Gates Open Res Gates Open Research 2572-4754 F1000 Research Limited London, UK 10.12688/gatesopenres.12972.1 Research Article Articles Timing between age at first sexual intercourse and age at first use of contraception among adolescents and young adults in Niger: What role do education and place of residence play?

[version 1; peer review: 2 approved, 1 approved with reservations, 1 not approved]

 Olakunle Abibatou Agbéké Methodology Writing – Original Draft Preparation https://orcid.org/0000-0003-0135-2211 a 1 Banougnin Boladé Hamed Conceptualization Formal Analysis Methodology 2 Department of Statistics, University of Ibadan, Ibadan, Oyo, Nigeria Pan African University, Life and Earth Science Institute (Including Health and Agriculture), University of Ibadan, Ibadan, Oyo, Nigeria abibatou.olakunle@outlook.com

No competing interests were disclosed.

 10 5 2019 2019 3 1463 2 5 2019 Copyright: © 2019 Olakunle AA and Banougnin BH 2019 This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background: Low contraceptive use among women in Niger is one of main causes of early childbearing and unwanted pregnancies, which affect maternal and child health. Education and place of residence have been cited as factors affecting modern contraceptive use.

Methods: We investigated the separate and joint effects of the place of residence and education on the time to modern contraceptive uptake among women aged 15-24 in Niger. The study used data from the second round of the 2016 Niger Performance Monitoring and Accountability 2020 (PMA2020) project. Survival analysis was applied for 830 women.

Results: Nelson-Aalen curves show that urban women had higher hazards of (and shorter delays in) modern contraceptive uptake as compared to their rural counterparts. Also, the higher the level of education, the higher the hazards of (and the shorter the delays in) modern contraceptive uptake. Findings from the multivariate (survival) analysis confirms these figures and provides the net effect of the place of residence on modern contraceptive uptake. Whether living in urban or rural areas of Niger, what matters more is the level of education.

Conclusions: Family planning programmes concerning adolescent and young women should focus more on women with no education and those that are illiterate.

 Education First sexual intercourse Niger place of residence survival analysis uptake of modern contraceptive women aged 15-24 Gates Foundation OPP1079004 OPP1181398 The PMA2020 project was supported by the Gates Foundation, Seattle [OPP1079004], who also supported publication of this article [OPP1181398]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Introduction

Niger presents worrying characteristics for youth sexual reproductive health. The country has the highest fertility rate as well as the lowest age for marriage and childbearing in Africa and the world ( Barroy et al., 2015). With a total fertility rate of 7.6 children per women and a population growth rate of 3.9%, Niger’s population has increased from 7,292,000 in 1989 to 21,311,000 in 2018 ( United Nations, 2019). Early marriage and, especially, childbearing are the main causes driving the high population growth and fertility in Niger ( Nouhou, 2016). In general, early marriage and childbearing is also associated with increased risk of maternal death and new-born death ( Gibbs et al., 2012). The 2018 Niger Performance Monitoring and Accountability 2020 (PMA2020) family planning brief reports that more than 70.9% of women aged 18–24 were married by age 18 ( INS - Niger, 2018). The brief also reports that one-third of Niger women aged 18–24 had their first birth by age 18. Moreover, estimates from UNICEF (2008) show that a woman’s lifetime risk of dying due to complications caused by pregnancy or childbirth was one in seven. This represents 14,000 deaths among Nigerien mothers from pregnancy-related causes. The maternal mortality rate in Niger was 553 per 100,000 live births in 2015 ( Roser & Ritchie, 2019b) and the under-five mortality rate was 84.5 deaths per 1,000 live births in 2017 ( Roser, 2019a). These figures reflect the poor health of both mother and child in Niger.

One of the major causes of maternal and childhood mortality remains the early age at first birth which typically depends on the age at first sexual intercourse. Evidence has shown that there is an inverse association between age at first sexual intercourse and early childbearing ( Gebreselassie & Govindasamy, 2013; Khatiwada et al., 2013; Meekers, 1993; Westoff, 2003). Data from the 2012 Niger Multiple Indicator Cluster Survey (MICS) and Demographic and Health Survey (DHS) show that 28% of women had already had their first sexual intercourse by age 15, and 45% had given birth by age 18 ( INS - Niger et al., 2013). These figures are among the highest in sub-Saharan Africa ( ICF International, 2018). This is an age at which adolescents are most susceptible to sexually transmitted infections, including HIV/AIDS and human papillomavirus and other health complications ( Fagbamigbe et al., 2015). The modern contraceptive is one of the means of promoting good reproductive health, especially among adolescents and young women.

Sub-Saharan Africa has one of the highest rates of teenage pregnancy ( Yakubu & Salisu, 2018) and the lowest contraceptive prevalence rates ( Radovich et al., 2018) in the world. The low contraceptive use among Niger adolescents (4%) and youths (14%) ( ICF International, 2018) is one of the main causes of early childbearing and unwanted pregnancies, which have many consequences including unsafe abortion, high maternal and child mortality and reduced earning potential and educational achievements. This calls for an increasing interest in expanding contraceptive use among adolescents and youths as advocated by the Family Planning 2020 global partnership ( Family Planning 2020, 2015).

The research literature has shown that the sooner the use of contraceptive methods, especially modern methods (pill, intrauterine device, injectables, female condom, male condom, female sterilization, male sterilization, implants and lactational amenorrhea), the higher the odds of preventing early childbirth ( Abma & Martinez, 2017; Guleria et al., 2017; Shu et al., 2016). Many factors affecting the use of contraceptive methods in the sub-Sahara African context have been observed. Many of them repeatedly cited women’s level of education and place of residence as major determinants of contraceptive use ( Fawole & Adeoye, 2015; Nair & Devi, 2017; Ochako et al., 2017). Most of these studies have examined the separated effect of the place of residence and education on contraceptive use. Likewise, few studies have been carried out on factors affecting the timing between first sexual intercourse and first use of contraceptive methods. Furthermore, the third target of Sustainable Development Goals (SDGs) aims to reduce maternal mortality and under-five mortality to less than 70 deaths per 100,000 live births and 25 deaths per 1,000 live births ( Gostin & Friedman, 2015). There is an expectation that through the use of modern contraceptive methods (from the very first sexual initiation), there will be an improvement in maternal and child health in Niger. This study aims to investigate both separated and joint effects of the place of residence and education on the timing between first sexual intercourse and first use of modern methods of contraception.

 Methods Data

The study draws on data from the Performance Monitoring and Accountability 2020 (PMA2020) project. PMA2020 supports low-cost, rapid-turnaround survey monitoring key indicators for family planning, water, sanitation and hygiene, and other health and development indicators in 11 low to middle income countries (for more details concerning the methods and objectives of PMA2020, see Johns Hopkins University, 2013). The PMA2020 project received approval to collect data from Institutional Review Boards at Johns Hopkins Bloomberg School of Public Health and the Niger Institute of Statistics.

We use data from the second round of Niger PMA2020 survey conducted from February to April 2016 ( Performance Monitoring and Accountability 2020 Project, 2016). The 2016 Niger PMA2020 survey used a two-stage cluster sample design with Niamey (the capital city of Niger), urban areas outside of Niamey, and rural areas as strata. This sample design enabled national and sub-national estimates to be produced. The study selected 84 enumeration areas (EAs) based on the 2012 Niger census’ sample frame, and randomly selected 35 households within each EA. Then, all resident women of reproductive ages within the sampled households were interviewed. The final dataset included 2,785 households (96.2%) and 3,048 women of reproductive ages (95.5%).

Our study population consisted of 1,238 women aged 15–24 at the time of the survey. Several studies in the areas of sexual and reproductive health have defined young women as those falling within the ages of 15–24 years ( Ali & Cleland, 2005; Dellar et al., 2015; Gouws et al., 2008; Lauby & Stark, 1988). Furthermore, young women undergo significant transitions in lifestyle, maturity, and legal rights between these ages; which places them at different vulnerabilities at different time points ( Dellar et al., 2015). Of the 1,238 women aged 15–24, 749 (60.5%) had already had their first sexual intercourse. Among these 749 women, we excluded 31 who reported having their first sexual intercourse before age 4, 26 who did not know when they first had sex, and 18 missing values. This resulted in a sample of 674 young women. From these, 671 were able to provide information on whether and, if so, when they have used a modern method of contraception for the first time. The corresponding weighted number of women aged 15–24 who had ever had sex is 830 1 (69.6% of women aged 15–24). None of the remaining 567 women (who had never had sex or did not give any consistent answer about the experience of the first sexual intercourse) use contraception.

 Measures

The 2016 Niger PMA2020 survey questionnaire was extensive and included sections on reproduction, pregnancy and fertility preferences, contraception, sexual activity, menstrual hygiene, and location. In particular, the questionnaire provided data on family planning use, of the type gathered in the Demographic and Health Surveys. Throughout this study, pill, intrauterine device, injectables, female condom, male condom, female sterilization, male sterilization, implants and lactational amenorrhea are referred as modern methods of contraception. The survey also collected information on the respondent’s socioeconomic characteristics, including educational attainment, marital status, and location.

The dependent variable of the study is the time to uptake of a modern contraceptive use since the first sexual intercourse. The time to uptake of a modern contraceptive use is calculated in years and is equal to the age at first modern contraceptive use minus the age at first sexual intercourse for modern contraceptive users. For non-users, the time to uptake of a modern contraceptive use is equal to the current age minus the age at first sexual intercourse. PMA2020 survey asks four questions to determine the time to uptake of a modern use: “How old were you when you first had sexual intercourse?”, “Have you ever done anything or tried in any way to delay or avoid getting pregnant?”, “How old were you when you first used a method to delay or avoid getting pregnant?”, and “Which method did you first use to delay or avoid getting pregnant?”.

The main independent variables are the place of residence and education. During the 2016 Niger PMA2020 survey, women were asked the following question about their educational attainment: “What is the highest level of school you attended?”. The possible response categories to this question were: ‘Never attended’, ‘Primary’, ‘Secondary’, and ‘Higher’. Only a very small proportion (0.4%) of young women had a higher level of education. To prevent non-accurate p-value for the test analysis, we combined the ‘Secondary’ and ‘Higher education’ categories into one (‘Secondary or higher’). The place of residence is divided into two categories: urban and rural areas. As mentioned above, the odds of modern contraceptive use usually depend (non-exhaustively) on the availability by location (urban/rural residence) and knowledge (education) ( Kandala et al., 2015; Kravdal, 2002; Rutenberg et al., 1991). In addition to education and place of residence, other variables (of control) such as age and household wealth tertile. Empirical data point out wealth as affordability factor, and age as confidence-to-buy factor affecting differences in modern contraceptive use across social groups ( O’Regan & Thompson, 2017; Wambui et al., 2009).

 Analytic methods

We used survival analysis since the main outcome of this study (time to modern contraceptive uptake) is a form of time to event. The survival time is assumed to begin when a woman has her first sexual intercourse until when she starts using modern contraceptives. Young women who have never used any modern contraceptive at the time of interview are right censored as of the date of the survey. Let T be a non-negative random variable representing the duration from first sexual intercourse to modern contraceptive uptake. The subjects at risk are the 671 (830 weighted) women aged 15–24, followed-up since their first sexual intercourse. The observation continues until time t. If a woman has taken up modern contraceptive, the time t is the time of modern contraceptive uptake; otherwise, the time t is the time of the survey (in 2016). The observation ends for a woman at time T = t if she has started using any modern contraceptive method. For example, T = 0 if a woman started using modern contraception before one year after her first sexual intercourse. Assuming that T is a continuous random variable with probability density function f( t), the cumulative distribution function (c.d.f), giving the probability that a woman has taken up modern contraceptive by duration t is:

F ( t ) = P r o b ( T ≤ t ) = ∫ 0 t f ( s ) d s ( 1 )

The complement of the c.d.f ( S( t) = 1 – F( t)), which represents the probability that a woman is yet to take up modern contraceptive by duration t, is the survival function. In this study, we defined the survival function in terms of the hazard function, which is the instantaneous rate of occurrence of the event of interest (here, the uptake of modern contraceptive since first sexual intercourse). The hazard function is defined as:

h ( t ) = d t → 0 lim ⁡ P r o b ( t < T < t + d t | T > t ) d t ( 2 )

The numerator of this expression is the conditional probability that a woman takes up modern contraceptive in the interval [ t,t + dt) given that she has not used modern contraceptive before. The denominator is the width of the interval. Dividing one by the other, we obtain a rate of uptake of modern contraceptive per unit of time (i.e. per year). Taking the limit as the width of the interval goes down to zero, we obtain an instantaneous rate of uptake of modern contraceptive.

The survival analysis for this paper proceeded in two steps. First, we used Nelson-Aalen nonparametric estimates of cumulative hazard rate function for the descriptive analysis. Especially, we drew cumulative hazard rate curves of time to modern contraceptive uptake for all women aged 15–24, and by groups (stratifying by education and place of residence). We used log-rank tests to test the equality of the cumulative hazard rate functions across groups. Second, we employed (multivariate) Cox semi-parametric proportional-hazards models to predict how the place of residence and education affect the time to uptake of modern contraceptive. The hazard rate in a Cox proportional-hazards model is defined as:

h ( t | x , β ) = h 0 ( t ) exp ⁡ ( x ′   β ) ( 3 )

where x is the vector of independent variables (education, place of residence, age, and wealth tertile), h 0( t) is the baseline hazard function, and β is the vector of coefficients. The multivariate analysis consisted of three different models. All models were controlled for age and wealth tertile. The first model (Model 1) presents the effect of the place of residence on the time to uptake of modern contraceptive use. The second model (Model 2) is equal to Model 1 plus the education variable. The third model (Model 3) shows how the interaction between the place and residence and education influences the timing to modern contraceptive uptake. We applied sampling weights for the analysis. Stata 13 was used to analyse the data.

 Results

We first report descriptive statistics of the variables used in the analysis in Table 1. Mean age at the first sexual intercourse among women aged 15–24 was 19.7 years. Just under one-quarter (197 out of 830; 23.7%) of the respondents (ever-sexually-active women aged 15–24) had used contraception. Over three-quarters of (first-time) contraceptive users (156 out of 197) had practiced modern methods. The pill is the most widely used modern method, accounting for more than three-fifths of modern contraceptive use and 11.8% of the respondents. Of all women aged 15–24, very few (0.6%) had used condoms for their first contraceptive experience. Table 1 also indicates that the vast majority of the respondents (86.4%) were living in rural areas. At the time of the survey, only 12.2% of ever-sexually-active women aged 15–24 had attained secondary or higher level of education, whereas more than two-thirds (68.3%) had never attended school. The mean time between first sexual intercourse and first use of modern contraceptive was 3 years with a standard deviation of 2.1 years.

Distribution of factors concerning contraceptive use and demographic variables.
Variable * Number of weighted observations = 830
Percentage Mean Std. Dev.
Contraceptive use
Modern methods 18.7
  Implants 1.8
  Injectables 4.6
  Pill 11.8
  Male condoms 0.6
Traditional 4.9
  Lactational Amenorrhea Method (LAM) 0.2
  Rhythm 0.1
  Withdrawal 0.1
  Other traditional 4.6
Total users 23.7
Nonusers 75.3
Unknown 1.0
Place of residence
  Urban 13.6
  Rural 86.4
Education
  No education 68.3
  Primary 19.5
  Secondary or higher 12.2
Wealth tertile
  Poorest 33.0
  Middle 31.9
  Richest 35.1
Age, years 19.7 2.3
Time to modern contraceptive uptake, years (a) 3.0 2.1
Time between first sex and interview date, years (b) 4.4 2.7
Total 100.0

*No variable contains missing value; (a) For women who have ever used modern contraceptive (18.7%); (b) For women who have never used modern contraceptive (75.3%). Std. Dev., standard deviation.

We then compared the incidence rate of modern contraceptive uptake among ever-sexually-active young women by place of residence ( Figure 1B) and education ( Figure 1C). Figure 1B reveals two important findings. First, the incidence rate of modern contraceptive uptake increases very fast in the first six years after sexual initiation. Between six and eight years post sexual initiation, the incidence rate continues to rise steadily though slower than the first six years. Secondly, the incidence rate of modern contraceptive uptake since sexual initiation was higher among urban women than rural women. In addition, the cumulative hazard curve of modern contraceptive uptake reached its peak faster among urban women (61.8% at the tenth year) than rural women (37.8% at the fourteenth year). As expected, the incidence rate of modern contraceptive uptake is higher among women with higher education than those with no education ( Figure 1C). Also, the higher the educational level, the faster the modern contraceptive uptake. Figure 1D shows how interaction between place of residence and education is associated with the chances of modern contraceptive uptake. Urban women with primary or higher education and rural women secondary or higher education had the highest incidences of and shorter delays in modern contraceptive uptake. The Log-rank test for equality of survival functions shows significant differences in the cumulative hazard function of place of residence and education.

Nelson-Aalen curve showing the cumulative hazard of modern contraceptive uptake among women aged 15–24 in Niger, according to the place of residence and education.

Table 2 presents the results of Cox semi-parametric proportional-hazards models of the effects of the place of residence and education on time to modern contraceptive uptake. We fitted three models for the survival data. Model 1 was restricted to place of residence (with wealth and age as control variables). Significant differences are observed between urban and rural women. The probability of modern contraceptive uptake was 1.73 (95% CI: 1.09–2.73) times higher for urban women as compared with rural women. According to the wealth status, women in the richest wealth tertile were more than twice as likely to take up modern contraceptive than women in the poorest wealth tertile. The chances of modern contraceptive uptake increased by 11% per one-year increase in age, all other factors being equal.

In Model 2, we introduced education, which was not included in Model 1. There was no significant difference in modern contraceptive uptake between urban and rural women after including education. Findings from Model 2 in Table 2 show that the higher the education the higher the hazard of modern contraceptive initiation. The hazard ratio (HR) of modern contraceptive uptake was higher among women with primary education (HR = 1.80, p < 0.01, 95% CI: 1.20–2.72) and women with at least secondary education (HR = 2.30, p < 0.01, 95% CI: 1.43–3.70) compared with their peers with no education.

Relative risks (hazard ratios) of modern contraceptive uptake among women aged 15–24 in Niger.
Variable Model 1 Model 2 Model 3
HR (95% CI) HR (95% CI) HR (95% CI)
Place of residence
    Urban 1.73 ** (1.09–2.73) 1.46 (0.92–2.33)
    Rural 1 1
Education
    No education 1
    Primary 1.80 *** (1.20–2.72)
    Secondary or higher 2.30 *** (1.43–3.70)
Place of residence * Education
    Urban * No education 1.72 (0.87–3.40)
    Urban * Primary 3.17 *** (1.54–6.53)
    Urban * Secondary or higher 2.87 *** (1.42–5.79)
    Rural * No education 1
    Rural * Primary 1.74 ** (1.08–2.79)
    Rural * Secondary or higher 2.82 *** (1.60–4.96)
Wealth tertile
    Poorest 1 1 1
    Middle 1.46 (0.91–2.33) 1.52 * (0.95–2.44) 1.50 * (0.94–2.40)
    Richest 2.18 *** (1.37–3.49) 1.99 *** (1.23–3.22) 1.88 ** (1.15–3.08)
Age 1.11 ** (1.02–1.21) 1.10 ** (1.01–1.21) 1.11 ** (1.01–1.21)
Subjects 773
Time at risk 3,409
Failures 140

*p<0.1; **p<0.05; ***p<0.01. HR, hazard ratio; CI, confidence interval.

Model 3 included the interaction terms between place of residence and education without the separated variables (education and place of residence). Urban women with primary education had the highest hazard of modern contraceptive uptake (HR = 3.17, p < 0.01, 95% CI: 1.54–6.53). They are followed by their peers with secondary or higher education (HR = 2.87, p < 0.01, 95% CI: 1.42–5.79) and rural women with secondary or higher education (HR = 2.82, p < 0.01, 95% CI: 1.60–4.96). Also, women with no education had similar hazard rates of modern contraceptive uptake whether living in urban or rural areas. In other words, the educational status is the most discriminating factor for modern contraceptive uptake.

 Discussion

Identifying investments needed to improve family planning programmes in developing countries remains a source of concern for health advocates and professionals, policies makers and funding agencies ( Sedgh et al., 2016). Improvement in modern contraception use should enable women—irrespective of their place of residence and socioeconomic status—to prevent reproductive health complications. This study used data from a nationally representative sample of Niger young women (aged 15–24 years) to examine how the place of residence and education, as well as their interaction, affect the time to uptake of modern contraceptives. Findings from the analysis show that residing in urban areas, having a high educational level and living in the richest wealthy tertile are associated with higher hazard rates of and shorter delays in modern contraceptive uptake. Moreover, the hazard ratio of using modern contraceptive among young women also increases with age.

Findings from multivariate (survival) analysis showed a minor effect of the place of residence on the time to modern contraceptive uptake when adjusted for education (and the main control variables: age and wealth tertile). Indeed, the difference between urban and rural women in modern contraceptive uptake become non-significant after adjusting for the level of education. Thus, the level of education seems to counterbalance the effect of increased time to uptake of modern contraceptives among rural women (ages 15–24) in Niger.

Therefore, the positive relationship between education and the hazard rates of modern contraceptive uptake confirms previous work on the association between education and contraceptive use ( Khan & Mishra, 2008; O’Regan & Thompson, 2017; Selassie, 2017). Better education is not only positively associated with early use of contraception, but also contributes to empowerment of women ( Family Planning 2020, 2014; Tadesse et al., 2013; Tuladhar et al., 2013). Also, women who attend at least a primary level of education (and those who are literate) are more likely to know about family planning services through media ( Fagbamigbe et al., 2015; Shapiro & Tambashe, 1994). Finally, we further our analysis by introducing interaction terms between the place of residence and education. In so doing, statistically significant differences emerged. Urban women with at least primary education and rural women with secondary or higher education have approximatively the same hazards of modern contraceptive uptake. This confirms the key role played by education in modern contraceptive uptake among young women in Niger.

 Conclusion

Our results suggest that the delay in modern contraceptive uptake is relatively high. This delay is higher among rural and less educated women than among urban and more educated women. Therefore, adolescents and youths in Niger, especially those living in rural areas and those with a low educational level, are at risk of early childbirth and its consequences. These consequences include among others, child and maternal mortality, school dropout and low participation in the labour market. In consequence, there is a need for policies and programmes that empower adolescents and youths through improving information about the access to, and utilization of reproductive health services. It is necessary to reduce all barriers to access and improve the contraceptive distribution system by integrating family planning activities into national policies on reproductive health with more focus on women with no education.

 Data availability

Underlying data used in this study were obtained from PMA2020 (Niger Round 2, 2016) ( Performance Monitoring and Accountability 2020 Project, 2016). All PMA2020 datasets are free to download and use, although users are required to register for a PMA2020 dataset account. The request form must include a brief description of the research or analysis that the user would like to conduct using the requested data. Information about the data and terms of use are available here.

1 The PMA2020 sampling design follows DHS’ standard, and as described in https://dhsprogram.com/pubs/pdf/DHSG1/Guide_to_DHS_Statistics_DHS-7.pdf, such data must be weighted. The PMA2020 also includes the sampling weight variable. Weighting the female data, the corresponding number of observations is 830.

 Abma JC Martinez G : Sexual Activity and Contraceptive Use Among Teenagers in the United States, 2011-2015. Natl Health Stat Report. 2017; (104):1–23. 28696201 Ali MM Cleland J : Sexual and reproductive behaviour among single women aged 15–24 in eight Latin American countries: a comparative analysis. Soc Sci Med. 2005;60(6):1175–1185. 15626515 10.1016/j.socscimed.2004.07.002 Barroy H Cortez R Karamoko D : Adolescent Sexual an Reproductive Health in Niger.2015. 10.13140/RG.2.1.4114.9927 Dellar RC Dlamini S Karim QA : Adolescent girls and young women: key populations for HIV epidemic control. J Int AIDS Soc. 2015;18(2 Suppl 1): 19408. 25724504 10.7448/IAS.18.2.19408 4344544 Fagbamigbe AF Adebowale AS Morhason-Bello I : Survival analysis of time to uptake of modern contraceptives among sexually active women of reproductive age in Nigeria. BMJ Open. 2015;5(12):e008371. 26671948 10.1136/bmjopen-2015-008371 4679946 Family Planning 2020: Family Plannign 2020: Rights and Empowerment Principles for Family Planning. 2015; Watts and Mayhew, “Reproductive Health Services and Intimate Partner Violence.” Washington, DC: FP2020.2014. Reference Source Family Planning 2020: 2020 (FP2020), Progress report 2014–2015: commitment to action. Washington, DC: FP2020,2015. Fawole OI Adeoye IA : Women’s status within the household as a determinant of maternal health care use in Nigeria. Afr Health Sci. 2015;15(1):217–225. 25834551 10.4314/ahs.v15i1.28 4370136 Gebreselassie T Govindasamy P : Levels and trends in unmet need for family planning among adolescents and young women in Ethiopia: Further analysis of the 2000, 2005, and 2011 Demographic Health Surveys. Retrieved from Calverton, Maryland, USA;2013. Reference Source Gibbs CM Wendt A Peters S : The impact of early age at first childbirth on maternal and infant health. Paediatr Perinat Epidemiol. 2012;26 Suppl 1:259–284. 22742615 10.1111/j.1365-3016.2012.01290.x 4562289 Gostin LO Friedman EA : The Sustainable Development Goals: One-Health in the World's Development Agenda. JAMA. 2015;314(24):2621–2622. 26660329 10.1001/jama.2015.16281 Gouws E Stanecki KA Lyerla R : The epidemiology of HIV infection among young people aged 15-24 years in southern Africa. AIDS. 2008;22 Suppl 4:S5–S16. 19033755 10.1097/01.aids.0000341773.86500.9d Guleria S Juul KE Munk C : Contraceptive non-use and emergency contraceptive use at first sexual intercourse among nearly 12 000 Scandinavian women. Acta Obstet Gynecol Scand. 2017;96(3):286–294. 28029175 10.1111/aogs.13088 ICF International: The DHS Program STATcompiler. Funded by USAID. STATcompiler.2018. Reference Source INS - Niger: Family Planning Brief. PMA2020. 2018. Reference Source INS - Niger, UNICEF/Niger, ICF International: Enquête Démographique et de Santé dans les Zones d'Intervention du Programme de Cooperation de l'UNICEF au Niger, 2012. Retrieved from Rockville, Maryland, USA.2013. Reference Source Johns Hopkins University: PMA2020: Performance Monitoring and Accountability 2020.2013. Reference Source Kandala NB Lukumu FK Mantempa JN : Disparities in modern contraception use among women in the Democratic Republic of Congo: a cross-sectional spatial analysis of provincial variations based on household survey data. J Biosoc Sci. 2015;47(3):345–362. 24911333 10.1017/S0021932014000212 Khan S Mishra V : Youth reproductive and sexual health. Retrieved from Calverton, Maryland, USA:2008. Reference Source Khatiwada N Silwal PR Bhadra R : Sexual and reproductive health of adolescents and youth in Nepal: Trends and determinants: Further analysis of the 2011 Nepal Demographic and Health Survey. Retrieved from Calverton, Maryland, USA:2013. Reference Source Kravdal Ø : Education and fertility in sub-Saharan Africa: individual and community effects. Demography. 2002;39(2):233–250. 12048950 10.1353/dem.2002.0017 Lauby J Stark O : Individual migration as a family strategy: Young women in the Philippines. Popul Stud. 1988;42(3):473–486. 10.1080/0032472031000143596 Meekers D : Sexual initiation and premarital childbearing in Sub-Saharan Africa. Retrieved from Columbia, Maryland, USA:1993. Reference Source Nair A Devi S : Knowledge and atitude of puerperal women towards family planning practices in a tertiary care hospital. J Evid Based Med Healthc. 2017;4(5):212–216. 10.18410/jebmh/2017/41 Nouhou AM : Projet de famille et processus d'autonomisation des individus en matière de fécondité au Niger. (Ph.D. Thesis), University of Geneva.2016. Reference Source O’Regan A Thompson G : Indicators of young women’s modern contraceptive use in Burkina Faso and Mali from Demographic and Health Survey data. Contracept Reprod Med. 2017;2(1):26. 29201431 10.1186/s40834-017-0053-6 5683538 Ochako R Temmerman M Mbondo M : Determinants of modern contraceptive use among sexually active men in Kenya. Reprod Health. 2017;14(1):56. 28449723 10.1186/s12978-017-0316-3 5408470 Performance Monitoring and Accountability 2020 Project, [Institut National de la Statistique (INS)]. [Niger]. Baltimore, MD PMA2020, Bill & Melinda Gates Institute for Population and Reproductive Health, Johns Hopkins Bloomberg School of Public Health.2016. Radovich E Dennis ML Wong KL : Who Meets the Contraceptive Needs of Young Women in Sub-Saharan Africa? J Adolesc Health. 2018;62(3):273–280. 29249445 10.1016/j.jadohealth.2017.09.013 Roser M : Child Mortality. from Our World in Data.2019a. Reference Source Roser M Ritchie H : Maternal Mortality. from Our World in Data.2019b. Reference Source Rutenberg N Ayad M Ochoa LH : Knowledge and use of contraception.1991. Reference Source Sedgh G Ashford LS Hussain R : Unmet Need for Contraception in Developing Countries: Examining Women’s Reasons for Not Using a Method.New York: Guttmacher Institute.2016. Reference Source Selassie TG : Determinants of contraceptive use among urban youth in Ethiopia. Ethiop J Health Dev. 2017;10(2):97–104. Reference Source Shapiro D Tambashe BO : The Impact of Women's Employment and Education on Contraceptive use and Abortion in Kinshasa, Zaire. Stud Fam Plann. Population Council.1994;25(2):96–110. 8059449 10.2307/2138087 Shu C Fu A Lu J : Association between age at first sexual intercourse and knowledge, attitudes and practices regarding reproductive health and unplanned pregnancy: a cross-sectional study. Public Health. 2016;135:104–113. 26927825 10.1016/j.puhe.2016.01.021 Tadesse M Teklie H Yazew G : Women's empowerment as a determinant of contraceptive use in Ethiopia: Further analysis of the 2011 Ethiopia Demographic and Health Survey. Retrieved from Calverton, Maryland, USA.2013. Reference Source Tuladhar S Khanal KR Lila KC : Women's empowerment and spousal violence in relation to health outcomes in Nepal: Further Analysis of the 2011 Nepal Demographic and Health Survey. Retrieved from Calverton, Maryland, USA.2013. Reference Source UNICEF: UNICEF supports maternal and newborn health in Niger.2008. Reference Source United Nations.2019. Reference Source Wambui T Ek AC Alehagen S : Perceptions of family planning among low-income men in Western Kenya. Int Nurs Rev. 2009;56(3):340–345. 19702808 10.1111/j.1466-7657.2009.00726.x Westoff CF : Trends in marriage and early childbearing in developing countries. Retrieved from Calverton, Maryland, USA:2003. Reference Source Yakubu I Salisu WJ : Determinants of adolescent pregnancy in sub-Saharan Africa: a systematic review. Reprod Health. 2018;15(1):15. 29374479 10.1186/s12978-018-0460-4 5787272 10.21956/gatesopenres.14079.r28037 Reviewer response for version 1 Mahande Michael J. 1 Referee https://orcid.org/0000-0002-7750-7657 Department of Epidemiology and Biostatistics, Institute of Public Health, Kilimanjaro Christian Medical University College, Moshi, Tanzania

Competing interests: No competing interests were disclosed.

 28 10 2019 Copyright: © 2019 Mahande MJ 2019 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation reject

Section

Comment, question, suggestions.

Abstract

Well-constructed clear and short abstract.

Easily understandable.

 Introduction

Background

Clear background with enough information.

A broader explanation as to what we can’t do by not having the data on timing would be important to the readers i.e. what is the knowledge gap?

 Objectives

It was better to clearly explain what was meant by separate and joint effects of residence and education.

 Methodology

WHO and UNICEF have standard definitions of adolescent and youths. It would better to use standard references from the WHO/UNICEF rather than citing papers.

The information regarding exclusion criteria would have been easier to understand if presented using a flow chart rather than narratives.

Why only focus on age and residence as the adjusting variables? What about other independent variables? Such as partner information (age and education levels), Family Planning media exposure, marital status, religion etc? All these have also been shown to affect modern contraceptive use. I am concerned with the aspect of residual confounding caused by leaving these variables out from the analysis.

The dependent variable was measured in years. Is it complete years? What if a woman started using a modern method 3 months after her first sexual encounter? How would she be recorded?

Do you mean multi-variate or multi-variable analysis? Please check these as they are totally different

I have some few concerns with the model building process.

Generally, model 1 should be an empty model containing the crude estimates of an exposure variable on the outcome, in this case a model with education alone or a model with residence alone.

Model 2 continues to start adjusting for other exposure variables and estimating confounding effect when an exposure var results in a 10%- 15% change in the estimates of the primary predictor of interest.

An interaction is looked at based on plausibility of its effect, for instance, do we have reason to believe that modern contraceptive use differs among urban women based on the levels of their education? Similarly, with women residing in rural areas.

Finally, was model diagnostics performed for the Cox-ph models? Was the Ph assumption holding or was it violated?

Was the data both st and svyset?

 Results

In the narratives you mention the frequencies of the variables, these are not shown in table 1. Would be better to present them as well since they are present in your narratives.

Would have also been interesting to show during the first sexual encounter, how many women utilized modern contraceptives?, and who are these women? (if this information is available in the dataset).

Why was it important to show in table 1 the time between first sexual encounter and the date of interview? What was this information telling the readers?

The information on median time to contraceptive use should appear earlier in the narratives since it is of much interest.

Please correct the y-axis in figure 1 to say cumulative hazards and not probabilities. Nelson Aalen curves are not probabilities.

Please indicate the log-rank p-values on the graphs in order to show the significance difference in modern contraceptive uptake.

Most of the interpretations are in-correct. For survival analysis cox Ph models we interpret the “Hazards”, and not probabilities/chance/likely. Please correct this -major comment.

For model 3: Why did you leave out the main effects and only included the interaction term alone? This information is necessary and important.

Why focus on a significance level of <0.1 rather than the tradition <0.05 for table 2? Please provide description in the methodology under data analysis section.

“In other words, the educational status is the most discriminating factor for modern contraceptive uptake.”

 

Since this is an interaction, it indicates that modern contraceptive use among urban women differs according to their levels of education and similarly, among modern contraceptive use, rural women differs according to their levels of education.

An interaction shows that the outcome among each level of an exposure variable differs across the levels of a third variable.

Which means that residence has an effect on modern contraceptive use differently across education categories.

Had you included the main effects in this model 3, it would also have added some information on the significance of the individual variables.

Discussion

Can be better synthesized to bring out the meaning of the results and their implications to policy and practice.

Strengths and limitations

Missed the section on strengths and limitations of the study.

Conclusion + Recommendations

Conclusion and recommendation were focused on variables which were not assessed in the study i.e. FP media exposure, roles of empowerment of women and removing barriers to access and distribution of modern contraceptives. All these were not measured in the present study.

Is the work clearly and accurately presented and does it cite the current literature?

Partly

If applicable, is the statistical analysis and its interpretation appropriate?

Yes

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Partly

Are sufficient details of methods and analysis provided to allow replication by others?

Partly

Reviewer Expertise:

Epidemiology and Biostatistics

I confirm that I have read this submission and believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

 10.21956/gatesopenres.14079.r28038 Reviewer response for version 1 Salisu Waliu Jawula 1 Referee https://orcid.org/0000-0001-9725-4936 School of Nursing and Midwifery, Tehran University of Medical Sciences, Tehran, Iran

Competing interests: No competing interests were disclosed.

 15 10 2019 Copyright: © 2019 Salisu WJ 2019 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve

The paper sheds light on the relationship between education and place of residence concerning modern contraceptive use among young women in Niger.

The author’s highlight on the gap in the literature however, seems to focus primarily on Sub Saharan Africa rather than the original study context. The gap in the literature that necessitates the study in Niger is unclear. A little emphasis on the literature gap in Niger is needed.

In the results section, the following statement meant to describe Figure 1 seems misleading since it did not feature any percentages on either axis of the figures. “…the cumulative hazard curve of modern contraceptive uptake reached its peak faster among urban women (61.8% at the tenth year) than rural women (37.8% at the fourteenth year)...”. I suggest the authors take note and amend it appropriately.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

I cannot comment. A qualified statistician is required.

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Nursing and Midwifery Education.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

 10.21956/gatesopenres.14079.r27393 Reviewer response for version 1 Abdoul Moumouni Nouhou 1 Referee Organiser l'avancement des solutions au Sahel, L'Initiative OASIS Niger, Niamey, Niger

Competing interests: No competing interests were disclosed.

 8 7 2019 Copyright: © 2019 Abdoul Moumouni N 2019 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve

Model 3 explains the modern contraceptive uptake by the (education)*(place of residence) interaction (controlling by wealth tertile and age). What about the marginal effects of both education and place of the residence in this model? These two factors may still have some marginal effects, something to be checked by adding them into the Model 3, along with the interaction terms (let’s call this Model 4). You could estimate this Model 4 and not to include the results in the Table 2, but rather explain and comment them in few sentences (what did you find? What are the implications). If this is not feasible because of a "small sample size" problem (preventing non-accurate p-value), it should be mentioned clearly.

The Model 3 shows that “urban women with primary education had the HIGHEST hazard of modern contraceptive uptake (HR = 3.17, p < 0.01, 95% CI: 1.54–6.53)”. Indeed, when comparing the different sub-groups from Model 3, the Relative risks (RR) could be ranked as (Urban * Primary) > (Urban * Secondary or higher) =~ (Rural * Secondary or higher). This important result was not sufficiently discussed. According to these results, when a woman attains a secondary or higher education level, the place of residence is of no importance whatsoever when it comes to modern contraceptive uptake. On the other hand, even with a primary level, women living in urban areas have a greater chance of turning to modern contraception faster after their first sexual intercourse. This result highlights an important role of the urban environment (accessibility to methods, quality, confidentiality, etc., see PMA2020 data on the quality of FP services in Niamey and complete the discussions section accordingly).

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Yes

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Fertility and family planning, statistics, demography

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

 10.21956/gatesopenres.14079.r27169 Reviewer response for version 1 Lindstrom David Philip 1 Referee Population Studies Centre, Brown University, Providence, RI, USA

Competing interests: No competing interests were disclosed.

 1 7 2019 Copyright: © 2019 Lindstrom DP 2019 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

The paper addresses an important topic and the analysis is reasonably constructed, but there are a number of problems with the presentation and interpretation of the results.

The purpose of the paper, which is stated at the end of the last paragraph before the Methods section, should appear sooner in the paper. Perhaps, at the end of the first paragraph.

For the table with the descriptive statistics, the number of observations with weights should be the same as the unweighted number of observations. Multiplying the weights by the ratio of the unweighted number of observations and the weighted number of observations will preserve the relative differences in the weights and not change the total n.

Figure 1 presents cumulative hazards. Cumulative hazards are not probabilities. The vertical axis title should be changed to cumulative hazard. In the text the authors interpret the cumulative hazards as percents - this is not correct). 

The log rank test is a test of the null hypothesis of no group differences. It is not a test of comparisons between groups when there are more than 2 groups. Rejecting the null hypothesis when there are 3 or more groups means that at least one group is different from the others.

The authors interpret the hazard ratios in Table 2 in terms of probabilities. Strictly speaking, the hazard ratio compares two hazards, which are instantaneous failure rates and not necessarily probabilities.

The authors find that place of residence is not significant when education is included in the model. Why then are interactions between place of residence and education tested?

A proper test of the interactions between place of residence and education would also include the main effects in the model. Including the main effects provides a clearer test of whether there are indeed any significant differences in the effects of education by place of residence than a model that only includes interaction terms.

Is the work clearly and accurately presented and does it cite the current literature?

Partly

If applicable, is the statistical analysis and its interpretation appropriate?

Partly

Are all the source data underlying the results available to ensure full reproducibility?

Partly

Is the study design appropriate and is the work technically sound?

Partly

Are the conclusions drawn adequately supported by the results?

Partly

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Reproductive health, adolescents, migration.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.