As Long as (I Think) My Husband Agrees…: Role of Perceived Partner Approval in Contraceptive Use Among Military Couples Living in Kinshasa, DRC


 Background: Male partner’s approval is a key determinant of contraceptive use for women living in Sub-Saharan Africa and improving men’s support and couple communication is a cornerstone of family planning programs. However, approval is often only measured through the women’s perception of their partner’s opinion.Methods: This study conducted in Kinshasa compares contraceptive approval variables from matched male and female partners to establish the frequency of (in)accurate perceptions by the woman, then test their association with modern contraceptive use. Additional regressions estimate individual and couple variables associated with (in)correct perceptions. Results: Results confirm women are poorly aware of their partner’s opinion but indicate that perceived approval or disapproval by the woman is a much stronger determinant of modern contraceptive use than her partner’s actual opinion. Higher educational achievement from the woman is the strongest driver of misunderstanding her partner’s approval. Conclusions: Women’s perceptions of partner’s approval are much stronger determinant of contraceptive use than the latter’s actual opinion, and stereotyping men’s opinion of family planning is a common error of appreciation. However, findings also suggest these misunderstandings might serve women’s capacity to negotiate contraceptive use.


Background
Men's involvement in supporting contraceptive decision and contraceptive use has been a leitmotiv of family planning programs since the 1994 ICPD [1], with strategies ranging from changing men's fertility aspirations and attitude towards contraceptive use [2], increasing women's agency in fertility related decision-making [3] and improving couple communication [4,5,6]. In Sub-Saharan Africa, the assumptions that men hold "traditional" desires for large families and that they play a dominant role in decision-making in the household positions them as key players in interventions designed to improve women's contraceptive use [7]. Indeed, research consistently nds that male approval is associated with increased contraceptive use independently of other factors such as women's sense of self-e cacy or perceived contraceptive accessibility [8,9,10,11,12]. Furthermore, the effect appears to be bi-directional since men's disapproval stand in some studies as the single more powerful deterrent towards use of family planning. [13,14,15] The key role of male partner's approval for contraceptive use is however complicated because existing evidence also strongly suggest that women tend to be inaccurate in the perception of their partner's actual opinion regarding their use of family planning [16]. The low validity of wives' report of husband's approval of family planning has been attributed to lack of intra-couple communication [17] and a tendency to stereotype their partner's fertility desire ("All men want large families") [18]. On the male partners' side, the widely held belief that men should have the nal say in family matters, but women are nonetheless solely responsible for managing their pregnancies [19], further deepens the possibilities for misunderstandings. Existing studies comparing men's and women's fertility desire indicate that most of these misunderstanding arise because respondents typically underestimate their partner's desire to limit the family size [15,18], or because women con ate their partner's desire for more children with their opposition to contraceptive use [20]. As a result, a trend towards "false negative" scenarios, where women perceive disapproval from their partners, but surveys among men suggest that a majority of them are supportive of contraceptive use has been noted, including in high fertility contexts such as Nigeria and Ethiopia [21,22] Few studies are available to compare female perception of approval and men's actual opinion of contraceptive use, and the sole availability of perceived partner preference rather than true partner preference as an explanatory variable if often cited as a limitation [12,17,23]. The research presented here provides insights on the aforementioned misunderstandings regarding contraceptive use by comparing matched male and female partners' independent answers to the same questions. The couples were all recruited within military camps dispersed throughout Kinshasa, the capital city of the Democratic Republic of Congo (DRC). The DRC, one of the poorest countries in the world, also has one of the world's highest fertility rates (6.0) and lowest modern contraceptive prevalence (14.2% on average in the country and 24.5% in Kinshasa) [24,25]. As an FP2020 focus country, the DRC and particularly the provinces of Kinshasa and Kongo Central, are regularly targeted by population surveys, and in 2016 the military authorities requested a speci c iteration of the PMA2020 (Performance in Monitoring and Accountability) survey to be implemented in military camps throughout Kinshasa to assess whether those populations had speci c need, compared to the general population living in surrounding neighborhoods. The smaller scope of the military study allowed for researchers to add speci c modules on topics which were of interest across the DRC general population but could typically not be integrated into the global PMA survey.
Results from the rst round of survey indicated among other ndings that women residing in military camps had high unmet need for family planning services [26]. The data presented in this article comes from the second round of PMA survey in military camps, implemented in January 2020. This round intended to capture contraceptive trends after four years of various interventions lead by family planning partners in military camps, but also included matching questions about fertility desires and approval of contraceptive use that were independently asked from male and female partners in established couples. The objective of the analyses presented here are rst to estimate the prevalence of correct perception, false positive and false negative perceptions of men's approval of contraceptive use by their female partner, and to compare the relative importance of actual and perceived approval with regards to women's contraceptive use. We then seek to determine individual and couple characteristics associated with correct and incorrect perception of male partner's approval by the woman, with a speci c attention to false negative and false positive scenarios.

Data
Data collection within the population of military camps used a multi-stage cluster design approach. There are 17 military camps in Kinshasa, which are home to an estimated 3% (305,000) of the total population of the city [26]. Households in these camps tend to have at least one person in active service duty, however extended families with no direct army a liation can also reside in the camps. All residents however have access to military health facilities within the camp, including family planning services offered by the Military Program for Reproductive Health (PMSR).
Out of the 17 military camps in Kinshasa, 10 were randomly drawn, proportionate to population size. These 10 military camps were then divided into enumeration areas (EAs) of similar size. One EA then was randomly selected in each of the 10 camps. Resident enumerators (REs) listed all the households in the selected EA. In each selected EA 33 households were randomly selected. Since we were interested in interviewing couples, we only interviewed women who were married or in union and currently living with their partners in those households. Women who were divorced, widowed or have never been married (about 2% of the total sample) were excluded from our analysis. Since the sample size was calculated based on the women, women in selected household were interviewed rst, followed by their male partners. Although some men indicated being in a polygamous relationship, no two women in our sample were linked to the same partner.

Analysis
This study focuses on two sets of outcome variables (1) correct or incorrect perceptions of male partner's approval of contraceptive use among female respondents, and (2) modern contraceptive use among female respondents. Perception of male partner's approval was de ned by the woman's response to the question "Do you believe your partner approves / would approve of your use of modern contraceptives?" Possible answers included "yes", "no", "depends on the method", "do not know", for which we categorized "yes" as "perceived approval", while all other were grouped as "perceived disapproval". The same binary categories were applied to the male respondents' responses to the question: "would you approve if your partner wanted to use contraception?" (Interviewers did not disclose to the man whether his female partner had indicated using contraception). Any other response than "yes" was recorded as disapproval. Comparing women's perceptions and men's actual opinions led to four main scenarios (see Table 2): Women correctly perceive approval (S1); Women correctly perceives disapproval (S2), women incorrectly perceive disapproval (false negative) (S3), and women incorrectly perceive approval (false positive) (S4), combinations of which led to "correct perception of partner's opinion" (S1 + S2 = S5) and "incorrect perception" (S3 + S4 = S6).
We rst conducted bivariate analysis to test the association of those scenarios with modern contraceptive use by the couple; then to test the association of modern contraceptive use and mentioned scenarios after controlling for other in uential independent variables, we included individual and couples' characteristics in a multivariable regression model. Finally, we conducted another set of regression analysis to determine whether certain individual or couple characteristics were signi cantly associated with perception scenarios we found likely to increase or decrease contraceptive use, including incorrect perception of partner's opinion, perceived approval and false negative perception. Independent variables included in all regressions were factors assumed to in uence power relations within the couple and contraceptive use, such as woman's age, individual education and educational difference within couple, number of living children, whether each partner wants more children, whether each partner has discussed his or her ideal number of children, the man's ideal family size and his military status (active military or not).

Results
The study sample included 252 couples, married or living in union. As shown in Table 1, women were on average younger and slightly less educated than men. About half of the men and only 2.9% of the women were active military members. Most individuals had completed at least secondary education (83.2% of women and 90.3 % of men) and educational achievements were similar within 57.0% of couples, whereas a quarter of men (25.9%) were more educated and 17.1% of women were more educated than their partner (See Table 1 below). Observed discrepancies in marital status, with men disproportionately reporting being married and with a slightly higher number of living children most likely come from the fact that their live-in partner in the camp is not always their o cial or primary spouse. Regardless, all family planning questions were asked in relation to said live-in partner, who was also the person interviewed as the other half of the couple. In Table 2, we compare male partner's approval of contraceptive use with women's perception of their partner's approval. When asked separately, 59.8% of men stated they disapproved of their female partners using contraception, but a larger proportion of women (73.4%) perceived disapproval. When women were mistaken in their perception of their partner's approval, it was more likely that thy perceived that he disapproved when he in fact approved, instead of vice-versa: a quarter (24.9%) of women erroneously thought their partner disapproved of contraceptive use (false negative). On the other hand, 11.3% of women stated their partner approved of contraceptive use, when he in fact did not (false positive). Overall, the woman's perception of their partner (dis)approval was correct 63.8% of the time. (See Table 2 below) We then looked at the association between the scenarios in Table 1 (i.e., the combination of partner's actual opinion and woman's perception of her partner's approval or disapproval) and modern contraceptive use by the couple ( Table 3). As would be expected, modern contraceptive use is highest (41.3%) when the man approves of it and his female partner correctly perceives his approval and is lowest (19.4%) when the woman correctly perceives her partner's disapproval (p-value approaches signi cance at 0.05).
Our results however indicate that modern contraceptive use is similarly high and low regardless of the man's actual opinion, provided that the female partner perceives it as approval or disapproval, and the associations are much more signi cant. Modern contraceptive use is 41.3% (pvalue < 0.05) for women who believe their partner approve of it, and 19.5% (p-value < 0.05) for those who believe their partner disapproves. In fact, 38.7% of women who incorrectly perceive their partner's approval (false positive) use modern contraception (p < 0.1), only 2.6 percentage points less than women who correctly believe their partner's support their use of a modern method. At the other end of the spectrum, women who incorrectly perceive their partner's disapproval (false negative) are among the lowest users of modern contraception, although the relationship is not signi cant. To further test the association of approval/perception scenarios with modern contraceptive use after controlling for socio-demographic characteristics of men and women, we performed a multivariate logistic regression on modern contraceptive use. Results from this analysis con rm the much more signi cant role of the woman's perception of her partner's approval as a determinant of contraceptive use. When their partner disapproved of contraceptive use, women who incorrectly perceived approval had 2.52 times greater odds of using a modern method than women who correctly read their partner's opinion, and the relationship was highly signi cant (p-value = 0.007). Conversely women who erroneously perceived their partner's disapproval (false negative) appeared slightly less likely to use modern contraception, but the relationship was not signi cant (OR = 0.92, p-value = 0.805).
Among other individual characteristics of women and men, few were signi cantly associated with modern contraceptive use: odds ratio for contraceptive use declined with age and women over 40 were signi cantly less likely to use a modern method (OR = 0.21, p-value < 0.05).
Women whose partner was in active military were also signi cantly less likely to use a modern method (OR = 0.44, p-value < 0.1). On the other hand, women who had four or more living children were three times more likely to use a modern method (OR = 3.01, p-value < 0.05). Differences in educational status, fertility desire variables such as either partner wanting more children and ideal family size or having discussed the desired number of children were not signi cantly associated with modern contraceptive use. Since perception of their partner's approval is so signi cantly associated with contraceptive use by the women, it seems important to determine whether speci c individual and couple pro les were associated with (in)correct perceptions, as well as false positive and false negative perceptions. To achieve that goal, we introduced three logistic regression models (see Table 5) testing the associations between women and couples' characteristics and three speci c scenarios. We chose the scenarios based on the strong association found in the previously shown model ( The results of this regression indicate that among all characteristics included in the model, number of living children was strongly associated with perception of approval for contraceptive use: women with 2 to 4 children and women with more than 4 children were twice to ve times more likely to declare that their partner approved of their contraceptive use (OR = 2.81 and 6.62 respectively, p-value < 0.05). The same was true of women who stated that they wanted more children (OR = 2.13, p-value < 0.05) On the other hand, educational differences were negatively associated with perceiving one's partner's approval: women who were more educated than their husband signi cantly less likely to perceive approval (OR = 0.35, p-value < 0.05). This last variable was also signi cantly associated with incorrect perception of their partner approval (OR = 1.9 p-value < 0.05) and even more so with incorrect perception of his disapproval: women who were more educated than their partner were almost three times more likely to fall in a "false negative" scenario (OR = 2.84, p-value < 0.01) Having discussed the number of desired children made women less likely to hold an incorrect perception of their partner's approval (OR = 0.69, p-value < 0.1), and the likelihood of a "false negative" decreased signi cantly when the man had discussed his desired number of children (OR = 0.27, p-value < 0.01) Difference between man and woman ideal family size, ever or current use of modern contraception were initially added to the model but did not yield any meaningful results.
Since differences in educational attainment and having discussed desired number of children were strongly associated with (in-)correct perception of partner's approval of contraceptive use, we ran a separate analysis to assess whether educational level was associated with increased communication about fertility desires. (Table 6 below) While higher level of individual education for either partner was associated with the likelihood of having discussed the desired number of children (p-value < 0.1), educational difference did not have a signi cant bearing on this outcome. Thus, women who were more educated than their partner were not more or less likely to be more informed about his actual fertility desires.

Discussion
Findings from this research among military couples residing in Kinshasa con rm that women tend to be poorly aware of partners' opinion regarding contraceptive use: almost half of the female respondents in our sample were incorrect in their perception, with the majority of them incorrectly perceiving their partner's disapproval (false negative), while about one in ten women was a false positive (erroneously perceiving approval from her partner). The false negative scenario, which concerned 24.9% of our sample, may be a product of stereotyping in the absence of spousal discussion regarding family planning, as observed in other high-fertility, low contraceptive use environments, where women tend to project on their partner the assumed attitude of most men living in those environments [11,21], These errors in assessing men's opinion regarding contraception are concerning because our results indicate that, regardless of men's actual approval of family planning use, it is their partner's perception of that opinion that acts as a determinant of (non)-use, thus con rming ndings from previous research conducted in Sub-Saharan Africa [11,28]. When introducing individual and couple variables in the models, only a few expected characteristics remained as independent predictors of contraceptive use (number of living children) or non-use (women older than 40, male partner in the army), but perceived approval, especially in the case of false positive, remained a strong independent predictor of contraceptive use. Therefore, with nearly a quarter of all female participants holding a "false negative" perception, improving couple communication on the subject might increase contraceptive use if the male partner's approval was more clearly shared.
Further analysis looking at variables associated with incorrect perception, perception of approval and "false negative" perception of partner's opinion con rmed the potential importance of communication: women were four times more likely to perceive approval and four times less likely to hold a false negative perception if their partner had discussed the number of children they wanted. However, the relationship did not apply if only the woman had expressed her fertility desires, which con rms the weight of men's preferences in fertility decisions [29].
Interestingly, women who had a large number of living children and those who stated that they wanted more children were also more likely to perceive approval from their partner to use contraception. This suggests that women who have ful lled their or their partner's fertility expectations, or at least who intend to do so, may feel they have earned the right to rest and thus assume they have their partner's approval to use contraception for spacing their next pregnancies.
Aside from individual and joint fertility desires and communication, the fact that the likelihood of perceiving approval for family planning increased when men were more educated than their partner compounds existing evidence that increase in male education plays an important role in improving spousal communication and favorable attitudes towards family planning [30,31]. On the other hand, women education, and particularly differences in education with their partner, played in a somewhat counter-intuitive ways in our results. While more educated women were more likely to have discussed their desired number of children, women who were more educated than their partners were also signi cantly more likely to be incorrect about his approval of contraceptive use and this error leaned heavily towards falsely perceiving disapproval. This could be because women with partners less educated than themselves are more prone to stereotyping their partner's attitudes. But another possibility lies in more educated women knowingly misrepresenting their partner's opposition to justify not using contraceptives. Whether they personally have high fertility desires, fear modern contraceptive side-effects, or hold other adverse beliefs towards contraception, more educated women are more likely to understand that partner's opposition is a more acceptable response, because it shifts the responsibility away from them, when asked about their reasons for non-use. Research conducted in Sub-Saharan Africa consistently indicates that women tend to have more conservative attitudes than men [28], and that the classic inverse relationship between women's education and lower demand for children is not as strongly marked as in other developing contexts [31].
This hypothesis further suggests that there might be some usefulness to the noted discrepancies in awareness of partner's opinion regarding considering cultural parameters rst [32].
This study of discordance in actual and perceived approval of contraceptive use within couples thus teases an important issue for family planning programs as they strive to improve male engagement while maintaining women's agency and rights-based family planning choices: whether women are honestly or purposefully mistaken about their partner's approval of contraceptive use, these misunderstandings might on occasion better serve the woman's particular contraceptive agenda than would perfect communication of family planning decisions in the couple.
Our study has several limitations however, particularly when it comes to desirability bias. Male respondents in particular may have been more likely to declare that they approved of their partner's contraceptive use because they knew it was the "correct" answer. In addition, we have used "ever discussed the desired number of children" has a proxy for communication about family planning but it is unclear whether agreement on fertility desires accurately translates into agreement regarding contraceptive use within the couple. Besides, our survey instrument did not permit us to establish the frequency, length and quality of these conversations, all variables we would expect to have an impact on the woman's perception of her partner's opinion. We also chose modern contraceptive use, rather than intention to use, as a more reliably reported key outcome but this choice relies on the assumption that women living in military camps all face similar barriers and opportunities for accessing contraceptive services. Finally, future research would need to explore whether the military environment of the study affected our results, particularly as higher unmet need for family planning and more unequal gender relations in this context than in the general population have already been recorded.