Menu
Menu
Menu
Menu
Background: In the eastern part of the Democratic Republic of Congo (DRC) Village Savings and Loan Associations (VSLAs) programs targeting women are implemented. In the context of the ‘Mawe Tatu’ program more equitable intra-household decision-making is stipulated by accompanying women’s participation in VSLAs with efforts to engage men for more gender equality, expecting a positive effect of this combined intervention on the household economy, on child nutritional status, on the use of reproductive health services including family planning, and on reducing sexual and gender-based violence (SGBV). Methods: A longitudinal parallel mixed method study is conducted among women participating in VSLAs in randomly selected project areas and among a control group matched for socioeconomic characteristics. Descriptive statistics will be calculated and differences between intervention and control groups will be assessed by Chi2 tests for different degrees of freedom for categorical data or by t-tests for continuous data. Structural equation modelling (SEM) will be conducted to investigate the complex and multidimensional pathways that will affect household economic status, child nutritional status and use of reproductive health services. Analysis will be conducted with STATA V.15. Concomitantly, qualitative data collection will shed light on the intra-household processes related to gender power-relations that may be linked to women’s participation in economic activities and may lead to improvements of maternal and child health. Focus group discussions and in-depth interviews will be conducted. All narrative data will be coded (open coding) with the help of qualitative data analysis software (Atlas TI). Discussion: Women’s empowerment has long been identified as being able to bring about progress in various areas, including health. It has been shown that men’s commitment to transforming gender norms is a sinequanone factor for greater equity and better health, especially in terms of reproductive health and child nutrition. This study is one of the first in this genre in DRC and results will serve as a guide for policies aimed at improving the involvement of men in changing attitudes towards gender norms for higher household productivity and better health.
This research will employ a parallel mixed-method design, combining a longitudinal cohort study (Study1) with a longitudinally designed qualitative study (Study 2). Quantitative findings will be triangulated with qualitative findings in order to deepen the understanding of the forces that trigger and sustain the expected change. A cluster-randomized, longitudinal intervention study compares VSLA participants in an intervention area with controls over a period of 12 months. The intervention districts are identified a priori by the implementing organization, but the intervention and control sites (villages) are randomly selected within these districts. In the intervention sites, all persons participating in a newly created VSLA are eligible for inclusion in the intervention arm. While participation in a VSLA is based on self-selection, a random sample of VSLA participants is selected based on VSLA members’ lists available from the Mawe Tatu project. A control group of participants is recruited in adjacent randomly selected villages where VSLAs are not offered. Participants self-select to take part in an information session on income-generation or a related theme in order to recruit participants with a similar socio-economic profile as the VSLA members. A random sample of participants of the information sessions is then included in the study as controls. In both the intervention and control areas, a community leader conveys the information about the upcoming activities in a similar way in order to attract a similar group of participants. Additional inclusion criteria for the study included being long-term residents of the study site (living in the household for at least 6 months) and being at least 15 years old. For participants with children, all children aged 1–5 years currently living in the household of an adult study participant are recruited for inclusion in the anthropometric study module. Children included in the study must be under guardianship of the adult study participant (Table (Table11). Study population by intervention and control group, quantitative study component The survey questionnaire includes questions about primary outcomes: household economy (income-generation, income, assets, housing, relative household status, health insurance), child nutritional status; and unmet need of family planning. Secondary outcome variables encompass gender norms and rights (perception of women’s rights and gender equity, women’s participation in decision-making and income-generation, women’s utilization of reproductive health services, and women’s perceived self-efficacy to speak out in community meetings). Further questions include information about the household structure (household composition and headship),, and individual socio-demographic information (age, education, marital status, number of children), and program-related variables (participant of VSLA, time in VSLA, partner participation in men’s reflection groups) (Table 2). A locally adapted composite wealth score is calculated based on structure of the house, type of fuel for cooking, toilet facility, food security, and having a mobile phone or TV. Questions from previously validated instruments are used if available (Demographic and Health Survey [7], Food-insecurity experience scale [50], Gender-equitable Men scale [51]). Anthropometric data is collected from children under 5 years of age living in the household including weight, height and mid-upper arm circumference (MUAC). An overview over the indicators used is provided in Table Table22. Study variables The power calculation is based on the hypothesis that establishing savings and loan systems at the village level lowers the risk of stunting of children up to the follow-up. Improved child growth is a result of increased livelihood and food security that is sustainable over some time. The surveys are conducted in 80 villages with an average of 15 households per village (1200 households). Assuming an attrition rate of 30% during follow-up this results in a final analytic sample size of 800. The control group is planned to be smaller with 40 villages (600 households). Assuming the same attrition rate there must be 400 households recruited to participate in the final control arm. As there are few data on the distribution of individual growth rates among children in the study area, we express the intervention effect in terms of a certain fraction of the standard deviation SD of individual growth rates. If the mean change in height during a given time period increases by z standard deviations as a consequence of the intervention, then this corresponds to a shift of the median of growth to the Φ (z)-quantile of the distribution in the control group (where Φ denotes the cumulative density function of the standard normal distribution. For instance, if z = 0.25, this corresponds to an intervention-related shift of the distribution whose new median is where the 60th percentile of the original distribution was. A team of local researchers fluent in the locally spoken languages is trained over a week in data collection methods, followed by a pilot study. Participation in the survey is voluntary and refusal to participate will have no repercussion whatsoever. The study information and consent forms are translated into local languages. Informed verbal and written consent is obtained from each individual prior to the beginning of data collection. Data will be collected strictly respecting confidentiality. No compensation will be offered in exchange for participation in the survey and no fees will be required from participants. The structured questionnaire is administrated using tablet technology and the Open Data Kit (ODK) software package. Data is stored on a secured server located at the Swiss Tropical and Public Health institute in Basel. Anthropometric measurements of children are taken by trained surveyors using a weighing scale, a tape measure and a (Mid-upper arm circumference) MUAC measuring tape [52]. An intention-to-treat analysis comparing all persons who were initially participating in a VSLA with a control group will establish the effect of participating in the project on household economies (composite wealth score; number of income-generating activities), on child nutritional status (height for age z-score HAZ, weight for age z-score WAZ, and weight for height z-score WHZ, mid-upper-arm circumference), and on the use of family planning (current use of modern family planning method; unmet need for contraception) as the primary outcomes. Household economic status will be assessed using an asset-based wealth score and number of income-generating activities. For the analysis and validation of scaling properties of the composite wealth score principal component analysis will be used. To measure child nutritional status, measures of chronic and acute malnutrition will be used. Stunting (small-for-age) as measure of chronic malnutrition will be measured as height for age index z-scores (HAZ): a HAZ < − 2SD was defined as stunted, a HAZ between -2SD and − 3 SD was defined as moderate stunting and a HAZ < − 3 z-score was defined as severe stunting. Underweight will be measured as weight for age index z-scores (WAZ): a WAZ < − 2SD was defined as underweight, a WAZ between – 2SD and − 3 SD was defined as moderate underweight and a WAZ < − 3SD was defined as severe underweight. Wasting, measuring acute malnutrition as weight for height z-scores (WHZ) < − 2 SD, a WHZ between – 2SD and – 3SD as moderate wasting and a WHZ < −3SD as severe wasting. A MUAC < 115 mm will also define a severe malnutrition [52–54]. To measure food security, the FAO food insecurity experience scale is used [50]. Family planning use is measured as proportion of women currently using modern contraceptives, and as unmet need for family planning. In addition to the primary outcomes, secondary outcomes include changing gender norms (attitudes towards women’s rights, gender-based violence, women’s roles) and women’s empowerment (participation in the economy; self-efficacy to express their views; intra-household decision-making; use of health services; gender-based violence). (Table 3). Power calculations for child anthropometric data Table 3 gives the achievable power for different effect sizes and intra class correlation coefficients. The expected power is given both under the assumption that a) a child under 5 years old will be found in each household, and b) that a child under 5 years old is found only in every other household. Intra-class coefficients reported in other African contexts range from 0.01 to 0.05 SD standard deviation of individual growth rates aProportion of variance explained by the factor village First, descriptive statistics will be calculated for the primary and secondary outcome variables, and for socio-economic characteristics. Assets, number of income-generating strategies, food security and child nutritional status, use of family planning and other reproductive health services, and related knowledge, and prevalence of perceptions of gender relations (beliefs and attitudes), and prevalence of different levels of knowledge on sexual and reproductive health including use of family planning methods and existing services will be documented. For this purpose, percentages, means and standard deviations will be computed. To assess self-efficacy and decision-making power, indices will be built using Mokken analysis, which is a nonparametric procedure based on item-response theory that has been used to assess similar scales in previous studies [55, 56]. Differences between education level, rural and urban populations will be assessed statistically using Chi2 tests for different degrees of freedom for categorical variables, or using t-tests for continuous variables. To assess the program effects, mixed-effect regression models will be run for each primary outcome variable to establish change in the outcome variables over time (baseline to endline) and by intervention and control group. The models will be adjusted for socio-economic confounders, and clustering will be considered at the level of villages. The same analysis as for primary outcomes will be conducted for secondary outcomes. In addition, the role of mediating factors will be investigated. It will be explored whether a change in the power of decision-making mediates changes in household economic status and use of family planning (FP). In addition, we will study whether children’s nutritional status improves if men are supportive of women’s economic activities. For this purpose structural equation modelling (SEM) with maximum likelihood estimation will be conducted to investigate the complex and multidimensional pathways by which the association of a positive masculinity and women’s empowerment directly or indirectly affect household economics status, children nutritional’s status and use of reproductive health services, and the potential role of mediating variables. Results will be discussed under consideration of the fidelity of the intervention. Women who complete the full cycle of a VSLA and men who participate in the full cursus of the peer-to-peer sensitisation group will determine the level of fidelity. Analysis will be conducted with STATA V.15. A qualitative study with households participating in a VSLA will be conducted, collecting data on gender relations, women’s economic participation and access to sexual and reproductive health services. Qualitative studies are by nature smaller and capable of providing in-depth insights into processes within selected households and couples of a particular study site [57]. With the qualitative study, women and their partners are closely followed through multiple interviews, including collection of information related to their income and expenditure, as well as information on gender-based dynamics within families and communities. Health-related behaviour and perceptions of family planning will also be explored in the context of men engaged and women’s empowerment to assess a change in behaviours and perceptions after men’s sensitization as well motivations for the change. For the qualitative interviews a guide is developed focusing on the impact participation in the project has on women’s families in terms of gender-relations, household economy, and health. Over time, the instrument is adapted based on the results from previous interviews to capture emerging themes. Focus group discussions (FGD) with 6–8 participants, with women and men separately will be organized as well as individual in-depth interviews (IDIs). We will use purposive sampling to recruit 18–49 years old women participating in VSLA whose husbands are also participating in either VSLA or a reflection group. Two focus groups will be organized with the same participants after a one-year interval. We plan to do five to seven FGDs coupled with IDIs carried out with same participants as FGD. Participants of the qualitative study are recruited from households where both partners are involved in the intervention: women participating in VSLA and their partner participating in a male reflection group. Women and men in qualitative are recruited in the same villages but are not part of the quantitative study. We plan to conduct 20–30 in depth interviews with women and men or until saturation is reached [58]. Different themes such as gender norms, roles and justice in society, communication between husband and wives, and gendered responsibility in health and in households’ economics will be explored. Observations of women in VSLAs and at their home during the visits will be done throughout the process. Attitudes, the ability to feel confident to speak about anything related to women’s health and the nutrition of children with special regards on the involvement of men will be collected. Interview guides will be developed to explore dynamics of participating in VSLA and men’s sensitization for traditional roles and decision making at the level of the household with a specific focus on gender, household economy, and maternal and children health. Interviews will be carried out by a researcher familiar with the local setting and language. Interviews will be conducted in Kiswahili after written informed consent is obtained. No compensation will be offered in exchange for participation in the study. Interviews will be conducted in a isolated place chosen by the participant either in their home or outside. The researcher will make sure that privacy and confidentiality is always granted. The narratives from qualitative data collection will be voice-recorded and transcribed in Swahili. Observations will also be done during VSLA discussion sessions; notes will be taken and transcribed in French language. VSLA discussions on health topics relevant to this study will eventuality be targeted and recorded for analysis. Observations of women in VSLAs and at their home during the visit will be conducted; notes will be taken and transcribed in French. Coding will be done with the help of qualitative data analysis software (Atlas TI). Latent themes will be identified by inductive analysis, reading and re-reading transcripts as well as notes from observations. At every round of iterative analysis emergent codes will be compared, grouped and contextualized. Finally, using a hermeneutic approach, the emerging hypotheses will be integrated in a wider contextual analysis. Quantitative data will provide associations between outcomes and different factors in the study. Throughout the analysis, qualitative and quantitative results will be discussed in the study team, and triangulation will be done between qualitative and quantitative results (convergent parallel design). The qualitative analysis will shed light on aspects that cannot easily be quantified. Qualitative research is needed in order to understand why people give a specific answer especially for sensitive topics like gender, power and decision-making as those are strongly linked to social norms and individuals perceptions. It will also help to generate hypotheses on how to construct the structural equation models during the quantitative analysis.
N/A
