Effects of Preconception Care and Periconception Interventions on Maternal Nutritional Status and Birth Outcomes in Low- and Middle-Income Countries: A Systematic Review

Pregnancy in adolescence and malnutrition are common challenges in low- and middle-income countries (LMICs), and are associated with many complications and comorbidities. The preconception period is an ideal period for intervention as a preventative tactic for teenage pregnancy, and to increase micronutrient supplementation prior to conception. Over twenty databases and websites were searched and 45 randomized controlled trials (RCTs) or quasi-experimental interventions with intent to delay the age at first pregnancy (n = 26), to optimize inter-pregnancy intervals (n = 4), and supplementation of folic acid (n = 5) or a combination of iron and folic acid (n = 10) during the periconception period were included. The review found that educational interventions to delay the age at first pregnancy and optimizing inter-pregnancy intervals significantly improved the uptake of contraception use (RR = 1.71, 95% CI = 1.42–2.05; two studies, n = 911; I2 = 0%) and (RR = 2.25, 95% CI = 1.29–3.93; one study, n = 338), respectively. For periconceptional folic acid supplementation, the incidence of neural tube defects were reduced (RR = 0.53; 95% CI = 0.41–0.77; two studies, n = 248,056; I2 = 0%), and iron-folic acid supplementation improved the rates of anemia (RR = 0.66, 95% CI = 0.53–0.81; six studies; n = 3430, I2 = 88%), particularly when supplemented weekly and in a school setting. Notwithstanding the findings, more robust RCTs are required from LMICs to further support the evidence.


Introduction
Awareness of the possibility for influencing preconception health to maximize benefits for mothers and babies started upon the release of a ground-breaking report from the Centre for Disease Control [1]. Following this, the World Health Organization (WHO) gathered experts to discuss how preconception care could potentially have a positive impact on maternal and child health outcomes, where the experts overwhelmingly agreed with this proposal [2]. Subsequently there has been growth in the awareness of how important the preconception period is, as well as initiatives to increase the awareness and promotion of reproductive health from adolescence and beyond. our review is thorough, we included non-randomized studies as contextual and supplementary evidence for the included RCTs [18]. The benefit of including these studies in systematic reviews includes the ability to demonstrate whether an intervention is applicable and effective in dissimilar populations, and allows us to explore possible interaction effects and to demonstrate long-term outcomes.
This review aimed to synthesize the current evidence on the effectiveness of preconception care interventions relating to the delayed age at first pregnancy; optimizing inter-pregnancy intervals; periconception folic acid; and periconception iron-folic acid supplementation on maternal, pregnancy, birth, and child outcomes using an approach of systematically reviewing primary studies, and meticulously appraising existing programs. This strategy enabled a comprehensive in-depth assessment of the effectiveness of these interventions to ameliorate the aforementioned outcomes. The evidence generated will be crucial to apprise both policy and programmatic decision-making in LMICs.

Materials and Methods
The protocol for this systematic review has been published with The Campbell Collaboration [19].

Objectives
The general objective was to assess how effective the following pre-and periconception interventions were at improving maternal nutrition, birth, and neonatal outcomes in LMICs when compared with no/standard intervention in terms of (1) interventions to delay the age at first pregnancy, (2) interventions to optimize inter-pregnancy intervals, (3) periconception folic acid supplementation, and (4) periconception iron folic acid supplementation.

Types of Studies
Primary studies, including large-scale program evaluations, were used to assess the efficacy and/or effectiveness of interventions using RCTs or quasi-experimental designs (natural experiments, controlled before-after (CBA) studies, regression discontinuity designs, interrupted time series (ITS)). Pre-post studies that lacked a control group were excluded. The language was restricted to English.

Types of Participants
Women of reproductive age (i.e., 10 to 49 years) were our target population. This included adolescent girls, regardless of health status, living in LMICs. The 2018 World Bank list of country economies [20] was used to classify country incomes; we consulted this document to ensure that studies found to be conducted in countries that were part of LMICs prior to 2018 were considered as such. While interventions were aimed at non-pregnant women, outcome measurements were for pregnant women, as well as for their children. For optimizing birth intervals, we considered interventions given to participants during pregnancy to optimize birth intervals for the next pregnancy.

Types of Intervention
The following interventions targeting women of reproductive age (10-49 years), including adolescent girls (10-19 years), during the pre-and periconception period in LMICs were included: • Interventions to delay the age at first pregnancy, such as curriculum-based sex education, abstinence alone programs, interactive computer-based interventions, etc.
o Educational interventions and contraceptive promotion given to adolescents and young women at the community, school, or household level by parents, colleagues, teachers, health workers, or social workers.
• Interventions to optimize inter-pregnancy intervals, such as introducing family planning methods, abstinence alone programs, etc.
o Educational interventions and contraceptive promotion given to mothers of reproductive age at the community, school, or household level by parents, colleagues, teachers, health workers, or social workers. • Periconception folic acid supplementation.
o Any folic acid supplementation given to either pubescent or menstruating women prior to conception that continued until the first trimester of pregnancy. • Periconception iron folic acid.
o Any iron folic acid supplementation given to either pubescent or menstruating women prior to conception and/or continued until the first trimester of pregnancy. Interventions were compared to either no intervention, standard of care (based on study setting and what was applicable therein), or placebo. Interventions where folic acid and iron-folic acid were only used during pregnancy were not included. We excluded multiple micronutrient powders for point-of-use fortification of foods; fortification of staple foods, water, condiments, or seasonings with folic acid or iron; and other micronutrient-or folic-acid-containing oral contraceptives. Fortification programs were excluded as, by their nature, they are administered universally and do not allow the exact period of starting and stopping of intake to be known and therefore generate evidence for recommendation. We also excluded oral contraceptives that contain folic acid as they warrant a separate review altogether.

Type of Outcome Measures
Primary outcomes: • Maternal: unintended pregnancy, anemia, and iron deficiency anemia.
• Neonatal: neural tube defects, still birth, perinatal mortality, neonatal mortality, and low birth weight. Secondary outcomes: • Maternal: reported changes in knowledge and attitudes about the risk of unintended pregnancies, initiation of sexual intercourse, use of birth control methods, serum folate, adverse effects, adherence to folic acid or iron folic acid supplementation, abortion or miscarriage, and maternal mortality. • Neonatal: preterm birth, small-for-gestational age, other congenital anomalies, and special care admission due to any reason. Studies were not included if they did not report outcomes of interests. If the outcomes were measured at any time points during pregnancy or the postpartum period, they were considered. This data was pooled based on a reasonable time point reported by most of the studies. The outcomes and their corresponding interventions are displayed in Figure 1.

Duration of Follow-Up
For interventions focusing on delaying the age at first pregnancy, we considered studies where the intervention was given in preconception. For interventions aiming to optimize inter-pregnancy intervals, we considered studies where interventions to optimize these intervals were given at any point in the course of the previous pregnancy, as well as interventions implemented after the birth of the last child. For folic acid supplementation, we considered studies where folic acid was supplemented during the pre-and periconception period. Lastly, for iron-folic acid supplementation, we considered studies where iron-folic acid was supplemented during the pre-and/or periconception period.

Literature Search
The search was performed on May 31, 2019 using the following electronic databases: CABI's Global Health, CINAHL, Cochrane Controlled Trials Register (CENTRAL), Dissertation Abstracts International, EMBASE, Epistemonikos, ERIC, HMIC (Health Management Information Consortium), MEDLINE, Popline, PsycINFO, Scopus, Social Science Index from Web of Science, Sociofiles, WHO's Global Health Library, WHO Reproductive Health Library, and the WHO nutrition databases (http://www.who.int/nutrition/databases/en/). Web sites of selected development agencies or research firms (for example, JOLIS, IDEAS, IFPRI, NBER, USAID, World Bank) and Google Scholar (Appendix A) were searched.
Moreover, we checked the reference lists of all included studies and systematic reviews for further references. We tried to contact pertinent organizations and field experts to locate relevant unpublished or ongoing studies. We also scanned the references of included articles, applicable reviews, and annotated bibliographies for eligible studies.

Data Collection and Analysis
Data collection and analysis was carried out in conformity with the Cochrane Handbook for Systematic Reviews of Interventions [21].
Two review authors (Z.S.L. and S.K.) extracted data separately and an additional review author (J.K.D.) was in place to check the data to ensure reliability and to resolve any conflict. Data for the study characteristics, including population details, setting, socio-demographic characteristics, interventions, comparators, outcomes, and study design, was extracted in duplicate. Primary study data was inspected for accuracy. Any disagreements that occurred were rectified through discourse with a third reviewer.
Once all the references were retrieved, two review authors (Z.S.L. and S.K.) independently screened their titles and abstracts. When at least one review author considered a citation possibly relevant, its full-text study report was retrieved. Two review authors (Z.S.L. and S.K.) independently screened the full text articles and identified studies for inclusion. They used a "characteristics of excluded studies" table to record the reasoning behind the exclusion of omitted studies. We resolved any disagreement through discussion or consultation with a third review author, if needed. We discovered and excluded duplicates, and to ensure that each study itself is the unit of interest, as opposed to each report, we collated various reports from the same study. We documented the entire detailed selection process to complete a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram [22].
Using a data extraction sheet, two review authors (W.T. and Y.J.) independently extracted data. We resolved any disagreement through discussion or in consultation with a third review author (Z.S.L.) if needed. We used a piloted data collection form for study characteristics and outcome data. In instances where any information was either unclear or missing, we contacted the authors of the original papers for additional details. A data extraction form was used to record data, which summarized principal characteristics of the review/studies including: • Methods: study design and study duration.
• Details of study participants (age, socioeconomic status, parity): numbers randomized, and inclusion and exclusion criteria. • Interventions: content, duration and timing of intervention, and comparisons. • Outcomes and time point.

Assessment of Risk of Bias in Included Studies
The risk of bias for each of the included studies was determined independently by two review authors (W.T. and Y.J.). We resolved any disagreements either by discussion or by consulting a third review author.
We used the Cochrane Collaboration Risk of Bias tool for all RCTs, encompassing cluster RCTs [23]. We assessed the risk of bias according to the subsequent domains. Each criterion was rated as high, low, or unclear risk.
• blinding of participants and personnel.
• blinding of outcome assessment for each outcome.
• other bias, such as validity of outcome measure and baseline comparability. For CBA and ITS, we used Effective Practice and Organization of Care (EPOC) methods [24]. Each criterion was rated as high, low, or unclear risk.
• random sequence generation • allocation concealment • baseline outcome measurements • baseline characteristics • incomplete outcome • knowledge of the allocated interventions adequately prevented during the study • protection against contamination • selective outcome reporting • other risks of bias.

Measures of Treatment Effect
We uploaded the outcome information for each included study into RevMan's data tables to ascertain the treatment effects [25]. For dichotomous outcomes, we used the risk ratio (RR). For continuous outcomes reported on the same scale, we used the mean difference (MD), and for continuous outcomes that reported the same outcomes, albeit on different scales in different studies, we used the standardized mean difference (SMD). We expressed uncertainty with 95% confidence intervals (CIs) for all effect estimates. When means and standard deviations were not given, we used other obtainable data, including confidence intervals, t-values, p-values, and assigned appropriate methods described in the Cochrane Handbook for Systematic Reviews of Interventions (21) to calculate the means and standard deviations. Where other obtainable reported data was not sufficient to calculate the standard deviations, we contacted the relevant study authors. When we could not enter the results in either way, we recorded and displayed them in the Supplementary Tables. We also considered the likelihood and implications of skewed data during analyses of continuous outcomes given that due to small sample sizes, they can provide delusive results. We also examined any relevant retraction statements and errata for information. Outcomes with multiple groups were analyzed appropriately to avoid overcounting participants by adding them to different sub-groups within the same plot. In these cases, we did not report the overall pooled estimate but instead the reported sub-group pooled estimate.

Unit of Analysis Issues
We performed a separate meta-analysis for each topic mentioned as a separate objective. We assessed the effectiveness of each intervention as a sub-group. When trials used clustered randomization, we predicted that the investigators would have presented their results after appropriately controlling for the effects of clustering (for example, variance-inflated standard errors and hierarchical linear models). In instances where it was uncertain whether a cluster RCT had appropriately accounted for clustering, we contacted the study investigators for further information. Where appropriate controls for clustering were not made use of, we sought an estimate of the intraclass correlation coefficient (ICC). In the unlikely event that authors did not reply to our request for ICC estimates, we took the ICC reported in similar studies with a similar context. Afterward, the effect sizes and standard errors were meta-analyzed in RevMan [21]. These were merged with estimates from individual level trials.

Dealing with Missing Data
Where possible, we contacted trial authors to confirm key study characteristics and secure absent numerical outcome data. When numerical outcome data, such as standard deviations (SDs), were not provided, and we were unable to procure these from the study authors, we calculated them from other available statistics, such as p-values, or using the methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions [21].

Assessment of Heterogeneity
Statistical heterogeneity was appraised using τ 2 , I 2 , and significance of the χ 2 test. We assessed heterogeneity visually and with forest plots. Based on prior theory and clinical knowledge, we expected clinical and methodological heterogeneity in effect sizes in this literature. Therefore, we used subgroup analysis to endeavor to explain any observed statistical heterogeneity.

Assessment of Reporting Biases
When sufficient studies were found, we drew funnel plots to explore any possible relationship between the effect size and study precision. Ten studies are usually considered sufficient to draw a funnel plot.

Data Synthesis
Statistical analysis was carried out using RevMan software [25]. A matrix was prepared for each intervention consisting of all the studies that outlined the dissimilarities in the studies at various levels (intervention, duration, timing, etc.), and was used to decide how best to pool the data. Random effects meta-analyses were used due to the diverse study circumstances, participants, and interventions. Findings for each comparison were descriptively summarized using contextual factors, including study setting, timings and duration of intervention, and people delivering interventions, to assess their impact on each intervention's effectiveness.

Assessment of Quality of Evidence
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) technique was applied to individual outcomes to assess the quality of the evidence [26], which involved deliberations regarding the within-study risk of bias, directness of evidence, heterogeneity, precision of effect estimates, and risk of publication bias. "High," "moderate," "low," or "very low" classifications were applied to the quality of the body of evidence for each key outcome. Nonrandomized studies were rated as "low" quality at first. When there were no serious flaws in methodology, we upgraded the evidence for studies with a large magnitude of effect, presence of dose-response relationships, and the effect of plausible residual confounding. GRADE was performed for all primary outcomes.

Subgroup Analysis and Investigation of Heterogeneity
When there were ample studies included in the outcomes, we conducted the subgroup analyses on the following domains: • Setting (home, facility based, community level, school, work).
• Type of intervention (school-based education, abstinence-only program, contraceptive promotion, etc.). The subgroup analyses were carried out using Review Manager 5.3 with a test for interaction.

Sensitivity Analysis
We used sensitivity analyses to assess the potential biasing effects of using the interclass correlation coefficients.
Sensitivity analyses were planned to factor the impact of the following; however, due to the limited number of studies in each outcome, we were unable to perform these.
• Imputed inter-correlation coefficients that were derived in different ways.

Study Selection
We identified a total of 8523 papers from the different search engines. After removing duplicates, 7123 abstracts were reviewed. Of those, 323 full texts were reviewed, and finally 45 studies were included [ Figure 2). Of these, 26 were on delay in the age of pregnancy, 4 on optimizing interpregnancy birth intervals, 5 on the supplementation of folic acid, and 10 on the supplementation of iron-folic acid. The characteristics of the included studies are summarized in Table 1    *The risk in the intervention group (and its 95% confidence interval) was based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
1. There is a high risk of attrition bias due to greater than 20% patients being lost before follow up from both intervention and control arms.

High risk of selection bias.
Optimizing inter-pregnancy interval-Education + provision of contraception + involvement of male partner vs. education alone *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
1. Heterogeneity not applicable as there was only one study under this comparison.
2. Total number of events was less than 300.

Periconceptional folic acid supplementation compared to placebo
Folic acid compared to placebo for periconceptional women (1 to 1) *The risk in the intervention group (and its 95% confidence interval) was based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect. Footnotes: 1. Two studies (Berry et al. 1999 [55]) (Vergel et al. 1990 [58]) did not have random sequence generation and allocation concealment.
2. Number of events was less than 300.

Periconceptional iron folic acid supplementation compared to placebo
Iron folic acid compared to placebo for periconceptional women *The risk in the intervention group (and its 95% confidence interval) was based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
1. Some studies use multiple micronutrients in the intervention arm.
2. Multiple studies with a large weightage are at high risk for bias.
4. Total number of events was less than 300. 5. One study used vitamin C along with iron-folic acid in the intervention arm.
8. One study was at high risk of bias. 9. Study was at risk of performance and reporting bias. 10. Study was at risk of other biases. 11. It was mostly unclear if study was at risk of bias. 13. Heterogeneity was 95%.
14. One study was at risk of attrition bias.

Description of Studies
A composite of 26 trials related to delaying pregnancy were included. All the trials focused on maternal outcomes, namely unintended pregnancy, reported changes in knowledge and attitudes about the risk of unintended pregnancies, initiation of sexual intercourse, use of birth control methods, and abortion.
Most trials investigated educational interventions, where 16 trials had education alone as the intervention [28][29][30][31]34,35,[37][38][39]41,43,45,47,49,50,52]. In six trials, education was combined with other strategies: provision of contraceptives [42,44,51]; peer referrals to health care providers along with the training of health care providers [46]; training of health workers and peer condom marketing [48]; referrals, family members' education, and improvement of contraceptive services [47]; skills training, referrals to micro savings and credit groups, and health care provider training [40]; and youth partnership group development and education subsidies [32]. Cash transfers were given to participants in the intervention group in two trials [27,36]. They were conditional in one trial [27], and unconditional in the other [36]. In one trial, there were three intervention arms: teacher training only, relative risk education only, and teacher training and relative risk education [33].
Two trials reported on the use of a modern birth control method, with a total of 2466 women [47,50]. One trial [50] reported the current use of modern birth control methods while the other reported whether they were ever used [47]. Education on contraception did not have a significant impact on the usage of modern methods of contraception when compared with no education (RR = 2.12, 95% CI = 0.64-7.07; two studies, n = 1028, random-effect, χ² p < 0.0001, I² = 94%). When this outcome was subdivided based on setting and trial type, there was no change in the risk ratio as both of the included trials had the same setting (community) and type (quasi-experimental). Eight trials [34,38,44,[46][47][48]50,51] reported on the use of condoms. Education on contraception did not have a significant impact on the current usage of condoms when compared with no education (RR = 0.93, 95% CI = 0.81-1.06, eight studies, n = 1175, random-effect, χ² p = 0.56; I 2 = 0%), but it did have a significant impact on whether condoms where ever used when compared to no education (RR = 1.54, 95% CI = 1.08-2.20; six studies, n = 1604, random-effect, χ² p = 0.004; I 2 = 71%).
One trial reported on the use of a traditional birth control method, with a total of 2061 women [47]. Education on contraception did not have a significant impact on the usage of traditional methods of contraception when compared with no education (RR = 1.70, 95% CI = 0.94-3.07; one study; n = 623; random-effect).
One trial reported on the use of contraceptive pills [34], and demonstrated that education had a significant impact on the use of pills when compared with no education (RR = 1.34, 95% CI = 0.89-2.01; one study, n = 288; random-effect]. The same trial also reported on the use of injectable contraception methods [34], and this intervention had a significant impact on the use of depot/injectable methods when compared with no education (RR = 1.58, 95% CI = 1.26-1.98; one study, n = 288; random-effect).
Initiation of sexual activity was described by four trials [31,45,48,52]. It was divided into three subgroups based on the time of follow up [three, six, twelve months and three years). Education on reducing sexual risks did not have a significant impact on sexual debut when compared with no education at the three-month follow-up (RR = 0.43, 95% CI = 0.04-4.51; one study, n = 56; randomeffect), the six-month follow-up (RR = 1.02, 95% CI = 0.57-1.83; two studies, n = 1443, random-effect; χ² p = 0.23, I 2 = 29%), or at the three-year follow up [RR = 0.95, 95% CI = 0.79-1.14; two studies, n = 1153, random-effect; χ² p = 0.36, I 2 = 0%]. There was a significant decrease in the age of sexual initiation at the twelve-month follow-up (RR = 0.70, 95% CI = 0.49-0.99; one study, n = 1387, random-effect). For the three-month follow-up outcome, there was one trial [52], which was an RCT that took place in a school setting. For the six-month follow-up outcome, there were two trials that were both RCTs [45,52], and one was a cluster RCT (cRCT) [45]. One trial took place in a school setting [52] and one took place in combined school and clinic settings [45]. For the 12-month follow-up outcome, there was one trial, which was a cRCT and took place in combined school and clinic settings [45]. For the three-year follow-up outcome, there were two trials [31,48], one was a cRCT [48], and one was a quasi RCT [31].
In Martiniuk 2003, the outcome was reported as a difference in knowledge and attitude, which was attributable to the intervention in change scores, post-test minus pre-test, between the two trial arms. For females' knowledge, the crude score was 2.11 with a 95% CI of 0.23-3.99. For females' attitude, the crude score was 0.05 with a 95% CI of −2.60 to 2.70 [43]. In Pandey [47], this was reported as the percentage of young women who had reported various risks to the mother or child associated with early child bearing. Almost 99.6% of young women given the intervention were aware of risks that a girl can encounter if she gives birth during adolescence at the age of 15-16 years, 38.8% were aware that underdeveloped reproductive organs lead to prolonged or obstructed labor, 50.2% were aware of the elevated likelihood of complications in pregnancy and labor/delivery, 43.0% were aware of the increased risk of maternal mortality, 6.1% were aware of miscarriage/still birth, 86.6% were aware of ill health of the mother, 1.2% were aware of anemia in women, 98.4% knew about the risks that an adolescent mother's child may encounter, 10.5% were aware of the increased possibility of an underdeveloped child, 3.0% were aware of the increased possibility of an underdeveloped child, 4.2% were aware of the possibility of a premature birth/baby, 90.6% were aware of the risk of a weak child, 39.5% were aware of the risk of infant death, 32.0% were aware of the risk of a low birth weight (LBW) baby, 16.0% were aware of the risk of a disabled child, and 92.5% of young unmarried women had the intent to practice contraception in order to delay their first pregnancy.
Two trials also assessed the impact of conditional cash transfers. Handa 2015 reported that cash transfers to adolescent vulnerable girls reduced the possibility of pregnancy by five percentage points, while the likelihood of early marriage was not significantly impacted [36]. Whereas, Baird 2010 reported that average condom use for school girls was -0.136 (0.075) and 0.031 (0.201) post intervention [27].

Description of Studies
Four trials were included in this review relating to optimizing inter-pregnancy intervals, which involved a total of 15,718 participants. There were two quasi experimental natural experiments [30,47,53] and one randomized controlled trial [54]. The primary outcome that was reported on by the included trials was unintended pregnancy [54], other outcomes comprised of changes in knowledge and attitudes about the risk of unintended pregnancies [47], initiation of sexual intercourse [47], use of birth control methods [30,47,54], and abortion [54].
There were a total of four trials in Asia: one in Bangladesh [53], one in China [54], and two in India [30,47]. All the trials took place in community settings. The minimum population size was 2336 participants [54], and the maximum population size was 3980 participants [47].
Education about reproductive health and related issues, such as family planning, was the main aspect of the intervention in three trials [30,47,54,71]. There were also additional elements of behavior change communication in Daniel's trial [30], counseling, referrals, family members' education, and the improvement of contraceptive services in Pandey 2016 [47], as well as referrals, counseling, free provision of contraceptive materials, and involvement of the male partner in Zhu 2009 [54]. Postpartum family planning, along with maternal and newborn care for birth spacing, was the intervention in one trial [53].

Effects of Interventions
For the analysis comparing education versus no intervention, two trials were included [30,47]. In Daniel's trial [30], the change in knowledge and attitude was judged via responses to certain statements. The number of married women aged 15-24 years that agreed that early childbearing is harmful to a mother's health increased after the intervention, from 17% to 74% in the intervention group and 12% to 65% in the control group. The number of married women aged 15-24 that agreed that contraceptive use is safe to use and required to delay first birth increased from 38% to 80% in the intervention group and 36% to 72% in the control group. In Pandey's trial [47], 29.7% of the young married women (who had given birth at least once) in standalone areas reported the use of any modern method of contraception, compared to 18.9% in control areas. A total of 19.8% of the young married women (who had at least had one birth) in standalone areas reported the use of any modern spacing method, compared to 9.2% in control areas.
For use of contraception, analysis of the trials [30,47] demonstrated that education on contraception did not have a significant impact on improving the use of contraception when compared with no education (RR = 2.72, 95% CI = 0.88-8.40; two studies, n = 2385, random-effect, heterogeneity:; χ² P < 0.0001; I 2 = 94%), but education with the provision of contraceptives and involvement of male partner did have a significant impact (RR = 1.83, 95% CI = 1.26-2.66; one study, n = 338; random-effect). One trial [47], involving a total of 2061 women, reported the use of modern methods of contraception. Education on the use of modern methods of contraception when augmented with supplying contraceptives and involving the male partner had a significant impact on the use of modern methods of contraception when compared with no education (RR = 2.45, 95% CI = 1.42-4.24; one study, n = 338; random-effect). Subgroups according to trial setting and type were not made for this as both of the trials [30,47] under this comparison were quasi-experimental trials that took place in community settings.
Zhu [54] compared education involving the male partner with contraceptive provision with education only. Education with contraceptive provision did not have a significant impact on the risk of unintended pregnancies when compared to the less comprehensive package (RR = 0.32, 95% CI = 0.01-7.45; one study, n = 45; random-effect; moderate certainty of evidence using GRADE assessments). Regarding birth control methods, there was no significant impact. These included the use of any contraceptive method (RR = 1.05, 95% CI = 0.91-1.21; one study, n = 39; random-effect), and the use of condoms, oral contraceptives, intrauterine devices, and implants (RR = 1.08, 95% CI = 0.88-1.26; one study, n = 39; random-effect).

Description of Studies
Five trials were included with a focus on folic acid, with a total of 254,746 women. The included trials mainly focused on maternal outcomes and a subset divulged outcomes related to neonatal health. There were two randomized controlled trials (RCTs) [56,57] and three quasi-experimental natural experiments [55,58,59]. The primary outcome reported was a neural tube defect. The secondary outcome reported was miscarriage.
Of the five included trials, two took place in Asia: two in China [55,56]. One trial took place in North America, in Honduras [57], and two trials took place in South America: one in Cuba [58] and one in Brazil [59]. All participants were non-pregnant women with ages ranging from 16 to 49 years, and the sample sizes ranged from 140 [57] to 247,831 [55].

Effects of Interventions
Of the included trials, two [55,58] underwent a meta-analysis comparing periconceptional supplementation with folic acid versus a placebo. All these trials had daily supplementation and the dosage of folic acid was 0.4 mg [55] and 5 mg [58]. Pooled analysis found that periconceptional folic acid supplementation reduced the risk of neural tube defects (NTDs) compared to placebo by 47% (RR = 0.53, 95% CI = 0.41-0.67; three studies; n = 248,056; random-effect; heterogeneity: χ² p= 0.36; I 2 = 0%; very low certainty of evidence using GRADE assessment). However, the impact of periconceptional folic acid supplementation on NTDs was only significant for 0.4 mg of folic acid and non-significant for 5 mg of folic acid.
Miscarriage was reported in one trial [58,72]. Vergel [58] reported miscarriages as examined versus not examined. In the not-examined category, six of 124 of the unsupplemented participants and 1 of the 81 fully supplemented participants had a miscarriage. In the examined category, one of the 20 partially supplemented participants had a miscarriage (Table 1).
Several different dosages of iron and folic acid were used in these trials. Three studies supplemented 60 mg elemental iron and 0.25 mg folic by incorporating a weekly schedule [62,64,69], one study supplemented each of 120 mg iron and 3.5 mg folic acid [61], 100 mg iron and folate 500 μg to two groups in a daily and weekly manner [60], daily 60 mg elemental iron and 0.5 mg folic acid [65], weekly 65 mg of elemental iron with 0.25 mg folic acid [66], once daily or weekly 350 mg iron and 1.5 mg folic acid [67], daily or twice weekly 60 mg iron and 0.5 mg folic acid [68], and weekly 200 mg ferrous fumarate and 200 mg folic acid [63]. One study had a daily supplementation group [65]. Six studies had weekly arms [61][62][63]66,69]. Two studies had a daily arm and a weekly one [60,64,67]. One study had daily and twice-weekly supplementation groups [68]. All trials provided the supplementation during the preconception period. Duration of supplementation: one study supplemented for 8 weeks [66], one study supplemented for 10 weeks [64,67], seven trials provided supplementation for more than 12 weeks [60][61][62][63]65,68,69].

Discussion
Regarding interventions aimed at delaying the age of first pregnancy, our review found three comparisons comprising of different interventions related to this intervention. Education on sexual health and contraception was the most commonly employed intervention, and the use of birth control methods was the most reported outcome. Education intervention alone showed an insignificant impact on the risk of unintended pregnancies. However, it showed a significant increase in ever having used a condom by 71%. The evidence on interventions on education and provision of contraception came from single study showing a significant impact on improving the usage of any method of contraception by 49% and the usage of condoms by 14%.
Interventions on optimizing inter-pregnancy intervals did not show a significant impact of education on and contraceptive provision along with male partner involvement on the risk of unintended pregnancies when compared to education only. However, educational intervention alone or with provision of contraceptives showed a significant improvement in the uptake of the use of contraceptives.
This review also investigated the impact of folic acid and iron-folic acid supplementation. Overall, folic acid use reduced NTD incidence by 47%. Folic acid supplementation had an impact on NTDs that differed by its dosage, as the meta-analysis showed a significant impact at 0.4 mg (reduction in incidence of NTDs by 47%) and a non-significant impact at 5 mg. For iron-folic acid, it was found that this supplementation reduced the prevalence of anemia by 34% when compared to a placebo. Weekly supplementation reduced the prevalence of anemia by 32%, while daily supplementation did not show any impact. Anemia prevalence was reduced by 34% at schools and did not show a significant impact at work when iron folic acid supplementation was given compared to a placebo. The current evidence does not support any significant difference between use of ironfolic acid and placebo to decrease adverse effects.
The results that interventions, such as the provision of contraceptives and education to delay the age at first pregnancy, are consistent with earlier reviews [17,[78][79][80][81]. Hindin included 21 studies from LMICs, which found that increased contraceptive use and delay in the age of sexual debut after interventions to prevent either unintended pregnancies or repeat pregnancies [78]. Similar to our review, Oringanje included 53 studies from LMICs and high-income countries (HICs), which also did not find evidence on reducing the risk of unintended pregnancies [17]. Similar to our review, two earlier reviews also showed that education interventions were effective at bringing about significant improvement in sexual knowledge, contraceptive use, and decreasing adolescent pregnancy and improving birth intervals [80,81].
The result that periconceptional supplementation of folic acid reduces the incidence of NTDs is consistent with other earlier reviews [12,82,83]. The results of our review on periconceptional ironfolic acid supplementation are consistent with Fernández-Gaxiola, although that review only assessed the use of iron among menstruating women (alone or with other micronutrients] in reducing the prevalence of anemia, regardless of dose used [84]. However, unlike Fernández-Gaxiola [84], our review showed that a reduction in anemia prevalence differs with the duration of iron-folic acid supplementation [no significant impact was found for supplementation for less than eight weeks).
Our review found that educational interventions that aim to delay pregnancy or optimize interpregnancy intervals can benefit in improving the contraceptive use and knowledge but there is insufficient evidence for the impact on unplanned pregnancy, which could be due to the shortage of included studies reporting on this outcome. Educational interventions aiming to optimize interpregnancy intervals were more effective than the provision of contraceptives; however, there were limited studies included in these analyses, therefore intervention strategies need to be investigated further to determine the most effective approach.
Regarding folic acid supplementation, our review further supports folic acid usage to reduce NTDs; however, the GRADE assessment determined that there was a very low certainty of evidence. Our review also favored the preconceptional use of iron-folic acid to reduce anemia, where weekly supplementation regimes were more effective. There is evidence that iron-folic acid supplementation is most effective when supplementation takes place in school settings compared to work settings, perhaps because students are more easily supervised and therefore adherence is more consistent. Therefore, our review suggests that vitamin supplementation in schools are an effective strategy to reduce anemia and that there is continuing evidence for folic acid usage to decrease NTDs although the optimal dosage is still unclear.
Further evidence is required for each of the intervention targets in this review. First, educational interventions that aim to delay pregnancy or improve inter-pregnancy intervals need to more consistently report on unplanned pregnancy. With reproductive and sexual health education often emphasizing the importance of avoiding unplanned pregnancies, studies need to ensure that these educational interventions are translating into behavior beyond contraceptive use to include the prevalence of unplanned pregnancies. Further research is required to determine whether school, community, clinical, or a combination of settings is optimal for these interventions, though there is evidence from our review that studies in community settings were more effective. A majority of the interventions had an educational component; while there was some adequate evidence from our meta-analyses of the use of educational strategies, there were fewer studies with multidimensional components (such as education and provision of contraceptives), and therefore these strategies need to be further investigated to determine whether they are just as effective or better than education alone.
Our review was consistent with other reviews investigating whether supplementing folic acid and iron-folic acid reduces the incidence of both NTDs and anemia, respectively, and therefore supplementation should be implemented before conception. Research is required to dictate the best methods to ensure consistent supplementation use of folic acid and iron-folic acid in order to determine which dosages and durations are most beneficial for maternal and neonatal outcomes. Regardless of study design, anthropometric measures should be collected when possible to ascertain the impact of micronutrient supplementation on related outcomes as this review could not report on neonatal outcomes, such as birth weight and small for gestational ageSGA. Since we included an array of study designs, it is therefore unsurprising that GRADE assessment varied from medium to very low quality.

Conclusions
While there is a growing body of evidence in support of the provision of preconception care, the effectiveness of interventions to delay the age of first pregnancy, optimize birth intervals, and to provide periconception folic acid and iron-folic acid supplementation warrant further inquiry. It is vital that we determine the most effective delivery mechanisms across different settings such that successful implementation of pre-and periconception interventions can take place in LMIC settings [85]. This review summarized and collated the present evidence for interventions executed during the pre/periconception period that aimed to delay the age at first pregnancy, optimize birth intervals, and increase the supplementation of folic and iron-folic acid with a particular focus on adolescent girls. The evidence for educational interventions focusing on delaying and optimizing intervals demonstrated a promising increase in the uptake of contraceptive use; however, no significant impact was reported for the primary outcome of unplanned pregnancies. For interventions focusing on folic acid supplementation, this review provides further evidence that this intervention in LMIC settings can successfully reduce the incidence of neural tube defects, whereas iron-folic acid supplementation can improve rates of anemia, particularly when supplemented weekly and monitored in a school setting. While we note it is important to include evidence beyond RCTs to ensure contextual factors are appropriately captured, further RCTs are required, especially for inter-pregnancy intervals and in broader LMIC locations, especially in the Americas.  [3] "reproductive aged women" OR "teenager" OR (MH "Adolescence") OR "girl" (MH "Puberty") OR (MH "Menarche") OR "pubescent" OR (MH "Puberty") OR (MH "Menarche") OR "pubescent girl*" OR "menstruating girl*" OR "menstruating women" [1] AND [2] AND [3] PsycINFO: Delay/Interval [1] ("preconception care" or "prepregnancy care" or "preconception" or "periconception").af.