Abstracts

Background

The adolescent fertility rate (AFER) is a key dimensions of adolescents' reproductive health, and is usually linked to socioeconomic status (SES).

But only a few studies follow the evolution of inequality in adolescent fertility across time, and they use follow-up data that is not available in most countries, and use absolute rather than relative measures of SES.

Absolute measures, such as education level and race, are likely to change their “prevalence” across time ¹1. Manlove J, Ikramullah E, Mincieli L, Holcombe E, Danish S. Trends in sexual experience, contraceptive use, and teenage childbearing: 1992-2002. J Adolesc Health 2009; 44:413-23. and are not standard measures across countries, jeopardizing the comparison of health outcomes. A relative measure, like the rank-position of the observation in a socioeconomic ladder, can be compared across time and places.

This study quantitatively describes the evolution of AFER by SES, and, to support the findings, the evolution of adolescents' sexual initiation (ASEX) by SES is also included. AFERs and ASEX are closely related. ASEX is usually regarded as a proximate determinate of AFER ²2. United Nations. Adolescent fertility since the International Conference on Population and Development (ICPD) in Cairo. http://www.un.org/en/development/desa/population/publications/pdf/fertility/Report_Adolescent-Fertility-since-ICPD.pdf (accessed on 10/Mar/2014).
http://www.un.org/en/development/desa/po... . In developed countries, AFERs are probably the only fertility rates where sex plays a major role. Indeed, the median age at sexual initiation has decreased in recent decades in Chile ³3. Rodríguez J. Variables intermedias de la fecundidad adolescente en Chile: Evolución, desigualdad e implicancias de políticas. Rev Chil Salud Pública 2009; 13:127-35., while AFERs have remained stable or even increased ⁴4. Instituto Nacional de Estadísticas. Demographic yearbook 2013. Santiago de Chile: Instituto Nacional de Estadísticas; 2014.,⁵5. Ministerio de Salud de Chile. Situación actual del embarazo adolescente en Chile. 2012.http://www.mineduc.cl/usuarios/convivencia_escolar/doc/201210251259030.situacion_actual_emba razo_adolescente_en_chile.pdf (accessed on 10/Mar/2014).
http://www.mineduc.cl/usuarios/convivenc... .

This study aims to describe quantitatively the evolution of the relative-SES ratio of adolescent (15-17 years old) fertility rates and sexual initiation in Chile on each year between and 1990-2011 or 1997-2012, respectively.

Methods

AFERs by SES are difficult to measure. AFERs are small (less than 4% of 15-17 year old women have a child in any recent given year in Chile ⁵5. Ministerio de Salud de Chile. Situación actual del embarazo adolescente en Chile. 2012.http://www.mineduc.cl/usuarios/convivencia_escolar/doc/201210251259030.situacion_actual_emba razo_adolescente_en_chile.pdf (accessed on 10/Mar/2014).
http://www.mineduc.cl/usuarios/convivenc... ), so either large surveys or vital records are required to estimate them. However, large surveys are costly and vital records usually do not include rich socioeconomic information, aside from a few variables regarding parents' education and employment status. Measuring adolescents' SES requires somewhat complex socioeconomic data ⁶6. Currie C, Molcho M, Boyce W, Holstein B, Torsheim T, Richter M. Researching health inequalities in adolescents: the development of the Health Behaviour in School-Aged Children (HBSC) family affluence scale. Soc Sci Med 2008; 66:1429-36.. However, if the vital records include complete data on place of residence, a geographic SES can be computed and used to compare across time and countries, in a relatively easy manner, the ratio of AFERs (or another health measure) between the X% of the population living in the lowest and highest income areas.

Chilean comunas, a very important territorial and administrative unit, are used here as geographic and socioeconomic areas. Comunas are very heterogeneous, ranging from isolated and unpopulated areas to metropolitan areas with over 400,000 individuals each. The socioeconomic comuna-level data is estimated using data from the Encuesta de Caracterización Socioeconómica Nacional (CASEN). CASEN is by far the largest and most influential socioeconomic survey in Chile, and it uses comunas as a second-level sampling unit. There are roughly 350 comunas in Chile, around two thirds of them included in the CASEN sample. Two groups are defined in this study, by sorting the average household per-capita-wage in each comuna and then forming two groups of comunas: a high and a low relative SES group. Comunas in between the two groups are not included in the analysis. Each year from 1990 to 2011 that CASEN is conducted, the grouping is carried out so that the percentage of females aged 15-17 in each group is as close as possible to 30% (it ranges between 28% and 31%). The 30% percent rule is chosen as an equilibrium between the size of each group and the size of the difference in AFER. AFER by SES is defined then as single-year live births over mid-year population among 15-17 year old females. Results are similar if a 15-19 age range is used instead. Deliveries by SES are taken from Chile's official micro data (missing data is estimated to be under 5% since 1990 and about 1% since 2004), which includes the comuna where the mother lives. Total population by SES is taken from official comuna-level estimations, and the percentage of the population who is a 15-17 year old female is taken from the CASEN survey.

Regarding ASEX, the data used in this study comes from a completely different source. The Chilean National Youth Survey ( Encuesta Nacional de la Juventud– ENJuv) is the primary survey for youth issues in Chile, with seven rounds conducted from 1994 to 2012. ENJuv had a sampling error of 1.1% and 3% in urban and rural areas, respectively, in 2012. The ASEX in this study is operationalized by the “proportion that has ever had sex” among 15-17 year old adolescents, as the survey collected data every third year starting in 1994. Due to sample sizes, male and females are pulled together. Comunas on ENJuv 1997 and 2000 are blind-coded, therefore a new exercise of ranking and grouping comunas is required. By choosing among the few household-level economic variables in ENJuv, relative SES is obtained by hierarchically sorting comunas by their “proportion of interviewed population that report having a household head with a professional occupation (i.e. managerial-level employees or at least four years of university education)”. Another option was the “socioeconomic level” variable, as included in ENjuv, but this variable changes substantially from round to round and is not clearly defined in the methodological documents of the survey. Results, nevertheless, are not very sensitive to these options, even if individual-based rather than geographical-based SES are used. The grouping is done so the percentages of the population living in either the higher or lower SES are as close to 30% as possible.

Results

The AFER goes from 0.037 to 0.043 births per person-year, between 1990 and 1997 (Table 1), and then goes down to 0.037 in 2011. The ratio of AFER by SES, on the other hand, decreased from 1.45 in 1990 to 1.10 in 2011 (Figure 1). In other words, in 1990 and 2011 the adolescent fertility rate appears to be around 45% and 10% higher, respectively, amongst the 30% of the population that lives in lower SES areas, as compared to the 30% that lives in higher ones. Moreover, the absolute gap between them also dropped from nearly 0.014 births per person-year in 1990 to 0.004 in 2011.

ASEX rises from 20% to 39% between 1997 and 2012 (Table 1). The ratio of ASEX by relative SES among adolescents aged 15-17 dropped from 1.24 to 1.01 between 1997 and 2012 (Figure 1). The absolute gap in sexual initiation also dropped, from 6 percentage points in 1997 to around zero in 2012.

At the same time, the relative household per-capita wage between the highest and lowest SES areas remained at 2.5 (approximately) from 1990 in 2011. Therefore it is probably not a mere closing gap in SES itself that is explaining the closing SES gap in AFER.

Conclusions

The size of some dimensions of socioeconomic inequalities in adolescent fertility and sexual initiation have decreased between 1990 and 2011. This exercise shows that even in Chile, the most unequal country among the Organisation for Economic Co-operation and Development (OECD) ⁷7. Gurría A. Going for growth: reforming for a strong and balanced recovery. Paris: Organisation for Economic Co-operation and Development; 2013., there are clear improvements in some dimensions of health inequality.

This evidence is consistent with a study by Comisión Económica para América Latina y el Caribe (CEPAL) ⁸8. Comisión Económica para América Latina y el Caribe. Panorama Social de América Latina 2005. Santiago de Chile: Comisión Económica para América Latina y el Caribe; 2005., where the comparative low demographic inequality (including adolescent fertility) in Chile is emphasized, among a group of seven Latin-American countries. Is also consistent with evidence regarding a closing gap in infant mortality in Chile (CEPAL ⁹9. Comisión Económica para América Latina y el Caribe. Mortalidad en la niñez. Santiago de Chile: Comisión Económica para América Latina y el Caribe; 2011. compares urban and rural areas; Castro ¹⁰10. Castro R. Indicios de igualdad demográfica en Chile: el caso de la mortalidad infantil. Papeles de Población 2015; 21:189-201. compares geographic relative SES).

Two general comments that are worth mentioning. First, greater attention of reproductive health programs to adolescents of higher relative socioeconomic level is necessary. Further exploration of the data shows that in higher socioeconomic areas ASEX decreased and AFER remained constant. Second, amidst a lack of standard population-level measures ¹¹11. Buhi ER, Goodson P. Predictors of adolescent sexual behavior and intention: a theory-guided systematic review. J Adolesc Health 2007; 40:4-21., opposite socioeconomic groups of equal relative size provides an opportunity for internationally and inter-temporally comparable studies.

Within the limitation of this study, it is important to highlight that selection effects might distort the results if, for example, adolescents that are less inclined to early reproductive lives are moving out of poverty across time. It is also important to mention that measuring SES by residential-area leaves out individual-level dimensions of SES that are relevant for the study of socioeconomic health inequalities.

References

Publication Dates

History