthan, Chilopa, Levy and Sikazwe2014) and has increased survival into adolescence for children perinatally affected by HIV (Mofenson & Cotton, Reference Mofenson and Cotton2013).

Upon parental death, adolescents’ living arrangements almost always change (Mojola, Reference Mojola2011). The transition to other households in the event of parental death disrupts interpersonal relationships, requiring children to build new relationships with their new care-givers, with varying degrees of success (Foster, Reference Foster2004; Juma et al., Reference Juma, Askew and Ferguson2007; Mojola, Reference Mojola2011; Puffer et al., Reference Puffer, Drabkin, Stashko, Broverman, Ogwang-Odhiambo and Sikkema2012). This transition may affect children’s social and emotional development, and has a bearing on their sexual development and sexual and reproductive health (Juma et al., Reference Juma, Askew and Ferguson2007). The characteristic and culturally sanctioned living arrangements for orphans is to seek care from kin networks, although orphans are increasingly living outside of family care (Foster, Reference Foster2004). Research in sub-Saharan Africa on the effects of orphanhood status on sexual and reproductive health outcomes such as early sexual debut, pregnancy, marriage and HIV infection and other sexually transmitted infections, although substantial, has not produced completely consistent findings. For example, a study by Palermo and Peterman (Reference Palermo and Peterman2009) found no evidence that orphanhood was associated with marriage or pregnancy, but with sexual debut. A study by Robertson et al. (Reference Robertson, Gregson and Garnett2010) suggested that the association between orphanhood and sexual debut existed for countries with high orphanhood levels and not those with low orphanhood prevalence. The association also varies by category of orphans, i.e. maternal, paternal or double orphans, and sex of the respective adolescent (McGrath et al., Reference McGrath, Nyirenda, Hosegood and Newell2009), in some cases existing for female adolescents only (Robertson et al., Reference Robertson, Gregson and Garnett2010; Chae, Reference Chae2013). A study by Magadi and Uchudi (Reference Magadi and Uchudi2015) did not find orphans to have a higher risk of sexual debut than their non-orphan counterparts.

The variation is partly due to the diverse control variables used in multivariate models, age cut-offs, geography and research design given the contextual variations in health. The age criterion used in data collection often introduces restrictions on the ages that are to be considered in research using secondary data. For studies using the publicly available Demographic and Health Survey (DHS) data, the age range for analysis of sexual debut is 15–17 years because these are the only ages with available data for both parental survival status and sexual and reproductive health outcomes (Palermo & Peterman, 2009; Robertson et al., Reference Robertson, Gregson and Garnett2010). For longitudinal studies and studies using primary or other sources of data, ages range from early adolescence to mid-20s (McGrath et al., Reference McGrath, Nyirenda, Hosegood and Newell2009; Chae, Reference Chae2013; Mkandawire et al., Reference Mkandawire, Tenkorang and Luginaah2013). Although the relationship is less conclusive, orphanhood seems to be an important factor for female adolescents’ sexual and reproductive health vulnerability (Operario et al., Reference Operario, Underhill, Chuong and Cluver2011). This relationship is likely to be influenced by demographic and socioeconomic factors, age, rural–urban residence or care arrangements for the orphan (Operario et al., Reference Operario, Underhill, Chuong and Cluver2011; Magadi & Uchudi, Reference Magadi and Uchudi2015; Embleton et al., Reference Embleton, Nyandat, Ayuku, Sang, Kamanda and Ayaya2017). Research also suggests the aforementioned factors may be important not only as confounding factors, but also in their own right (Mkandawire et al., Reference Mkandawire, Tenkorang and Luginaah2013). Nevertheless, the highlighted inconsistency between orphanhood and sexual and reproductive health outcomes may also be because the longstanding patterns of fosterage in the region are rarely considered.

Parental absence is not exclusively due to parental death, and may be prompted by other circumstances (Meintjes et al., Reference Meintjes, Hall, Marera and Boulle2010; Gaydosh, Reference Gaydosh2015), so the orphaned children may not have been living with the deceased parent to begin with. Children may be ‘virtual orphans’ when parents are living elsewhere (Case et al., Reference Case, Paxson and Ableidinger2004). While parental non-co-residence may not always imply child–parent non-contact, frequency of contact is almost certainly likely to vary, and ultimately to affect the extent of parental influence. There is emerging research considering the effect of adolescents’ living arrangements on sexual and reproductive health, taking into account orphanhood status, but this research is sparse and geographically restricted.

A longitudinal study in a district of Uganda showed that parental absence, particularly of fathers, whether due to death or otherwise, resulted in early sexual debut among female adolescents (Pilgrim et al., Reference Pilgrim, Ahmed, Gray, Sekasanvu, Lutalo and Nalugoda2014). A cross-sectional study of male and female adolescents in two districts of Sierra Leone found that living with a surviving parent, a mother, father or both parents, delayed sexual debut (Stark et al., Reference Stark, Tan, Muldoon, King and Lamin2015). A study using a two-phased global Well-Being of Adolescents in Vulnerable Environments (WAVE) study data, suggested that in Johannesburg, South Africa, there was no association between parental absence and sexual debut for male adolescents (Mmari et al., Reference Mmari, Kalamar, Brahmbhatt and Venables2016). Contrary to other studies, female adolescents who were raised by other persons besides their parents were less likely to have sex compared with their counterparts who were raised in households in which both their biological parents were present. In addition, the risk of sexual debut was higher for maternal and double orphans, but lower for paternal orphans. For males who were paternal orphans, the risk of sexual debut was higher than for their counterparts who were non-orphans (ibid).

It is against this background that the present study sought to contribute to the discourse by focusing on female adolescents’ sexual debut and parental presence/absence, taking into account parental survival status. The study used nationally representative and comparable data from seven southern African countries. The study had the following hypotheses: 1) orphans have a higher risk of sexual debut compared with non-orphans, and 2) female adolescents who are living with both parents have reduced risk of sexual debut than their counterparts who do not. The objectives of the study were first to estimate the timing of sexual debut across the countries, second, to examine whether there was a significant difference in sexual risk across an array of living arrangements and third, to estimate the association between orphanhood status, living arrangements and timing of sexual debut.