FACTORS ASSOCIATED WITH THE LACK OF USE OF CONTRACEPTIVE METHODS IN FEMALE ADOLESCENT IN PERU ENDES 2018-2020

95%. Resultados: 28,4% utilizaban MC, hubo asociación signi�cativa para las variables no tener pareja (Rpa: 2,49 IC: 1,99-3,11) y no tener hijos (Rpa: 2,73 IC: 2,22-3,35). Conclusiones: Los factores asociados a no uso de MC en mujeres adolescentes de 15 a 19 años fueron no tener pareja y no tener hijos, con mayor prevalencia de no uso de MC. ABSTRACT Introduction: In Peru, adolescents tend to use contraceptive methods (CM) less frequently, constituting a subgroup at risk of unwanted pregnancies and sexually transmitted diseases. Methods: Quantitative, observational, analytical-cross-sectional study corresponding to a secondary analysis of data from women aged 15 to 19 from the ENDES 2018-2020. To quantify the association, Poisson regression was used for robust variances, calculating the Prevalence Ratio (PR) in its crude (PRc) and adjusted (PRa) forms, with a 95% con�dence interval. Results: 28.4% did not use MC, there was a signi�cant association for the variables without a partner (Rpa: 2.49 CI: 1.99-3.11) and zero children (Rpa: 2.73 CI: 2.22-3 .35). Conclusions: The factors associated with the lack of CM in female adolescents were without a partner and zero children, with a higher prevalence of the lack of CM.

In 2020, the WHO estimated that every year 21 million adolescents aged 15 to 19 living in developing countries become pregnant. Some teens plan and want to get pregnant, but in many cases they don't; approximately ten million pregnancies of adolescents in this age range living in developing countries are (1) unwanted .
The promotion of the use of CM has been shown to be effective in reducing early and unwanted pregnancies, maternal and neonatal morbidity and mortality, and cases of unsafe abortions. In addition, male and female condoms offer double protection against unwanted (2) pregnancy and sexually transmitted diseases .
Regarding women, who are considered a vulnerable group to infection of sexually transmitted diseases, there are different studies that report the prevalence of frequency of CM use, as is the case of the study carried (3) out in Congo in 2020 by Casey et al. who reports a prevalence of 51.7% in the age group 15-19 years, compared to those 20-24 years of 36.5%. (4) In Uganda in 2020, Otim carried out a study in the Central, East, Northeast and West regions of the country, in which it reports that there are high percentages in these regions of non-use of CM in women who do not have children compared to those who do (73.9; 79.9%; 91.0% and 86.7% respectively); it also observed that in the Northeast region, adolescents who had no education had a higher percentage (65.8%) of not using CM. (5) In the case of Latin America, Gomez et al. in 2019 reported low prevalence's in Haiti (31.3%) and Bolivia (34.6%) in contrast to countries such as Brazil, Colombia, Costa Rica, Cuba and Paraguay with more than 70% of (6) CM use. In Colombia in 2018, Angarita et al. found that women who have not been victims of intimate partner violence (OR 0.75, 95%CI 0.69-0.82), physical (OR 0.87, 95%CI 0.80-0.94) and psychological (or 0.75, 95% CI 0.69-0.82) are more likely to use some CM than women who were. In Peru, a study based on the demographic and family health survey (ENDES) 2017 reports a prevalence of 52.3% for non-use of CM, and women living in rural areas have a prevalence ratio of 1.8 (95%CI: 1.5-2.1) for non-use of CM, compared to those (7) living in urban areas ; However, it does not specify the types of CM or how the dependent variable was de ned.
An observational, analytical and cross-sectional study was carried out using the ENDES 2018-2020 database in (8) Peru . The study had a two-stage, probabilistic sampling of balanced, strati ed and independent type, at the departmental level and by urban and rural area.
As mentioned, all these factors can in uence a lower use of CM, and being adolescents a risk group of unwanted (2) pregnancies with unattended contraceptive needs , the objective of this study was to identify the factors associated with the non-use of CM in adolescent women of the ENDES 2018-2020.

Variables
The variables were dichotomized and evaluated using (8) the ENDES 2018-2020 which we did not modify. They were dichotomized in order to better represent the state of the participant and facilitate the interpretation (9) of the coefficients of the statistical model .
The dependent variable was no use of CM (question V313), dichotomized as use of CM (folkloric method, traditional method and modern method) and no use of CM (no method). The covariates were categorized into sociodemographic, reproductive and violence. Among (10) the sociodemographic factors were considered age (question V012), average adolescence (14 to 17 years) and late adolescence (18 to 21 years) were considered; (7) area of residence , (question V025), rural and urban; (11) wealth index (question V190), poor (very poor, poor) and non-poor (medium, rich and very rich); higher (12) education (question V149), higher (non-university higher, university higher and postgraduate) and nonhigher (initial/pre-school, primary and secondary); and (13) the couple (question V501), having a partner (cohabiting and married) and not having a partner (single).

METHODS
The study population consisted of 117 476 women and the sample after applying the inclusion criteria (participants aged 15 to 19 years with complete data for the dependent variable) was 1 706 adolescents aged 15 to 19 years, which was collected based on the sample design of the ENDES 2018-2020. In the bivariate analysis, prevalence ratios were calculated, crude (ORc) and adjusted (ORa) as a measure of association, both accompanied by their respective 95% con dence intervals. A value of p<0.05 was considered. The ORc and ORa were calculated using a Poisson regression model with robust variances. All estimates were made considering the expansion factor of the sample design. A power of 99% was obtained for risk factors identi ed as signi cant, with the overall mean power of all factors being 56%.

Pág. 347
(4) For reproductive factors, the number of children (question V201), not having children and having one to more children (if you have live children); age at rst ( 4 ) sexual intercourse (question V5125), early adolescence (11 to 13 years) and middle adolescence (4 (14) to 17 years); and sexual partners (question V836), one sexual partner (if you have had one sexual partner in your life) and two more sexual partners (if you have had several sexual partners in your life) were taken into account.
For this study we used a secondary data source from the ENDES 2018-2020. Statistical processing and analysis (15) were performed using Stata software . The univariate analysis was performed from the calculation of frequency and percentage.

RESULTS
Approval was obtained from the ethics committee of the Institute of Biomedical Sciences of the Faculty of Human Medicine of Ricardo Palma University; however, this was a study based on secondary sources.

Ethical aspects
In the multivariate analysis, it was found that not having a partner had a prevalence ratio of 2.49 (95%CI: 1.99-3.11) of not using CM with respect to having a partner; and not having children had an ORa of 2.73 (95%CI: 2.22-3.35) of not using CM with respect to having one or more children. Table 3 In the bivariate analysis, the variables not having a partner and not having children were signi cant (p <0.001). Table 2 Of the 1,706 women aged 15 to 19 years, 28.4% did not use CM. It was observed that the adolescents were mostly in late adolescence (72.5%), lived in urban areas (53.1%), were poor (77.8%), had no higher education (93.7%), had a partner (88.0%), had one or more children (82.9%), had their rst sexual relationship in late adolescence (84.1%), had a sexual partner (69.2%), More than half reported domestic violence (54.4%) and psychological violence (50.1%), while the majority did not report economic, physical or sexual violence 85.4%, 80.7% and 96.9% respectively.

DISCUSSION
The research found that 28.4% of adolescents aged 15 to 19 years did not use CM, a similar result is observed in (18) the study of Guerrero where adolescents of the same age did not use contraceptives in 22.6%; However, Mejía in his study in four universities in Latin America reports that only 7% do not use any CM.
These ndings may be explained by the fact that adolescents without a partner may not consider the use of CM necessary, since they would not have sex; or some of them, when they have a formal partner with an apparent stability of the relationship, abandon the use of CM, putting their sexual health and the possibility of having an unwanted pregnancy at risk.
A statistically signi cant association was found for the variable couple (p <0.001). Not having a partner had 2.49 times higher prevalence of not using CM than having a partner; while having a partner had a higher percentage of use of CM (76.1%), however, Valladares et (13) al.
report association between marital status and non-use of CM, where those who never married had 63.2% non-use of CM.
It was also found that 71.6% of adolescents used CM; (3) However, Casey et al. report that only half of adolescents aged 15 to 19 years use CM (51.7%). Gomez (5) et al. found that the use of reversible long-acting contraceptives among adolescent women aged 15 to 17 years is 1.1% and in women aged 18 to 19 years 2.0%.
The variables age, area of residence, wealth index, higher education, age at rst sexual intercourse, sexual partners, violence (domestic, psychological, economic, physical and sexual) showed no statistical signi cance; This could be explained because the statistical signi cance between two variables depends on two fundamental components. The rst is the magnitude of the difference to be tested, that is, the greater the difference between two variables, the easier it is to demonstrate that the difference is signi cant. On the other hand, if the difference is small, it is difficult to detect differences between the variables. The second component is the sample size, because the larger the sample size, the easier it is to detect differences between variables. Small differences can be detected with large sample sizes and large differences between (20) variables need far fewer individuals to be studied .
In this investigation, a signi cant association was found between the variable number of children and not using CM. Not having children had a prevalence of 2.73 times of not using CM compared to having one or more children; but having one or more children had the highest percentage of CM use (77.5%), (21.3%); Salazar found signi cant association for area of residence, 61% of women who did not use CM lived in (14) rural areas; Noll et al. reported signi cance for the variable number of sexual partners, and also observed that adolescents with four or more sexual partners had a higher frequency of not using CM, such as condom (15) (43.5%); Finally, Quispe et al. found an association between women who do not use CM and domestic violence (16.4%).
Within the limitations, the frequency of use of each CM An association was found between non-use of CM and the reproductive factor not having children (ORa: 2.73 CI: 2.22-3.35) and the sociodemographic factor not having a partner (ORa: 2.49 CI: 1.99-3.11). No association was found with the violence factor. CONCLUSION and the evaluation of its correct use were not determined; and others speci c to analytical crosssectional studies, such as the fact that causality cannot be established; only association between the variables studied.

Authorship contributions:
The authors, YDR and CLM participated in the genesis of the idea and design of the draft; YDR, CLM and LRA in data interpretation, analysis of results, and preparation of the manuscript of the present research work; plus YDR on data collection.

Con icts of interest:
There was no con ict of interest between the authors during the process of conducting this research.