The association between men’s family planning networks and contraceptive use among their female partners: an egocentric network study in Madagascar

Ensuring universal access to family planning (FP) services is a priority in the 2030 Sustainable Development Goals [1]. Removing barriers to access FP services, as well as making sure women have the information and support they need to seek reproductive health services, receive respectful care, and be able to effectively use contraception when desired will allow women to exercise their reproductive autonomy [2]. Meeting women’s FP needs has been shown to help reduce maternal mortality, which is particularly high in sub-Saharan African countries such as Madagascar [3, 4]. While there is growing research showing that women’s social networks, that is their ties with family members, friends, and health care providers and the nature of these ties, affects their contraceptive use [5‐14], male partners can also influence whether women use contraception [15]. Studies have shown that perceptions of whether male partners are supportive of contraceptive use can influence whether women use contraception [16‐22]. For example, concern about opposition to contraceptive use, mainly from male partners, was identified as a barrier to contraceptive use among women in Madagascar [23]. Additional research has found that involving men in contraceptive counseling was associated with increased contraceptive use [24‐27], though other studies have found negative or no effects [28, 29]. Most research on the role of men in contraceptive decision-making has focused on their individual characteristics, such as age, education, urbanicity, as well as social and cultural factors such as religious beliefs, desired family size, gender roles and male identity [30‐34]. However, very few studies have moved beyond looking at men’s individual characteristics to consider whether men’s social networks influence women’s contraceptive use.

Insofar as women’s contraceptive use may be influenced by their male partners, women’s contraceptive use may also be influenced by the networks of their male partners, i.e., the ties that men have with their partner, family members, friends and health providers [35, 36]. Based on the nature of these ties, networks can provide informational, emotional, and financial support, offer opportunities for peer learning, and/or social engagement, while being influenced by the socio-cultural context within which they exist. Whereas researchers have increasingly used social network theory and analysis to explore how women’s reproductive health decisions are tied to others in their networks [37‐39], very little research exists examining how the composition of men’s social networks may influence, in part, female partners’ contraceptive use. A longitudinal network study in Kenya found that couples were significantly more likely to use contraception if the male partner had at least 1 contraceptive user in their own network, adjusting for unmeasured confounding and the tendency for individuals to have ties to other individuals with similar characteristics (i.e., homophily) [9]. Another longitudinal study found that men in Ghana who received positive encouragement from their social network to use contraception were more likely to report using contraception with their partners [40].

Men may differ in the types of social ties they rely on for seeking information and advice about FP including contraceptive use. Some men may rely on health providers (e.g. community health workers [CHWs], doctors, nurses, or mid-wives) for FP and contraceptive information and advice, while others may rely on social ties, such as friends, partners, and family members. Men may also rely on a combination of both of these, or on no one at all. In-depth interviews among married men in Uganda found that men preferred to rely on partners for contraceptive information as opposed to health providers, peers, or mass media [41]. Focus group discussions with men in Kenya revealed that men did not trust either their health providers or their partners for FP information [33]. The extent to which the type and composition of ties in men’s FP networks impacts contraceptive use among couples is unknown.

While relying on health providers for advice is expected to be positively associated with couple’s contraceptive use, the influence of other kinds of ties within men’s social networks on couples’ contraceptive use is more ambiguous. A study among women in the United States showed that women tend to hear negative or incorrect information about contraception from their social relations compared to providers, but it is unclear if this might be the case for men as well [42]. However, this association has not been studies among men, especially within a resource-limited setting. Additionally, networks can constrain individual autonomy in reproductive health decisions by inducing individuals to align their behaviors with other network members [43, 44]. Information about the extent to which the composition of men’s networks for FP information and advice may be associated with couple’s contraceptive use is limited. Addressing this gap in the literature could inform the design of FP interventions that leverage men’s social networks to foster their engagement in contraceptive decision-making, support their partner’s choices in relation to contraception use and facilitate the goal of universal access to FP for all women.

In this egocentric network study among men in Madagascar, we collected data from men about who they turn to for information and advice regarding FP and how these FP ties are associated with their female partners’ contraceptive use. We assessed 1) whether men were more likely to report that their partner was using modern contraception if they personally had at least one FP tie to turn to for information and advice about FP, 2) whether the composition of men’s networks was differentially associated with couples’ contraceptive use (that is networks made up of social connections versus health care providers), 3) whether men who knew other individuals using contraception were more likely to report that their partner was using modern contraception, and 4) whether the characteristics of men’s ties, such as type of interpersonal relationship and the perceived support of contraceptive use by that person, was associated with couples’ contraceptive use.

There exists very little evidence examining men’s networks for information and advice about FP and how these networks are associated with women’s contraceptive use. The findings from this novel egocentric study show that many men (82%) report having at least one person who influences their FP use by providing information, advice or guidance. Although men are twice as likely to be influenced by providers by relying on them for information, advice and guidance about FP (54%) compared to family and friends (26%), men having any kind of FP network were 2 times more likely to report their partner using modern contraception compared to men with no FP network. Having more than one tie was not differentially associated with contraceptive use compared to having just one tie. Overall, the findings demonstrate that having at least one tie who influenced their FP use, be it a social relationship or a provider relationship, is what differentiates men who report that their partner is using contraception. Interventions to ensure women can exercise their reproductive autonomy and use contraception if desired should adopt a multi-pronged approach that includes male partners and their networks to foster their supportive engagement in contraceptive decision-making.

Men who reported being influenced by their partner for FP decision-making, by turning to them for FP information, advice and guidance, are even more likely to use contraception. This finding is consistent with evidence from Uganda where married men were less likely to rely on health providers or their peers, and they preferred to rely on their partners due to their partners’ experience with side effects, knowledge gained from their own peers, and exposure to mass media campaigns and health providers [41]. Men who nominated their partner as their social tie influencing their FP decision-making likely have more open couples’ communication about contraception, a critical factor affecting contraceptive use [52, 53]. The absence of significant coefficients for joint couples’ roles in decision-making about contraception may mean that actual decision-making resides with the woman, but more work needs to be done to explore these mechanisms in depth. In addition, future research should also examine concordance within couples regarding perceptions of contraceptive use, since this study relied on men’s perceptions of their partner’s contraceptive use, which may differ from their actual contraceptive behaviors.

Relying on other social ties, specifically siblings, was also associated with higher contraceptive use as a couple, compared to relying on doctors, nurses and mid-wives. Similar to the role of women’s social ties in Madagascar which play a positive role in increasing the likelihood of contraceptive use [14], men’s social ties also play a supportive role for contraceptive decision-making. These results contrast with other research finding that, for women, they were more likely to receive negative and incorrect information about contraceptives through social ties [42] and that social networks can exert a negative influence on adoption of reproductive behaviors [43, 44]. Our findings are consistent with the network study in Ghana showing that couples’ contraceptive use was higher among men who received support through their social networks to use contraception [40]. Despite the positive influence of social ties, only half as many men turned to social ties compared to providers which could point to the more sensitive nature of this topic in these communities.

There were no differences in the likelihood of contraceptive use depending on the type of health provider. Men were just as likely to use contraception whether they relied on nurses, mid-wives or doctors for information and advice about FP, compared to when they relied on CHWs and health educators. While another study found that women in Madagascar were more likely to use contraception if they relied on CHWs in their FP network compared to relying on nurses, mid-wives and doctors [14], the association in this study was not statistically significant for men in this study. The lack of precise estimation is possibly due to a small sample size. The findings also demonstrate that men primarily relied on women rather than men for FP information and advice. This tendency reflects the high proportion of women who occupy the roles of providers including mid-wives, nurses, CHWs and health educators tend to be women, in addition to partners and female siblings.

Similar to past work [8], men in this study were more likely to use contraception if they knew someone else using contraception. Yet, the associations were attenuated when network composition was included. Having FP network that is supportive of contraceptive use may be more important than having a network that includes contraceptive users [6, 8, 9, 11, 12]. Valente et al. (1997) show that, among women, their perceptions of ties’ approval of contraceptive use was more important than their ties’ actual use [8]. While our data did not allow for a distinction between alters’ active use versus ego’s perception of their use, future research should explore the role of both actual and perceived contraceptive use by alters on contraceptive use. Specifically, socio-network studies examining discordance in perceptions between egos and alters would be able to show whether perceptions of contraceptive use matter more than actual contraceptive behaviors. Research on social norms in sub-Saharan Africa has found that perceptions about whether friends or most people within a local reference group engage in a health-related behavior are potentially more important than peers’ actual behaviors [8, 13, 54, 55].

There are several study limitations. First, the relationship between having a FP network, including the composition of that network, and the decision to use contraception may be endogenous. For example, the reverse relationship could exist where couples decided to use contraception and men then selectively formed their FP network to validate that decision. Due to feasibility, it was not possible to conduct a complete network study to examine dynamic selection in this study population and establish the causal role of networks on contraceptive use. It is plausible that the main findings reflect a combination of selection (i.e., forming networks with individuals that have similar contraceptive behaviors to validates one’s contraceptive decisions) versus the influence of those networks on contraceptive behaviors. As with other egocentric network studies, this could not be teased out with the data. For this reason, we consistently referred to associations rather than using causal language. If it was possible to distinguish between selection and influence of networks, our estimates are likely to be smaller in magnitude though not zero, as put forth by Fowler and Christakis (2008) [56]. In addition, the opportunity for men to selectively form their FP network may be limited in this particular study context. Half of the men had a provider-only network and were constrained in selectively forming their FP network. In the study area, there are typically only 1 health educator or CHW per village and 1 nurse at the public health center providing FP. Men may have more ability to selectively form their social-only network. Future research using multiple data waves would facilitate identifying the causal influence of FP network composition on contraceptive use. Furthermore, if there was an omitted variable which influenced couples’ decision to use contraception (e.g. men’s personal preference for contraceptive use) and this factor also influenced FP network formation, the relationship between that omitted variable and these two variables would need to have a RR of at least 3.2, based on the e-value sensitivity analyses proposed by VanderWeele and Ding (2017) and Oster (2019) [50, 51], to completely explain away our main findings. The example of another unobservable (access to health services) did not meet this threshold for ruling out the main findings.

Second, men’s networks may be closely tied to their female partner’s social networks, by virtue of being in a couple. One of the study’s limitations is that the data were not collected by couple, therefore it is not possible to assess correlations in a couples’ FP networks. The results could be confounded by the female partner’s FP network which we cannot adjust for. Due to potential sensitivities of interviewing couples about FP and sexual health/practices in Madagascar, it was more feasible to recruit men without their partner. There is no study to our knowledge comparing differences by FP networks within couples and future studies examining couples would be beneficial. Third, there may be concern that including female partners within men’s networks is problematic because the outcome is simultaneously measuring those female partners’ contraceptive behaviors. Yet, when female partners were excluded from the definition of men’s FP networks, the results remained consistent.

Fourth, it is not possible to rule out the influence of potential unmeasured confounders such as shared environmental factors that may influence both men’s FP networks and contraceptive use. One such unmeasured factor might be access to FP services. The analyses adjusted for differences by fokontany (sub-district) which were statistically significant and suggest that there may exist differences in ease of access to FP services. Whether men have a FP network and the composition of that network may vary with their access to FP providers, thereby potentially biasing our estimates away from the null. However, there is no known research demonstrating such an association, and the participants in our study were all equidistant from FP providers since they were recruited from the same geographic area. One study from Madagascar showed that integrating FP services at the community level to improve access resulted in a two-fold increase in contraceptive use [57], but the effect size estimated from that study is smaller than our estimated e-value of 3.2. Therefore we believe it is unlikely that failure to account for confounding by access to FP services could completely explain away the main findings. Fifth, as an egocentric study, the data did not include alter-level contraceptive use for comparison to ego’s perceptions. Sixth, the sample may not be representative of rural populations in Madagascar since the villages were within a 5 km radius of a well-resourced semi-urban town. Couples’ contraceptive use in the sample (66%) was higher than reported contraceptive use (36%) in Madagascar as a whole [45], potentially due to ease of access to providers in the study region. Lastly, while the difference in missingness of covariates just misses statistical significance at the 5% level, it suggest that the analysis likely provides conservative estimates of the relationship between FP network and contraceptive use because the men without a FP network who were not included in some of the analyses may have been more difficult to engage with the survey and thus had missing data and biased the results towards the null.

Whether men have a FP network distinguishes if couples are using contraception. Men vary in terms of whether they relied on providers or social ties for FP information and advice. Yet all men who had a FP network were significantly more likely to use contraception as a couple compared to those with no FP network. While men are more likely to rely on providers than social ties, these social ties can be just as important in providing support for contraceptive decision-making. There is very limited research examining whether men’s networks influence women’s contraceptive use, even though existing research has shown that men can play an important role in either facilitating or impeding women’s contraceptive use [16‐29]. This study broadens the examination of men’s roles by moving beyond a focus on individual-level characteristics, to examining men’s FP networks and their role in contraceptive decision-making. The study’s findings point to several important ways to foster men’s positive engagement in contraceptive decision-making. First, reaching men through providers continues to be important. There remains a critical need to foster couples’ open communication about contraception. Interventions that engage men’s broader social network, including friends and family members, are an untapped avenue for supporting women’s access to and use of contraception in pursuit of universal access for all women.