Evaluating the Effectiveness of Educational Interventions in Family Planning for Men in Developing Countries: A Systematic Review

Introduction: Unintended pregnancy is associated with national socioeconomic development and gender inequality. In addition to contraception, educational interventions that promote family planning and address gender dynamics are considered important in preventing unintended pregnancy. While the importance of encouraging men’s participation in family planning has been advocated, most studies have focused on the application of interventions to women or populations in high-income countries only. Therefore, we conducted a systematic review to evaluate the effects of educational interventions on men in low- and middle-income countries in terms of knowledge, attitudes, practices, and gender dynamics. Methods: Three electronic databases (CINAHL, Ovid MEDLINE, and Web of Science) were searched for studies published from January 1980 to October 2022. Keywords such as “men/husband,” “family planning,” “contraception,” and “education” were combined to identify studies. Two independent reviewers conducted screening and data extraction, and the risk of bias was assessed using the Risk of Bias 2 tool. The quality of evidence was evaluated according to the GRADE Handbook. Results: The database search identified 16,086 articles, of which 4 cluster randomized controlled trials (RCTs) and 1 RCT were ultimately included. Each of them was conducted in four different countries: Malawi, Guatemala, Tanzania, and India. Changes in knowledge, attitude, family planning, and gender dynamics were the outcomes used to assess the effectiveness of interventions. The five selected articles exhibited an effect on ≥1 indicator for each outcome. However, the quality of evidence was determined to be low or very low owing to the risk of bias, heterogeneity, and imprecision. Conclusions: Determining the effectiveness of educational interventions in family planning for men in low- and middle-income countries requires additional high-quality intervention studies. As family planning is influenced by various background factors, it is important to develop appropriate interventions for each context and define relevant indicators that can be compared across contexts.


Introduction
Unintended pregnancy is associated with social and economic development and gender equality in a country (1) ; hence, the inclusion of men in family planning is imperative (2) .Indeed, unintended pregnancies occur worldwide, with an estimated 121 million cases annually (3) .
Unintended pregnancy is defined as "a pregnancy occurring to a woman who was not planning to have any (more) children, or that was mistimed, in that it occurred earlier than desired (1) ."According to Bearak et al., the average rate of unintended pregnancy worldwide was estimated to be 64 per 1,000 live births among women of reproductive age (15-49 years) in the years 2015-2019 (3) .Although this average has been improved from the estimated unintended pregnancy rate of 79 reported during 1990-1994, disparities exist when regional and socioeconomic conditions are considered.The World Bank estimates rates of 34, 66, and 93 unintended pregnancies per 1,000 live births in high-, middle-, and low-income countries, respectively, indicating an inverse relationship between unintended pregnancy and national income (2) .
Supporting sexual and reproductive health and achieving gender equality are part of the United Nations' Sustainable Development Goals and the United Nations Population Fund, a subsidiary agency of the United Nations that works to ensure reproductive rights for all and supports access to a wide range of sexual and reproductive health services (1) .Specific measures to combat unintended pregnancy have been attempted through material means, including the provision of contraceptives and contraceptive pills and expansion of contraceptive services (1) .While contraceptive use is increasing worldwide (4), (5) , progress may vary between countries.According to Kantorová et al., contraceptive needs are not being met in developing regions, which have an estimated 214 million people of reproductive age (4) .Therefore, providing access to contraceptives in populations with unmet needs can prevent unintended pregnancies.However, rather than the lack of access to contraception, barriers, including the lack of knowledge, concerns regarding side effects, social stigma, and opposition from partners and others, are the major obstacles to contraceptive use in low-and middle-income countries (1), ( 5), ( 6), (7) .Therefore, contraceptive coverage alone is insufficient to reduce the rates of unintended pregnancy; educational interventions, such as promoting family planning and addressing underlying social norms, including gender dynamics, are also essential (5) .
Family planning can be defined as "the information, means, and methods that allow individuals to decide if and when to have children (1) ."It can prevent unintended pregnancies or pregnancies in young girls, increase educational opportunities, empower women and girls, and promote a sustainable population growth (1), (8) .The incidence of unintended pregnancy is closely related to gender inequality.Thus, gender dynamics are important in family planning (1), (9) ; it is crucial to promote men's participation in family planning and establish communication between couples and partners to promote family planning (2) .
Previous studies have reported that approximately 59% and 42% of men in rural Ethiopia and in pastoralist communities, respectively, are involved in family planning (10), (11) .Furthermore, 48% of men in rural Ghana reportedly participate in family planning services (12) .Factors related to men's participation in family planning include communication with their spouses concerning sexual and reproductive health, family planning, and agreement with family planning (10), (11) .In addition, a systematic review in the Philippines found that the most common barrier to male participation in family planning was the lack of knowledge regarding the proper use of contraception and its side effects (13) .
Educational interventions can promote family planning.A cluster randomized controlled trial (RCT) based on pastoralist communities in Ethiopia reported that education among men regarding family planning had promoted family planning use and attitudes toward family planning (14) .An RCT involving married women in Jordan reported an increase in modern contraception use in the group that underwent educational counseling for couples compared with the group that did not (15) .However, the impact of education on men remains unclear as these studies have mainly focused on women; their responses were applied to assess impact.An RCT of a fishing community in Uganda reported that education regarding family planning using easy-to-understand handouts increased knowledge and the use of family planning methods (16) .However, because that study included both men and women, the results were mixed, and the impact of education on men was unclear.A systematic review of the effects of educational interventions focused on contraception in family planning showed that while various educational approaches can increase contraceptive knowledge, there is insufficient research to evaluate their effects on contraceptive attitudes, selection of effective methods, and contraceptive use.Although the included studies had high quality (17) , they were conducted in high-income countries, with only a few focused on men in low-and middle-income ones.
Therefore, we conducted a comprehensive systematic review of the effectiveness of educational interventions on family planning in low-and middle-income countries, particularly focusing on their impact on men.

Materials and Methods
The study was conducted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).Particularly, the Japanese version of the checklist translated by Hiroharu et al. was used in this study (18) .The English version is listed in Appendix 1.The systematic review was registered with PROSPERO (CRD42022374221), and no protocol article was published.

Inclusion/Exclusion criteria
We included primary quantitative RCTs or cluster RCTs published in English after 1980.Promotion of family planning programs began at the International Conference on Population and Development in 1974, and the concept of sexual and reproductive health/rights was introduced in 1994.These worldwide discussions on family planning as well as sexual and reproductive health/rights have led to increased research on these topics since the 1980s.Studies not falling under the above categories and those in the form of review articles, conference proceedings, and protocols were excluded.
We included men from low-and middle-income countries (i.e., countries with Gross National Income < $13,205 based on the World Bank's Gross National Income per capita income classification) (19) .Men from high-income countries were excluded.No restrictions were imposed on the number of children, history of contraceptive use, history of unintended pregnancy by the partner, or education on family planning before the intervention.We excluded interventions aimed at preventing sexually transmitted diseases (STDs) or the recurrence of STDs in populations at a high risk of STDs.Sex workers, sexual minorities, and men from low-and middle-income countries who had migrated to high-income ones were excluded.
We included all educational interventions based on family planning, contraception, and sexual and reproductive health (i.e., interventions aimed at promoting family planning practices, such as contraceptive use, and improving attitudes toward family planning).No restrictions were imposed on the intervention format (remote/face-to-face), duration, or providers.Interventions targeting couples or communities that included men were also included, as were interventions targeting men alone.
As regards the comparison of this review, we did not impose a limitation on a specific population as a comparison group.However, studies without a comparison group were excluded.
As regards the outcomes of this review, we focused on knowledge, attitudes, and practices concerning family planning and gender dynamics in decision-making.These outcomes were set with reference to two systematic reviews that evaluated the effects of educational interventions on family planning (17), (20) and explored the factors associated with family planning in low-and middle-income countries (6) .Gender dynamics refer to the relationships and interrelationships between men and women.

Search strategy
Electronic database searches were conducted on the CI-NAHL, Ovid MEDLINE, and Web of Science (last search date: 2022/10/11).The search term was a combination of keywords related to men and spouses (men, partner, husband) as terms for the target population; keywords related to sex education and family planning (such as sex education, family planning, contraception, and birth control) as terms for the intervention type; and keywords related to education (education, intervention).Furthermore, we used a combination of keywords related to sex education (sex education), family planning (family planning, contraception, birth control), and education (education, intervention) as terms to indicate the study design.Furthermore, keywords related to the study design (RCT, clinical trials) were used.We referred to the search formulas of Pazol et al., Sharma et al., and King's College London's "Searching for Systematic Reviews: Advanced Search Techniques" to develop the search formulas (17), (20), (21) .The search formulas are presented in Appendix 2.

Study selection
Five reviewers performed the study selection.First, electronic databases were searched using a search formula, and the literature from each database was imported into Covidence (https://app.covidence.org;Covidence Inc., Melbourne, Australia), a tool for systematic reviews.Duplicate articles were excluded.Titles and abstracts were screened, followed by fulltext screening by two independent reviewers.The full-text screening was performed in two stages, with the first stage focusing on the population and intervention-based content and the second focusing on outcomes.Disagreements were resolved through discussion between the two reviewers, and when disagreements persisted, a third reviewer was included to resolve them.

Data extraction
Data were extracted independently by two reviewers.A format created in Microsoft Excel (Microsoft Corp., Redmond, WA) was used for data extraction and organization.The following data were extracted: study author, year, study country, study design, study objectives, population characteristics (eligibility criteria, characteristics, sample size), randomization process, intervention group/comparison group characteristics (implementer, duration, composition), outcomes (indicators, measurement methods), data analysis, main results, limitations, and implications for the future.

Risk-of-bias assessment
Risk-of-bias assessment was independently conducted by two reviewers.The Cochrane Risk of Bias 2 tool, which has models for RCTs and cluster RCTs, is available on the Web as an open resource (22) .The risk-of-bias assessment items for bias included five domains for RCTs: bias arising from the randomization process, bias owing to deviations from the intended intervention, bias owing to missing outcome data, bias in the measurement of outcomes, and bias in the selection of reported outcomes.For cluster RCTs, aside from the five domains of RCTs, the domain of bias owing to the timing of participant identification or recruitment in cluster randomization was also included.Based on these assessments, the reviewers judged the bias on a 3-point scale of "low," "some concerns," and "high," with any disagreements resolved by discussion between the two reviewers.A narrative synthesis was used to integrate the results.

Quality-of-evidence assessment
The GRADE Handbook (23) was consulted to assess the overall quality of evidence of all the included studies.As we focused on RCTs and cluster RCTs as the study design of interest, the level was reduced by one or two when there was a risk of bias, inconsistency, indirectness of evidence, imprecision, or publication bias and finally rated as "high," "moderate," "low," or "very low."

Results
Overall, 16,086 articles were identified during the database search.Of these, 1,956 were excluded because of duplicates, resulting in 14,130 articles eligible for screening.The title and abstract screenings excluded 14,017 articles as they did not meet the eligibility criteria, and 113 articles were included in the full-text screening.The full-text screening excluded 108 articles; thus, 5 articles were finally included in the study.The reasons for exclusion, ranked by priority at the time of screening, were as follows: not an original article, abstract only, different setting (high-income country), setting unknown, differ-ent study design, different population, different intervention, different outcome, and results obtained only from men. Figure 1 presents the process of study selection.
Two interventions (24), (25) were delivered by peers living in the study area.In other studies, interventions were delivered by local healthcare providers (27) , a combination of people (including peers, teachers, and healthcare providers) (26) , or by an online curriculum (28) .Comparisons were made with groups receiving no intervention (three studies) (24), ( 25), (26) , only referrals to public family planning services (one study) (27) , and only cash benefits (one study) (28) .The outcomes were categorized into knowledge, attitude, practice, and gender dynamics of family planning.3. Table 4 summarizes the results of the effectiveness of the interventions.Outcome measures were described as they appeared in each study as follows: 1)Knowledge: One of the three studies examining the effect of educational interventions on knowledge demonstrated statistically significant improvement in one outcome measure (family planning knowledge).The study targeted the community and included interventions for men and women, as well as for couples (25) .The intervention comprised informational and role-playing sessions.After 2 months, the number of known contraceptive methods increased.The risk of bias in the measure was assessed as "some concern."2) Attitude: One of the two studies examining the effects of educational interventions on attitude demonstrated a statistically significant improvement in one outcome measure (communication frequency).The study targeted men, and peer support on contraception was provided as motivation and information (24) .The questionnaire at 6 months reported a significant increase in the frequency of communication regarding family planning with their partner.The risk of bias in the measure was assessed as "some concern."3) Practice: Two of the three studies examining the effects of educational interventions on practice demonstrated statistically significant improvements in a total of three outcomes measured: sexual initiation, condom use, and contraceptive uptake.The school-based intervention reported significant differences in condom use and age of sexual initiation at 12 months (26) .Interventions targeting men of reproductive age also reported significant increases in contraceptive uptake (24) .The risk of bias was rated as "low" for sexual initiation and contraceptive uptake and "high" for condom use.4)Gender dynamics: Two of the three studies that examined the ef-fects of educational interventions on gender dynamics found statistically significant improvements in two outcome measures: equitable attitudes toward household decision-making and gender attitudes.An intervention on family planning and gender ideology for couples reported a significant improvement in "Equitable attitudes toward household decision-making" at 9 months (27) .Conversely, no significant increase was observed at 18 months, indicating that the effect was not sustained.Interventions for communities reported significant increases in scores for equal attitudes toward gender (25) .The riskof-bias assessment in the outcome measures was "high" for equitable attitudes toward household decision-making and "some concerns" for gender attitudes.Because the study design for all outcomes was RCT or cluster-RCT, the GRADE rating was reported as "high." 1) Knowledge: According to the GRADE Handbook, the three sources of evidence reporting knowledge were rated as "very low" with downgrades for risk of bias, imprecision, and inconsistency.First, the risk of bias was downgraded by one grade to "moderate" as all outcome measures were rated as "some concerns."Inaccuracy was downgraded by one grade to "low" as it included articles for which 95% confidence inter-Table 3.2.Risk-of-Bias Assessment: Attitude.vals (CIs) could not be accurately evaluated.Heterogeneity was downgraded to "very low" as it was difficult to interpret the consistency of the effects.Indirectness was not determined because no population or outcome showed indirectness.Publication bias was not determined as the number of studies was insufficient for evaluation.2) Attitude: The two pieces of evidence reported on attitudes were evaluated holistically and downgraded to "very low" for risk of bias, imprecision, and inconsistency.First, the risk of bias was downgraded by two grades to "low" as two outcome measures were rated as "high" and four as "some concerns."The risk of inaccuracy and inconsistency was downgraded to "very low" for the same reasons as above.Publication bias was not determined owing to the insufficient number of studies.3) Practice: The three studies were rated as "low" with downgrades for imprecision and inconsistency.First, concerning the risk of bias, one outcome measure was rated as "high," whereas the other five were rated as "low" and, thus, were downgraded based on the GRADE Handbook guidlines that most information across studies comes from studies with low risk of bias.Therefore, downgrading was not performed.Next, the imprecision was downgraded by one grade to "moderate" because the study in-cluded articles for which 95% CIs could not be accurately evaluated.Heterogeneity was downgraded to "low" as it was challenging to interpret the consistency of the effects.Indirectness was downgraded to "low" because no population or outcome showed nonindirectness, and publication bias was not determined because of the insufficient number of studies.4) Gender dynamics: The three evidence sources reported on gender dynamics were holistically evaluated and downgraded to "very low" for risk of bias, imprecision, and inconsistency.The risk of bias was downgraded by two grades to "low" given that two outcome measures were rated as "high" and two as "some concerns."The risk of inaccuracy and inconsistency was downgraded to "very low" for the same reasons as above.Publication bias was not determined because of the insufficient number of studies.

Discussion
This systematic review investigated the effects of educational interventions on family planning for men in low-and middleincome countries and evaluated outcomes, including changes in family planning knowledge, attitudes, practices, and gender Table 3.3.Risk-of-Bias Assessment: Practices.dynamics.Five articles were selected for the review, including four based on couples and one based exclusively on men that exhibited effects on ≥1 indicator for each outcome.However, the evidence quality was rated as "very low" for knowledge, attitudes, and gender dynamics and "low" for practice because the heterogeneity of the measures did not allow for a consistent evaluation of the intervention effects, bias risk remained, and data, including CIs, were insufficiently described.
Community-based approaches to educational interventions have yielded positive results, as demonstrated by Schuler et al (25) .According to findings from this review, community education on family planning can contribute to increased family planning knowledge (20) .However, the study by Schuler et al. may have been limited to rural areas, where contraceptive use is low and family planning needs are unmet compared with urban areas.Indeed, approaching groups with unmet contraceptive needs enhances the promotion of contraceptive use (5) ; hence, we believe that approaching groups with high needs is more effective in terms of intervention.A study of fishing communities in Uganda, where family planning needs are considered high, found an increase in knowledge among men and women, which suggests that community education for high-need groups has the potential to improve knowledge; however, in-depth studies are warranted.
Shattuck et al. reported that interventional effectiveness was obtained for attitude; Mmbaga et al. and Shattuck et al. reported that practice was the outcome of their respective interventions.By outcome, more effect indicators were obtained for practice than for attitude (24), (26) .The behavioral change stage model suggests that attitudes change from a period of indifference to a period of interest, followed by behavioral change (33) .Therefore, it can be inferred that attitude change is observed before behavioral change.However, our findings did not conform to this assumption.The first reason is that these outcome measures were not obtained from the same study.Second, the age and marital status of the participants and timing of outcome measure differed in each study.Thus, the results may differ depending on the characteristics of the target population.Third, all studies used self-reported measures of practice, which may have been more effective than other outcome measures as they used simple measures such as using family planning methods and were also considered to be more directly affected by intervention content.In low-and middleincome countries, increased knowledge of family planning positively impacts other aspects of contraceptive use, including positive attitudes toward contraception and contraceptive Table 3-4.Risk-of-Bias Assessment: Gender Dynamics.use retention (17) .While the effects of knowledge, attitudes, and implementation are important, exploring their interactions will create more effective educational interventions to promote behavioral change in family planning in the future.
Fleming et al. and Schuler et al. reported the effects of educational interventions on gender dynamics and included couples (25), (27) .According to Tilahun et al., educational interventions on family planning, such as including couples and promoting communication between spouses, can help promote contraceptive practices (34) .Furthermore, a systematic review of behavioral change techniques suggests that the most effective interventions involve the male partner in contraceptive decision-making and combine multiple behavioral change techniques (35) .Therefore, interventions involving men and couples can affect gender dynamics, which is a fundamental aspect of the decision-making process.However, according to Ruane-McAteer et al., evidence on intervention studies to improve gender inequalities in sexual and reproductive health and rights is generally of low quality and inconclusive (36) .Thus, further evidence on the effects of interventions on gender dynamics in family planning is needed.terpreted based on the characteristics of the setting and target population.

Conclusion
It is necessary to conduct more high-quality intervention studies that consider various factors and generate evidences through more comprehensive searches in the future.As various background factors influence family planning, it is important to develop appropriate interventions in each context.In this work, the findings suggest that interventions targeting high-need populations and educational interventions that include community and couple interventions may be helpful.
In addition to developing interventions that fit various contexts, it is important to provide comprehensive evidence in low-and middle-income countries.Thus, it is necessary to develop consistent and appropriate outcome measures that can be adapted across diverse contexts to ensure accuracy in reporting.
Four outcome measures were extracted for knowledge: knowledge of contraception, knowledge limited to modern contraception, and knowledge of family planning.Outcome measures for attitude included attitude toward family planning, self-efficacy, communication with partners about family planning, and action plans for delaying sexual activity and condom use.Practice outcomes included indicators of contraceptive use, condom use, and initiation of sexual activity.Gender dynamics outcomes included gender norms and fairness in decision-making within the household.Outcomes were measured once or twice following the intervention, and the timing varied from 2 months after the baseline survey to 18 months.The results of the risk-of-bias assessment for each outcome are presented in Table

Table 1 .
Summary of Studies Included.

Table 2 .
Supplementary Information of Each Country.