Our study included 97,161 teenage mothers who gave birth to and registered a child between 2016 and 2020 in South Africa and had a paternity acknowledgement. Our sample showed only two mothers aged 11 at the time of delivery had a paternity acknowledgement, 17 mothers who were 12-years-olds and 131 mothers who were 13-year-olds had a paternity acknowledgement during the child’s birth registration. Table 2 further shows paternity acknowledgement of children born to teenage girls declined over the years, with the lowest period in 2020 (n = 13,436) when compared to 2016 (n = 25,133). The majority (85.26%) of birth registrations happened within the year (a period not exceeding 12 months) in which the birth occurred.

Table 2 Frequency and percentage distribution of teenage mothers who had a birth between 2016 and 2020 in South Africa and their male partners Who are the men who father children with teenage girls in South Africa

The mean age of fathers included in our study is 24 years. Our findings show the majority of men who fathered children with teenage girls are men between the ages of 20 and 24 years (54.41%, n = 52,865) and 25 and 29 years (23.56%, n = 22,891), this is presented in Table 2. Of the teenage boys also fathered children with teenage girls during this period, our findings show 10.29% (n = 10,010) of fathers of children born to teenage mothers were fathers under the age of 20 between 2016 and 2020.

Our findings from the Pearson’s chi-square tests (Table 3) revealed that—apart from the two 11-year-old mothers—the younger the age of the mother, the higher the proportion of peer fathers. The proportion of girls who had children fathered by their peers reduced as the age of teenage girls increased, where 29.41% of peer fathers had children with 12-year-old mothers and 5.75% peer fathers fathered children with 19-year-old mothers (Table 3). The younger the mother, the higher the likelihood that the father of the child they had was a teenage boy. This is further illustrated by the results from our bivariate regression analysis presented in Table 4, where for every unit increase of mothers age, so too increase the odds of having children with a non-peer father (older men). The results show 36% reduced odds for a 12-year-old mother having a child with a non-peer father (UOR 0.54, CI 0.19–1.53, p value 0.0001) and double the odds for an 18-year-old mother (AOR 1.84, CI 1.69–1.90, p value 0.0001) having a child with a non-peer father (Table 4).

Table 3 Distribution of peer fathers and non-peer fathers of children born to teenage mothers between 2016 and 2020Table 4 Results of bivariate regression to establish the association between father category (non-peer father) and independent variables (unadjusted odds ratio) and the results of the binary logistic regression model (adjusted odds ratio) to establish the relationship and odds of non-peer fathers (older men) having a child with teenage girls Age gap between fathers and teenage mothers

Figure 1 presents the results of the variable created using fathers age and mothers age to illustrate the age gap between teenage mothers and their partners. The majority of parents of children born to teenage mothers had a three-year age gap (12.06%), four-year age gap (12.02%) and five-year age gap (11.07%). More than half the parents of children born to teenage mothers (54.91%) had a five year age gap. Very few (0.91%) fathers are younger than the teenage mother whom they fathered children with. Our findings also reveal that n = 2693 parents are age mates, this means the teenage mother and father are exactly the same age. Figure 1 shows cumulatively, 64.17% had a 5-year age gap, 85.55% a 10-year age gap, 94.77% a 15-year age gap and 97.98% a 20-year age gap between teenage mothers and the men who fathered children with them.

Fig. 1

Age gap in years between teenage mothers and the fathers of children born to teenage mothers in Bojanala between the years 2016 and 2020

Where do the men who have children with teenage girls live

The majority of men who fathered children with teenage girls in South Africa are located in South Africa’s rural provinces of KwaZulu Natal (39.02%), followed by Mpumalanga (19.36%) and Limpopo (14.13%), as presented in Fig. 2. Figure 2 is a map of where men who have fathered children with teenage girls reside, presented by percentage. The district with the highest number of men who have fathered children with teenage girls is Ehlanzeni municipal district (11.08%) in Mpumalanga, followed by Vhembe municipal district (9.8%) in Limpopo and Umkhanyakude municipal district (6.95%) in KwaZulu-Natal (Fig. 2). The Person’s chi-squared tests conducted show the largest proportion of peer fathers reside in the Western Cape (13.07%), KwaZulu Natal (11.02%) and Mpumalanga (10.67%) provinces (Table 3). In all provinces, the results from the bivariate regression tests and the binary logistic regression model show non-peer fathers have increased odds of fathering children with teenage girls when compared to peer fathers. Stratified analysis shows the North West province has the highest odds of older men having children with teenage girls in South Africa; non-peer fathers in the North West province have close to threefold odds of fathering children with teenage girls (OR 2.81, CI 2.17–3.64, p value-0.0001).

Fig. 2

Heat map of fathers of children born to teenage mothers between the years 2016 and 2020 by district municipality

The profile of fathers over time

Our study observed a pattern concerning the age of men who have children with teenage girls over time. The results of the Pearson chi-square test showed the number of older men having children with teenage girls increased between 2016 (83.99%) and 2020 (99.49%), while peer fathers having children with teenage mothers declined during the same period between 2016 (16.01%) and 2020 (0.51%) as presented in Table 3. Our results further showed that non-peer fathers had 19% increased odds (UOR 1.19, CI 1.13–1.25, p value 0.0001) of having a child with teenage mothers in 2017, 57% odds in 2018 (UOR 1.57, CI 1.48–1.66, p value 0.0001), threefold increased odds in 2019 (UOR 3.69, CI 3.42–3.98, p value 0.0001) and almost 37-fold increased odds in 2020 (UOR 36.94, CI 29.09–48.91, p value 0.0001). The same trend is observed in a binary logistic regression model, where the odds of older men fathering children with teenage girls in the country increased annually.

Time from birth occurrence to birth registration

The time between birth occurrence and registration of births was analysed. Our findings show a positive outcome concerning birth registration patterns for children born to teenage mothers in the country; 85.26% of births were registered within the year of the child’s birth. Our findings show children of peer fathers had a higher representation in the five years to registration category as opposed to non-peer fathers who had the highest representation in the zero time to registration category. The unadjusted odds ratios presented in Table 4 show children with non-peer fathers had increased odds of registering within the zero years to registration (UOR 2.02, CI 1.64–2.50, p value 0.0001) and one year to registration (UOR 1.99, CI 1.60–2.48, p value 0.001).