Barriers to family planning use in the Eastern Democratic Republic of the Congo: an application of the theory of planned behaviour using a longitudinal survey

Abstract

Objective

In the Democratic Republic of the Congo, there is a low adherence of the population to the use of family planning (FP) due to various social barriers. This study aimed to understand the drives from social barriers to the use of FP in women in the Kivu, a region particularly affected by poverty and many years of conflicts. A theory of planned behaviour (TPB) using a generalised structural equation modelling has been applied to understand the complex sociocultural drivers to the intention and the ultimate decision to use FP.

Design

Longitudinal study.

Setting

A community-based approach was used to investigate FP use in the North and South-Kivu regions.

Participants

Overall, 1812 women 15 years and older were enrolled in the baseline study and 1055 were retrieved during the follow-up.

Primary and secondary outcomes

FP use and intention to use FP.

Results

The mean age was 36±12.9 years, with a minimum of 15 years old and a maximum of 94 years old. Among sexually active participants, more than 40% used a modern contraceptive method at the last sexual intercourse. Education was positively and significantly associated with intention to use FP (β=0.367; p=0.008). Being married was positively and marginally significantly associated with intention to use FP (β=0.524: p=0.050). Subjective norms were negatively and significantly associated with intention to use FP (β=−0.572; p=0.003) while perceived control was positively associated with intention to use FP (β=0.578; p<0.0001). Education and perceived control were positively and significantly associated with the use of FP (respectively, β=0.422, p=0.017; and β=0.374; p=0.017), while Intention to use FP was positively and marginally significantly associated with the use of FP (β=0.583; p=0.052).

Conclusion

TPB helped understand sociocultural barriers to FP use and it can be useful to define adapted strategies in different contexts.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study is relevant to embrace the complexity of sociocultural factors associated with family planning use.

• Large community-based survey with random selection.

• The components used to define attitudes and subjective norms may not be complete.

• Generalised structural equation models design do not allow causal inference.

• Most deprived women could not have been included due to lack of enough resources to integrate the intervention group, but they could have been found in control area.

Introduction

Family planning (FP) refers to policies, services and methods aimed at controlling fertility. The use of FP can help sub-Saharan Africa (SSA) countries to fight poverty and hunger.¹ In African countries including the Democratic Republic of the Congo (DRC), researchers have identified many barriers to adhering to FP, all methods confounded, including financial aspects, religious beliefs or couple dynamics.^{2 3} On the other hand, children are seen as a sign of wealth; the more children a family has, the wealthier it is considered to be.³ Moreover, many authors found that the fear of side effects was a principal barrier to the uptake of FP methods.^3–5

To overcome economic barriers to FP use in DRC, non-governmental organisations (NGOs) are working with the government to promote and provide FP. Social marketing approaches using social networks or mobile apps and media campaigns are widespread to provide information on FP even in remote areas. Many programmes give incentives such as small gifts (soap, etc…) or money for the use of FP or support health facilities so that women and couples can access free FP methods.^{6 7} The assumption often underlying such interventions is that FP methods are unavailable or difficult to access for those who want to use them.^{8 9}

Even though FP methods are now more available and affordable throughout countries in SSA, increasing FP use has remained challenging. Despite many individuals’ intentions to use a FP method, intention is not always followed by an uptake of FP. Qualitative research has highlighted that social factors that may act as barriers to uptake of contraception are seldom considered when implementing FP programmes.^{3 5 10}

The decision to use FP is complex. Therefore, understanding the drivers of behaviour change requires a comprehensive model, such as the theory of planned behaviour (TPB), to understand how knowledge, social norms and beliefs can interfere with women’s intentions. Ajzen developed the TPB in 1991.^11–13 TPB states that attitudes, subjective norms and perceived behaviour control are the three components shaping an individual’s behaviour and intentions. Attitudes are defined by behavioural beliefs that represent an individual’s favourable or unfavourable evaluation of the behaviour. Subjective norms refer to perceived social pressure from society/individual’s community. Perceived behaviour control refers to the ease or difficulty of performing a behaviour in different situations.¹¹

Many authors used the TPB to study health-related behaviours, including the use of condoms and other contraceptives.^14–17 Its efficacy has been demonstrated for years¹² although some authors postulate that the TPB is out-of-date, raising controversy to understand health behaviours.^18–20 The definition of perceived control was among authors’ criticisms although it is an important aspect of health behaviour. However, supporters of the theory state that TPB does not need to be retired, rather, extended. They recommend that the application of the theory must be adapted in every context for better operationalization.^{21 22}

Using TPB as a framework for analysis, this study provides quantitative evidence on the importance of considering the full range of factors, which influence contraceptive behaviour. This analysis brings to light barriers women encounter when attempting to use FP methods, contributing to the development of adapted strategies to meet first-time contraceptive users’ needs.

Methods

Study setting and design

This study is embedded in the impact evaluation of an intervention called Mawe Tatu in North and South-Kivu Provinces, DRC. Mawe Tatu is a gender transformative approach associating women’s empowerment through village savings and loans associations (VSLA) with a ‘men engage’ component to improve household economies and maternal health. The Mawe Tatu project targeted specifically economically vulnerable women in the community. In order to establish associations of sociocultural factors with the use of FP prior to the intervention, this analysis uses the data set of the Mawe Tatu evaluation (baseline and follow-up). A longitudinal cohort study was conducted from May to December 2017, before any sensitisation for gender issues and/or FP had taken place (baseline) and from July 2018 to January 2019 (follow-up), with at least a 1 year interval (exposure) for each participant between the first and last interview. For more details on the Mawe Tatu project, the impact evaluation protocol was published elsewhere.²³

Study population and sampling

Women eligible for the project or for the control group were identified by using community-based targeting, that is, approaching community leaders to provide listings with vulnerable households and invite women from these households to an initial meeting—but participation in the project was voluntary.^{23 24}

The initial sample size calculation was based on the hypothesis that VSLA lower the risk of stunting in children.²³ A two-stage cluster-randomised sampling was used to select first 120 villages in eight health zones (80 in the intervention area and 40 in the control) for inclusion in the study.²³ A second level of random selection was done using the list of all participants in the Mawe Tatu project and from the control groups. A control group consisted of a sample of women in villages with no intervention, that is, economically vulnerable women invited to informative sessions on household economy held by community leaders were considered controls, based on a list of those present. For villages with VSLA, 15 (one VSLA) to 25 (>1 VSLA) women participating in a VSLA were then randomly selected for inclusion in the study. In the control villages, 15 women were randomly chosen from the lists of those present at the information meetings.^{23 24} More details on the selection were provided elsewhere.^{23 24}

Twenty-three thousand women were expected to take part in the Mawe Tatu project. The calculated sample size was 1200 women (respectively 800 in the intervention group and 400 in the control group). Ultimately, 1812 women were enrolled in the baseline study and an individual identification number (ID) was assigned to every participant. To ensure that they would be easily retrieved, addresses, the mobile number of participants when available (or mobile number of someone living in the house) and the names of the head of the household were taken. Out of the initial sample, 1055 were retrieved during the follow-up (12 months interval) using the unique ID for individual and households: 748 women were retrieved in the intervention group (93.5% of the calculated sample size) and 307 participants in the control group (76.7% of the calculated sample). The loss to follow-up was mainly due to security reasons in the North-Kivu. Two villages were insecure and could not be reached during the entire period of the follow-up. The other reasons involved the natural movement of the population looking for better places to survive. Sample size calculation was maximised taking into account a high rate of loss to follow-up described in the region.²³

All women participating in the study aged 15 years and older were included in the present analysis.

Patient and public involvement

The public was involved in the conduct and dissemination plans of the research. Stakeholders and the community guided the researchers in the field and dissemination sessions were planned to obtain the community’s feedback on the findings and ways forward.

Data collection

Survey data were collected via electronic tablets using the Open Data Kit software package. To control bias, additionally to the random selection, a team of local researchers fluent in the locally spoken languages was trained over a week in data collection methods, followed by a pilot study. Data were directly saved on a secure server located at the Swiss Tropical and Public Health Institute in Basel, Switzerland.

Instrument

The impact evaluation questionnaire included Demographic Health Surveys-validated questions about the use of FP and knowledge of different contraceptive methods: ‘Did you ever use a contraceptive?’; ‘Did you use any contraceptive at last sex intercourse?’; ‘Use of modern contraceptive at last sex intercourse?’; ‘Who in a couple should, according to you, decide the use of a contraceptive?’

The content validity of the questionnaire was assessed by Public health experts from the Swiss Tropical and Public Health Institute and Catholic University of Bukavu (Content validity index=0.95). Test–retest reliability of the questionnaire was obtained by administering the questionnaire two times to the same 30 households after 2 weeks interval in a semi-urban rural health zone, Casha.

To control sociodemographic factors that may affect the use of FP, we also recorded participant’s age, education level, religion (catholic or other), provenance (urban, semirural or rural) and marital status (married/in couple or single).

Data analysis

Descriptive statistics were calculated for sociodemographic variables, and FP use-related questions (uptake of contraceptives, decision-making on the use within the couple).

FP pathways have been studied in the past using many types of regression analyses. However, to address the complexity of FP uptake, especially in the study region, the use of generalised structural equation modelling (GSEM) appeared appropriate when building the model of the TPB. GSEM is a multivariate statistical analysis technique used to analyse structural relationships between variables and latent constructs. GSEM is a combination of factor analysis and multiple regression analysis using endogenous and exogenous variables. Structural equation modelling has been used previously to determine the utilisation of condoms or other contraceptive methods.^{25 26} Data with missing values were discarded. Analyses were performed using Stata V.15.0.

Model specification

To build the model of planned behaviour, qualitative findings from previous researches were used^{3 27 28} (figure 1)

Figure 1.

TPB model built from existing qualitative findings. FP, family planning; VSLA, village savings and loans associations; TPB, theory of planned behaviour.

This TPB model analyses pathways linking personal beliefs (attitudes, subjective norms in the society and perceived control), intention to use or not a FP method and the uptake of the behaviour (use of a FP method) (figure 1).

To identify and confirm the underlying structure of latent variables in the GSEM, different factors were tested using factor exploratory analysis. Factor exploratory analysis identified the number of latent constructs and the underlying factor structure of the TPB attitudes, subjective norms and perceived control. The confirmatory factor analysis (CFA) examined the appropriateness and generalisability of the measurement portion of the GSEM. Factor loadings >0.4 were retained.

Two ‘yes/no’ items reflecting knowledge and behaviour beliefs on contraceptives measured ‘Attitudes’: (1) ‘Contraceptive methods obtained at a health facility, are of good quality’ and (2) ‘I trust in what health professionals in this community say about contraceptives’.

Three ‘yes /no’ items constructed ‘subjective norms’: (1) ‘Children are a gift of God. Contraceptives should not be used’; (2) ‘The use of contraceptives is a sin’ and (3) ‘The use of contraceptives encourages infidelity’.

To assess ‘the perceived control on behaviour’, two ‘yes/no’ items were considered: (1) ‘If my partner and I want to have sex and I want to protect myself I can always convince him/her to use protection’; and (2) ‘If your partner wants to have sexual relations with you, do you think you can convince him/her to use a contraceptive method?’

Primary and secondary outcomes

The ‘intention to use a contraceptive method’ was assessed by the question: ‘Even if you have never used a method, we would like to learn about your preferences: which contraceptive method do you or would you prefer?’ Intention to use any FP was coded ‘1’ if the woman cited one modern contraceptive method and ‘0’ otherwise. In the present analysis, we consider baseline values to define the intention to use.

The behaviour, the ‘use of a FP method’, was defined by the use of a modern FP method during the last 12 months. For the final analysis, the outcome included all new users of modern FP at follow-up, excluding those who were already using it at baseline.

The coefficient of determination of 0.951 and the Standardised Root Mean Square Residual of 0.074 (less than 0.08) indicated that the model fit our data.²⁹ The potential confounding effects of the intervention (being member of VSLA) and sociodemographic characteristics such as age, religion, provenance (rural or urban), education and marital status were controlled. Only education and perceived self-control were significantly associated (at level of 0.10) with the use of FP in a binary association controlling for intention, therefore, we kept these associations in the final model.

Results

Descriptive results

Table 1 indicates the sociodemographic characteristics of the participants in the study.

Table 1.

Sociodemographic characteristics

Variables	Baseline		Follow-up
	Frequency Mean/SD	%	Frequency Mean/SD	%
Age (years)	36.0±12.9		37.8±12.5
Education	n=1806		n=1055
None	345	19.1	217	20.6
Primary	599	33.2	316	30.0
Secondary	765	42.3	469	44.6
University	97	5.4	51	4.8
Marita status	n=1740		n=1055
Single	420	23.2	218	20.7
Married/couple	1320	76.8	837	79.3
Number of children	5±3	5±3	6±3
Religion	n=1806		n=1055
Catholic	959	53.1	606	57.4
None catholic	847	46.9	449	42.6
Urban	n=1471		n=1055
Urban	317	21.6	234	26.4
Semiurban	940	63.9	555	62.6
Rural	214	14.5	97	11.0
Intervention	n=1806		n=1055
VSLA	1225	67.6	748	70.9
Control	587	32.4	307	29.1

VSLA, village savings and loans associations.

The mean age was 36±12.9 years, with a minimum of 15 years old and a maximum of 94 years old at the beginning of the study. Over 76% of the study participants were married (table 1).

Table 2 details participants’ uses of contraceptive methods. The results are presented separately for single women and married women/living together with their partners.

Table 2.

Use of contraceptives

Variables	Baseline		Follow-up
	Frequency	%	Frequency	%
Did you ever use a contraceptive?	n=1709		n=1015
Yes	583	34.1	365	36.0
No	1126	65.9	650	64.0
Use of contraceptive at last sexual intercourse	n=567	%	n=359	%
Yes	365	64.4	204	56.8
No	202	35.6	155	43.2
Use of modern contraceptive at last sexual intercourse	n=567	%	n=359	%
Yes	243	42.9	130	36.2
No	324	57.1	229	63.8
Who in a couple should, according to you, decide the use of a contraceptive?	n=1673	%	n=997	%
Myself	121	7.2	55	5.5
My partner	299	17.9	150	15.1
The couple together	1253	74.9	792	79.4
Who decides within your couple the use of a contraceptive?	n=573	%	n=357	%
Myself	117	20.4	82	23.0
My partner	85	14.8	55	15.4
The couple together	371	64.8	220	61.6
FP last 12 months	n=583	%	n=365	%
Yes	171	29.3	124	34.0
No	412	70.7	241	66.0
FP new users	–	–	72	19.7

FP, family planning.

Approximately 76.9% of the participants in the study thought that the couple must decide together whether to use a contraceptive, while only 70.6% thought the same (table 2). Among participants who were sexually active, more than 40% used a modern contraceptive method at the last sexual intercourse (table 2). At the baseline, 29.3% used modern FP during the past 12 months, whereas 34.0% used FP during the follow-up; 19.7% were defined as new users during the endline. Approximately, 57.1% of the participants at baseline thought that the couple must decide together whether to use a contraceptive compared with 63.8% during the follow-up (table 2).

Factor analysis results

Three factors (eigenvalues>1.0) emerged as results in the factor analysis (table 3): attitudes (two indicators), subjective norms (three indicators) and perceived control (two indicators).

Table 3.

Factor analysis for indicators of the TPB

Latent construct	Items included in the survey	Factor loadings EFA	Determinant of the correlation
Attitudes	Contraceptive methods, which are obtained at a health facility, are of good quality.	0.539	0.838
	I trust in what health professionals in this community say about contraceptives.	0.500
Subjective norms	Children are a gift of God. Contraceptives should not be used.	0.574	0.767
	The use of contraceptives is a sin.	0.488
	The use of contraceptives encourages infidelity.	0.445
Perceived control	If my partner and I want to have sex and I want to protect myself, I can always convince him/her to use protection.	0.694	0.638
	If your partner wants to have sexual relations with you, do you think you can convince him/her to use a contraceptive?	0.695

All factor loadings were significant at p<0.05 (table 3).

EFA, exploratory factor analysis; TPB, theory of planned behaviour.

SEM results

The final adjusted SEM results are presented in figure 2 and table 4. The standard path coefficients and p-values in the standardised metric are reported (figure 2 and table 4).

Figure 2.

GSEM results explaining the TPB. GSEM, generalised structural equation modelling; TPB, theory of planned behaviour.

Table 4.

Standardised path coefficients of the latent variables GSEM

Variables
Intention	β (95% CI)	P value
Endogenous variables
Attitudes	−0.136 (−0.573 to 0.299)	0.539
Subjective norms	−0.572 (−1.013 to 0.131)	0.003
Perceived self-control	0.578 (0.293 to 0.862)	<0.0001
Exogenous variables
Age	−0.006 (−0.024 to 0.011)	0.474
Marital status	0.524 (0.0007 to 1.048)	0.050
Religion	−0.029 (−0.462 to 0.403)	0.893
Urban	−0.295 (−0.660 to 0.069)	0.113
Education	0.367 (0.097 to 0.636)	0.008
Intervention	0.563 (0.108 to 1.019)	0.015
Contraception use
Education	0.422 (0.076 to 0.767)	0.017
Intervention	−0.025 (−0.771 to 0.32)	0.419
Perceived self-control	0.374 (0.067 to 0.682)	0.017
Intention	0.583 (−0.005 to 1.171)	0.052

Statisticaly significant results (p < 0.05) are presented in bold

GSEM, generalised structural equation modelling.

Education (high level) was positively and significantly associated with intention to use FP (β=0.367; p=0.008). Being married was positively and marginally significantly associated with intention to use FP (β=0.524; p=0.050) (table 4). Among latent constructs, subjective norms was negatively and significantly associated with intention to use FP (β=−0.572; p=0.003), while perceived control was positively associated with intention to use FP (β=0.578; p<0.0001). Education and perceived control were positively and significantly associated with the use of FP (respectively, β=0.422, p=0.017; and β=0.374; p=0.017), while Intention to use FP was positively and marginally significantly associated with the use of FP (β=0.58; p=0.050) (table 4).

Discussion

The percentage of those who used a contraceptive method in our sample population was low (36% during the follow-up) and is similar to other study findings elsewhere in DRC.^{30 31} In the last decade, many NGOs made contraceptive methods free or available at a lower cost in the Kivu region, which can explain why the prevalence of contraceptive use is higher compared with other regions.^32–34

Education was significantly associated with the use of contraceptives, corroborating findings from existing literature that higher education enhances women’s ability to make reproductive choices including contraceptive use.^{32 35–37} Education ‘corrupts’ peoples’ normative values and ways of seeing the world, which can include changing women’s perspective on fertility. Fertility preference becomes a ‘calculus of conscious choice’^{38 39} rather than a predetermined destiny. In addition, schooling changes the objective conditions under which women and couples’ decisions are made such as by increasing the opportunity cost of women’s labour.^{40 41} To enhance their career, educated women are more likely to use FP because they can decide how many children they want and when to have them.

The model constructed fits our data well and significantly predicted contraceptive use. In a given society, as Ajzen¹¹ postulated, behaviour is shaped by the interaction between attitudes, subjective norms and perceived behaviour control.¹¹ These factors lead to intention before adoption—or not—of the designed behaviour. Authors found that misconceptions and misinformation were associated with lower contraception use.^{42 43} However, in our survey, (mis)trust of health providers or knowing where to get information (defining attitudes) was not statistically associated with the use of FP. This reflects that an individual’s knowledge and favourable evaluation of FP might not necessarily lead to its use. Some other factors rooted in social life may be more decisive.

In our study context, subjective norms are first rooted in religion as contraceptive use is considered a sin. People choose not to use a contraceptive to avoid feeling guilty, judgement by society or punishment by God. Many other studies have also found that religious beliefs were barriers to contraceptive use.5 31 44 45 The literature describes religious barriers in terms of blanket prohibition and many religious communities are against the use of FP in the study region.^{45 46}

We also found, like other studies in comparable contexts, that children are considered as ‘gifts from God’, and therefore it is wrong to use contraceptives to avoid pregnancy.^{10 27}

Moreover, our findings suggest that power dynamics within the couple are a strong predictor of contraceptive use. While FP use ideally requires the implication of both partners, the male partner is more often the key decision-maker.⁴⁷ Existing literature states that partner’s opposition is a limiting factor for women to use FP.28 33 48 49 In these male-dominant societies, partner’s opposition can result from many reasons like personal beliefs against FP use or society’s pressure.

Across low-income and middle-income contexts, the communities perceive there is no need to use FP within marriage. Furthermore, the conviction that someone using contraceptive methods is susceptible to being unfaithful is another limiting factor to use FP.²⁸ When a married woman uses contraceptive methods (mainly without her partner’s knowledge) that implies she has extraconjugal affairs and does not want to be caught with an unexpected pregnancy.^{28 50} However, previously, researchers highlighted that some women are ready to use contraceptive methods without their partner’s knowledge²⁷ to protect themselves from approaching births and consequences.

The third element of TPB, perceived behaviour control, was illustrated in our study as perceived control over the decision to use any FP method. Women with enough confidence and capacity within their partnership to negotiate the use of contraceptives are most likely to use a contraceptive method, corroborating other findings.^{51 52} Through education and financial independence, women acquire negotiation skills empowering them to raise and fight for their own opinions within the couple.³⁷

The intervention was not statistically associated with an increased in the use of FP. The results also demonstrated that before any intervention, women who joined the VSLA were already more likely to have an intention to use FP. The short-term interval (1 year between the baseline and the follow-up) might explain why we did not see any change in the use of FP despite the expressed intention. However, it is not excluded that subjective norms could have prevented them from taking action. A long-term evaluation could help understand better all the interactions.

Many FP interventions focus on health system-based solutions such as information campaigns and increasing availability of FP methods. This study contributes to the growing literature base and shows that health system-related factors do not seem to play a dominant role in contraceptive uptake. Rather, subjective norms such as religion and couple power dynamics, namely women’s low bargaining power, remain crucial considerations to increase FP uptake.

Limitations

The selection of the two groups (intervention and control) was done randomly assuring the representability of the population of selected areas. However, due to the self-selection in VSLAs groups, it is not excluded that some women could have been marginalised either because they were too poor to afford the weekly contribution in VSLA or because of a poor social network. These analyses cannot necessarily be generalised to all the population given the self-selection of participants in the groups.

Another limitation is that the definition of certain concepts like subjective norms could not have been exhaustive. We only considered the measurements reported in an existing questionnaire, which was broader and covering many aspects. However, others have used the same scales or items to understand the use of FP in similar settings. Finally, some control variables such as employment status were not taken in account because they did not fit in the GSEM. However, before discarding them, we ran univariate and bivariate regressions of employment status over intermediary and dependent variables (intention and/or use of FP) and the associations were not statistically significant, which reinforced our decision of not considering them.

Policy implications

To circumvent barriers to FP use, programmes should:

• Promote women’s education to empower women by raising their self-efficacy to negotiate the use of contraceptives when needed.

• Consider addressing gender norms and other cultural aspects with religious leaders and other stakeholders.

• Fight against subjective norms by providing accurate information on FP use to the two partners and the community.

Conclusion

A TPB, using a GSEM, allowed us to demonstrate that the use of contraceptives is deeply related to one’s perception of contraceptive use and many other sociocultural constructs. Education, subjective norms (children are gifts from God, the use of contraceptive is sin, contraceptive use encourages infidelity), and perceived control (self-efficacy to negotiate the use of contraceptives) influence a woman’s intention to use FP. Programmes promoting FP use must take into account the empowerment of women to give them accurate information and involve the male partner in sensitisation sessions.

Data availability statement

Data are available upon reasonable request. Data set generated and analysed are not publicly available due to confidentiality and anonymity of study population. Data sets are stored on a secure Alfresco website and are available from the corresponding author on reasonable request.

Ethics statements

Patient consent for publication

Not required.

Ethics approval

This study involves human participants and was approved by The EKNZ Ethical committee (Ethik Kommission Nord-und Zentralschweiz No EKNZ BASEC UBE 2016-01878) in Switzerland and the Ethics Committee of the Catholic University of Bukavu (No UCB/CIE/NC/019/2016) approved the main research proposal for the impact evaluation of the 'Mawe Tatu' intervention in 2016. Informed verbal and written consent was obtained from each individual before the beginning of the survey. Confidentiality was guaranteed. For participants below 18 years old, written consent from the parent or legal tutor were obtained. Participants gave informed consent to participate in the study before taking part.