Women's economic empowerment and health related decision-making in rural Sierra Leone.

Maternal mortality rates during childbirth in Sierra Leone are amongst the highest globally, with 1360 maternal deaths per 100,000 live births. Furthermore, the country's neonatal mortality rate is estimated at 39 deaths per 1000 live births. There is growing recognition of the health consequences of gender inequality, but challenges in addressing it. Gendered power dynamics within households affect health outcomes, with men often controlling decisions about their family's health, including their family's use of health services. The Government's Free Health Care Initiative, which abolished user fees for pregnant women, lactating mothers and children under five is promising, however this reform alone is insufficient to meet health goals. Using in-depth interviews and focus group discussions with men and women, this study explores women's economic empowerment and health decision-making in rural Sierra Leone. Findings show the concept of power related to women's income generation, financial independence and being listened to in social relationships. Whilst women's economic empowerment was reported to ease marital tensions, men remained household authority figures, including regarding health decision-making. Economic interventions play an important role in supporting women's economic empowerment and in influencing gender norms, but men's roles and women's social empowerment, alongside economic empowerment, needs consideration.


Introduction
Sierra Leone has one of the world's highest maternal mortality rates during childbirth, recording 1360 maternal deaths per 100,000 live births in 2013 compared to 546 deaths on average across Sub-Saharan Africa (WHO et al. 2018). The country's neonatal mortality rate is estimated at 39 deaths per 1000 live births, the infant mortality rate is 92 deaths per 1000 live births and the under-five mortality rate is very high at 156 deaths per 1000 live births (Statistics Sierra Leone (SSL) and ICF International 2014).
Adolescents aged 10-19 years constitute about 20% of the population and many reach adulthood with little to no information about reproductive and sexual health (Statistics Sierra Leone (SSL) and ICF International 2014). According to the 2013 Sierra Leone Demographic Health Survey (DHS), the majority of young people aged between 10 and 25 years have incorrect information about HIV prevention methods such as condom use or limiting number of sexual partners (Statistics Sierra Leone (SSL) and ICF International 2014;WHO et al. 2018).
To address these poor health indicators, in 2010 the Sierra Leone Government launched the Free Health Care Initiative (FHCI). This enabled pregnant women, lactating mothers and children under five to access medical care and services free-of-charge (Government of Sierra Leone 2009). The introduction of the FHCI was the culmination of a growing commitment from the Sierra Leone government, and the Office of the President, to tackle the high levels of maternal and child mortality (Donnelly 2011). The FHCI addressed the issue of out-of-pocket health care expenses that had previously been shown to reduce access to care (Ensor and Ronoh 2005).
FHCI evaluations are promising, showing associated improvements in service use, particularly amongst the most disadvantaged in rural areas, despite the shocks of the 2014 Ebola crisis (Witter, Wurie, and Bertone 2016). Furthermore, the 2013 DHS indicated that 97% of women had attended at least one antenatal care appointment with a skilled provider during their last pregnancy (Treacy, Bolkan, and Sagbakken 2018;Statistics Sierra Leone (SSL) and ICF International 2014). However, information about the quality of care received is not provided. Despite the high percentage of women accessing antenatal care, and regardless of the FHCI and introduction of informal by-laws, in 2013, 46% of women did not deliver at a health facility (Statistics Sierra Leone (SSL) and ICF International) 2014).
Challenges remain with the uptake of health services, particularly in rural areas where primary health care infrastructure is weak, including prohibitive costs (Sharkey et al. 2017). For individuals and households, there are still costs associated with accessing care such as transport, informal charges or contributions to health care staff, official user fees and lost income. These factors can considerably affect household financial security, with multiple difficult financial decisions often needing to be made about household health (Treacy, Bolkan, and Sagbakken 2018;McNamee, Ternent, and Hussein 2009).

Relationships between gendered power and health decision-making
Gendered power relations play a key role in how health-related decisions are made, and who makes them (Gerber 2013;Rodin 2013). In households and communities, these power dynamics can affect health-seeking behaviours and outcomes. In many and particularly rural communities globally, men often control decisions about the health of their wives and children, including the family's use of health services (Barker et al. 2011;Ensor and Cooper 2004). For example, studies in Ethiopia and Iran have found links between husbands' approval and women's use of modern contraception and maternal and child health services (Mohammed et al. 2014;Rahnama et al. 2010). In Sierra Leone, inequitable gender norms are among many factors, including the fragile health infrastructure following the 2014 Ebola outbreak and high poverty and illiteracy rates, that continue to undermine the availability and accessibility of health services for the most vulnerable children, women and men.
In order to address issues of gender inequality, the past decade has seen an increase in programmes focussed on women's empowerment (Seymour and Peterman 2018). Sustainable Development Goal (SDG) 5 identifies achieving gender equality and empowering all women and girls as one of the key steps towards achieving inclusive and sustainable development by 2030. However, the meaning and measurement of the concept of empowerment has been contested (Gram, Morrison, and Skordis-Worrall 2018). Most social science research has based its conceptualisation of empowerment on the notion of 'agency', defined by Sen as the 'ability to use those capabilities and opportunities to expand the choices they have' (Sen 1999, 10), and on women's ability to participate in decision-making over certain vital matters (e.g. major household purchases or personal health care) (Seymour and Peterman 2018). Evidence shows that increasing the availability of economic resources alone, however, is insufficient and also requires increased women's agency, in order to meaningfully impact their decision-making and control of household resources (Peterman et al. 2015).
There is substantial interest from practitioners and researchers in the interactions between increasing women's economic empowermentwomen's increasing power to make economic decisionsand positive health-seeking behaviours. A substantial body of literature supports the idea of gendered behavioural patterns in allocating resources (Peterman et al. 2015) and that when women hold greater influence over or control income they are more likely to make spending decisions which improve family welfare and have positive health impacts (Buller et al. 2016). Women as key actors in improving health outcomes largely results from common cultural norms, such as the role of women as carers within societybeing more likely than men to be responsible for food sourcing/preparation, and care of infants, children, and the sick and elderly.
However, evidence also suggests that women are less likely to generate significant proportions of household income, and to control or contribute to household financial decision-making (Hughes et al. 2015). This is due to factors including inheritance and land rights, gender-based violence, limited access to education and restricted access to credit and financial services necessary to develop income-generating activities (Hughes et al. 2015). Women's work is more likely than men's to be unwagedin domestic work, and care of children and the elderly, or within informal sectors such as small-scale agriculture and hand crafts. However, globally, whilst all women are not impacted equally, female-headed households are on average poorer than male-headed households (Heise 2012).
Related to this, feminist theory posits that societal-level power imbalances within patriarchal societies create structural factors that work directly and/or indirectly to validate a male-dominated family structure. This often results in men exercising power and control over women in several ways, including through violence (Antai 2011). This leads to the argument that enhanced female autonomy and power can challenge men's control and domination over women in these patriarchal societies (Kabeer 2005). However, gender norms and roles matter. Despite being economically empowered, men may use their power against their spouse when household roles, conditioned by traditional gender norms according to the social context, are either not fulfilled or transgressed by one party (Zegenhagen, Ranganathan, and Buller 2019). Applied to decision-making, this means that an increase in household tension is more likely when either women assume a decision-making role considered traditionally male, or when men are unable to fulfil their role as household decision-maker or breadwinner (Zegenhagen, Ranganathan, and Buller 2019).
Against this background, this study aimed to explore the relationship between women's economic empowerment and health decision-making in rural Sierra Leone to understand better the implications for women's health-seeking behaviour, and to support the design and implementation of interventions to address maternal, child and new-born health. Our research questions emerged from discussions with project implementing staff who highlighted a need for in-depth understanding of the conflicting pressures on women's lives regarding health decision-making. This study offers a systematic way to explore research participants' core concerns and makes an important contribution to national-level debates and associated programme design and implementation.

Study context and setting
The international development and humanitarian charity, Christian Aid, in partnership with Rehabilitation and Development Agency, Sierra Leone (RADA-SL), has implemented a 3-year project across 22 communities in Pujehun, Sierra Leone. The project aimed to improve reproductive, maternal and child health outcomes for women and adolescents through strengthening the capacity of communities and health systems. The project's activities have included facilitating participatory community planning processes to develop action plans to identify and address community needs and build resilience. In the process, economic empowerment has emerged as a key issue for the communities in relation to health care. Responding to the action plans, RADA-SL facilitated a process of establishing village savings and loans associations (VSLA), using a microcredit model to address both credit needs to deal with health demands, and to finance small business development. Furthermore, Christian Aid and RADA-SL staff working on these projects to improve health had identified gendered inequalities in household income and financial decision-making as contributing factors to multiple health problems and health-seeking behaviour. We conducted our qualitative study in a subset of the communities (n ¼ 10) that Christian Aid and RADA-SL are working in to explore the relationship between women's economic empowerment and health decision-making.
In the Pujehun district where this study was undertaken, poverty is widespread, and there is a low population density and largely rural population, with only 8.1% of people in the district living in an urban area (Weekes and Bah 2017, 33). Pujehun district also borders Liberia, and the border is porous, with communities easily moving across it with little distinction. The population of Pujehun is predominantly Mende (one of two majority ethnic groups in Sierra Leone, and the largest in the Southern province) and majority Muslim. Nationally, 71.1% of household heads are men and polygamy is common; about two thirds of the married population (67.9%) are in a polygamous marriage (Statistics Sierra Leone 2016, 71). Health service provision in Pujehun is of poor quality, with an estimated 74 health facilities including one government hospital, 13 Community Health Centres (CHCs), 15 Community Health Posts (CHPs) and 45 Maternal and Child Health Posts (MCHPs) covering a geographically dispersed population (OCHA Sierra Leone 2015). Digital connectivity in the area is low, with only 7.9% of the population over ten years of age having internet access (Statistics Sierra Leone 2016).

Data collection
For this qualitative study, we purposively selected 10 communities in Pujehun district from the 22 communities that RADA-SL works with to ensure maximum diversity in terms of geography (access to larger population centres, e.g. Pujehun town) and livelihoods (e.g. access to quarries or fishing, as well different agricultural settings). Within these communities, three women were randomly selected from RADA-SL's database of VSLA participants for IDIs, with six 'alternates' also drawn where the first choices were not available.
In total, we conducted 29 in-depth interviews (IDIs) with women in a rural area of Pujehun district to explore the realities of women's day-to-day lives and how they negotiate health spending decisions. We explored how women perceived economic empowerment, and how experiences of power impacted their health decision-making, for example, who in their household decides how and when medical assistance is sought for themselves and their families. This was further explored through five focus group discussions (FGDs) with women, and five with men, complementing the individual data with an exploration of community norms and perceptions. Each FGD had between 6 and 8 people, thus involving a total of 40 men and 38 women.
The IDIs took between 40 and 60 min, were conducted in the Mende language and followed a semi-structured, thematic interview guide. FGDs also followed a semi-structured schedule, with some participatory exercises used to initiate discussion, and lasted between one and two hours. The semi-structured interviews explored five domains: (1) participant experiences of income-generation activities; (2) financial management and decision-making; (3) conceptualisations of power; (4) health problems and decision-making; and (5) experiences of changes through the VSLA. Within these broad domains the interviews explored participant relationships with partners and family members.
The FGDs took between 60 and 90 min and followed the same domains. We used participatory exercises to prompt discussion of power and decision-making. The first was a ranking exercise, with participants using pebbles to apportion decision-making authority to different family members in response to prompts about health. The second was a body mapping exercise to metaphorically explore different type of power (e.g. movement through feet, work through hands, speech through mouth). This led to a discussion of who had access to different kinds of power, and how it impacted health decision-making. These exercises were drawn from Christian Aid's approach to power analysis, as well as their own toolkits (Christian Aid 2016). We used FGDs because they are effective in yielding insight into how people construct collective notions of power dynamics (Bandali 2011). Rather than eliciting detailed information about individual stories, the FGDs focussed on community dynamics; they aimed to draw out norms and instances of consensus or disagreement.
Data were collected by two teams (made up of two women and one man) of research assistants from Njala University. Research assistants spoke Mende, but received additional training from RADA programme staff on local dialect and cultural nuances, and in qualitative research methods. The IDIs were conducted by women researchers, with the FGDs facilitated by mixed teams. The research assistants translated interviews and FGD materials from Mende to English. All the interviews and FGDs were digitally recorded and translated and transcribed.
It is important to mention the complex and multi-level power dynamics within the research context of the investigative team. Unequal power dynamics were particularly apparent in the uneven access to information technologies. Those based in the betterconnected cities of Nairobi, Abuja and London found it easier to connect remotely to group conference calls. Connection with Freetown and with academics at Njala University were more difficult. There was also a discrepancy in resources between research teams and interviewees: academics at Njala University had better resources and communication skills than research participants. Moreover, diversity within the research team (in terms of location, access to resources and research involvement) led to different expectations of the research. Navigating this required additional time for discussion, reflection and team-building that may have at times affected the data collection and research process.

Data analysis
The approach to analysis was both inductive and deductive, drawing from the available data to generate reasonable explanations (Saldana 2009;Ritchie and Lewis 2003). The academic lead in Sierra Leone with support from the research assistants recorded their initial findings and impressions from the data and field work. Using these notes and existing insights from the literature on women's economic empowerment, we developed an initial tentative coding framework, identifying some key themes and codes. The IDIs, male FGDs, and female FGDs were analysed as separate datasets, using the same initial coding framework, but each generating their own refinements, in response to the different data.
Four steps were followed in the analysis of the transcripts for each of the data sources (1) Reading for content. The analysis started with reading and re-reading transcripts until the content became familiar; (2) Development of the coding framework. The initial codes were assigned to specific sections of text; and the framework was refined with the addition of new codes and refining of existing ones through an iterative process. Once transcripts were coded, these were displayed, and used to identify new themes and interrogate initial themes identified; (3) Data reduction and display. We used matrices that categorised and displayed data to help us understand and explore further dimensions, identify sub-themes that reflected finer distinctions in the data, and to also facilitate comparisons between the different data sources; and (4) Interpretation.
Once the themes and central ideas were extracted, we identified and explained the data's core meanings, bringing together the different themes from the FGDs and IDIs. We also searched for relationships among themes or concepts identified from the analysis (Braun and Clarke 2006;Creswell 2003). The first author also discussed the emerging findings with senior Christian Aid programme staff, who played a critical role in verifying data analysis and interpretation.

Ethics and informed consent
Ethical approval for the research was granted by the Office of the Sierra Leone Ethics and Scientific Review Committee and the London School of Hygiene and Tropical Medicine's Research Ethics Committee. The research assistants attended a 5-day training at Njala University. All participants gave written informed consent with verbal explanations and co-signatures from witnesses where participants had limited or no literacy. As part of our informed consent procedures we asked the participants if they would agree for us to include anonymised quotes from their interviews (included as pseudonyms) in potential papers.

Participant socio-demographic characteristics
In total, 29 women participated in the IDIs, and 40 men and 38 women participated in FGDs. The IDI respondents were mothers, aged between 22 and 65 years, although several were unsure of their exact age and many had lost children. The majority were married at the time of the interview (n ¼ 24, 83%) and of the remaining five, three were widows, one was separated from her husband, and the other reported that her husband had left her. All of the women (n ¼ 29) were Muslim; 26 were Mende, one was from the Temne tribe, one was of Susu and Mandinga heritage, and one did not respond. FGD recruitment involved convenience samples, taken from participants in the VSLA schemes, or in the case of men, the male relatives of those participating in the VSLA. All participants were over the age of 18 and predominantly from Muslim, Mende backgrounds, as recorded by the research assistants. Ages and detailed demographic information for individual FGD participants were not collected due to the project demands in challenging environments.

Principal themes
We identified four main themes from the analysis: first, participants expressed appreciation of their own economic activity and their husband's appreciation of their financial contribution; the second provides insights into notions of power in the households, particularly linked to women's sense of financial autonomy; the third shows women's perceived role as custodians of family income; and the fourth illustrates women's decision-making role and its intersection with household power dynamics.

Generating income; increasing harmony within the household
In the IDIs, most women reported engaging in agricultural activitieseither growing food for direct consumptionto generate their own income by selling at the market, and as they themselves described it, labouring on cash crops for their husbands. Many also reported being engaged in petty trading (e.g. buying larger amounts of oil or flour and selling on for a small profit). Others mentioned processing crops to generate more incomepalm oil, or a green culinary herb that is dried and groundfor sale. Some women mentioned additional sources of income such as breaking rocks or gathering firewood. FGDs with both men and women confirmed that it was usual, and expected, for women to be engaged in income-generating activities. Most of those women with husbands reported that their husbands worked as farmers, mostly on cash or bulk crops including upland or lowland rice, cassava, oil palm or groundnut, rather than garden vegetables. Several reported additional sources of income for their husbands, ranging from rock breaking to teaching.
When asked about activities they enjoyed, many women (n ¼ 18) reported enjoying workingeither in regard to growing food to eat or feed their family, or in seeing their business grow or being able to send children to school or make specific purchases. As one woman explained: I do so many things … When I do weeding, planting, harvesting and the selling of the produce to get money for us to pay our children's school fees, I feel happiness within myself. (Jilloh, W 30 y) When am healthy I am able to do a lot of work, I do this with happiness. (Tenneh, W 36y) I enjoy doing business because I touch money, whenever you touch money you feel good. (Marie,W,30y) Most of the women (n ¼ 27) described their husband's reactions to income generation as positive, and this was confirmed in the FGDs. They described how the increased income gave them some financial independence, strengthened their confidence and meant that they were less likely to ask their husbands regularly for money. This helped improve relationship quality with a common refrain that it increased unity and 'peace' within the relationship. This was through the mechanism of reduced financial stress and reduced tensions within the relationship, as illustrated by the following quote: I think he thought it was a good idea because before, there were conflicts between us whenever I ask him for money for salt, Maggi [culinary seasoning] or soap. However, now I do not ask him for all of these things. He is very happy about me and I am happy about him also. (Marie, W, 30y) The women saw themselves as supporting their husbands and reducing the burden on the man as the provider for household needs.
To remove our husbands from shameful situations that face them when they do not have money. (Jeneba,FGD 203 F) Men in FGDs also mentioned improved relationships through the women earning money: There was no peace between my wife and I before, when she wasn't earning money, but now I give thanks to God. (Ibrahim,FGD 201 M) Other men in the FGDs also mentioned that with women earning money, the family's health and hygiene needs are better taken care of:

Musa:
The positive part is that when women generate money is good because, it is what has help us to live a better married life. That is why we are happy with women working to generate their own income.

Sadik:
The positive part of women's generating income is that these women have promoted good health in our village. They have used their money to make our community very clean. They have also made our homes very clean, especially for sleeping.

Notions of 'power' within households
In the interviews, women were asked about their understanding of power and when they felt most powerful. A common theme from almost all women was the relationship between power and feelings of independence to earning or having money. One participant described power as 'when you have money' and when prompted by the interviewer about a time she felt powerless she responded: It was also common for the women to link power to good healthhaving the strength to work and take care of responsibilities. There was a strong emphasis of being able to 'solve problems' with money. Increased income meant being able to support themselves and children and extended family experiencing hardships. It was also used to make contributions to the healthcare and funeral expenses of family membersthe most commonly described situation in which women experienced greater income as a source of strength.
It happened even when my mother's relative died, I was unable to give my own contribution. As a result, I felt powerless. (Fatama, W, 25y) Some women also associated feeling powerful with 'being able to generate their own livelihood' and having control in the household as illustrated by the comment, 'I am always powerful in my home. When I tell my children to stop doing something, they stop, that shows I am powerful'. Jilloh W, 30 y). Several women (n ¼ 7) also described power as coming through social relationships, and mostly with their husbands; for example, power when their husband listens, or when they can help their husband.
If your partner agrees in whatever good thing you suggest, it means you have power. To me power is when your partner agrees with what you say. It means you have power. (Marai W,31y) In a KII with an NGO worker, the respondent explained that women gaining access to more resources improved their confidence to participate in community decision-making.
… that is providing confidence to the women, because they have their resources now, so they can speak and challenge from an informed point of view. Because initially even with their husbands they haven't been recognised by them. Because they think they are not economically empowered, so they are just subjects in the homes. They don't have a stake. (KII 101, RADA). Some women mentioned power coming from having a place in the community and being respected by people in the community.
A woman with money can command a lot of power here because this woman is doing a lot of things for herself without the help of anyone. People respect her, so she has power. She can solve a lot of problem on her own. That is why I think she is powerful. (FGD 202 M) A few women (n ¼ 3) indicated that having money to buy clothes and shoes increased their respect in the community. The FGDs with men also reflected the ideas that power equated having good health, possessing money and having resources. In addition, several men in the FGDs spoke about the benefits of 'togetherness' or being cooperative whether at work or in the household with partners.

Konneh:
Power is about togetherness or cooperation ('Ngoyila' in Mende). If you are together you can break stones out there and transport them for sale. Another kind of power is effort/seriousness, if you are serious, and make more effort you can break those rocks for sale; that shows you have power of will.
John: … If there is no togetherness we would not have gathered here as one, in order to answer your questions or take part in this discussion. Even if it pertains to money, we won't come there, but because we are together that is why we are sitting here with you.

Hassan:
In the household it is also togetherness e.g. if my wife and I want to do something, we first hold consultation before proceeding to undertake the activity.

Women as custodians of family income
A strong theme was that women are better able than men to safeguard household income, save money and to ration spending on household finances. Women provided multiple reasons for this that were also reflected in their FGDs and several KIIs; men might use the money to court other women, or loan it to friends, or are simply more likely to spend it on items such as tobacco. There were several different models of managing household income discussed in the IDIs and FGDs. Mostly the women held their own earnings, and whilst they might 'show' their money to their husbands, and the husbands would be aware of it, the women would keep it themselves. To some extent, this arrangement is likely to reflect the polygamous social structure whereby multiple families may be supported by one husband. In many cases, the husbands would give money to a wife for safe keeping, and the wives would hold this money, often keeping it separately from their own income. In a few cases there would be shared household finances, with both spouses contributing and the money held by one of them. Husbands were also more likely than their wives to be mobiletravelling for work opportunities or into town.
The data suggested that women have more 'pro-social' saving habits; they were more likely to save money and spend on family welfare, food, education and investment into business. Conversely, men were commonly perceived as being less reliable. This VSLA 'box' [a safe box that holds the contributions to the VSLA and has two key holders and an account book] has brought peace in our home because I no longer ask my husband for salt or Maggi to cook. This box has taught our husbands that we women also deserve respect. (Jeneba,W 25y) In situations where the husband is not financially strong, the home will suffer; but if the women is empowered, she is able to adequately support the home. Therefore, when women are empowered, the home and the entire community will be safe. (KII, Pujehun District Council).
Overall, there was a sense that empowering women is beneficial for their relationships, for their individual confidence and for their communities more generally.

Decision-making roles and expenditures around health care
Despite women being custodians of household income, certain types of decisions were still considered as within the domain of men. This is specifically decisions regarding health-care decisions as demonstrated by this comment from a woman about having to pay extra for good health care treatment.
You cannot depend on free medical healthcare. I pay for treatment for me and my children. The treatment that is free is not good, but the one you pay money for is better for you and your children. (Hawanatu, W, age unknown) In one IDI, a woman described having to pay out-of-pocket expenses for her infant's health by paying nurses.
If you give money to the nurse for treatment, she will treat your children well, but if you do not give them money, the treatment will be less. When money is given, proper medication is given to cure your children. (Kula, age 25 years) Decision-making responsibility regarding when to seek health care and the payment for it was frequently reported to be the man's responsibility (n ¼ 17).

JTK:
Who typically makes decisions on whether or not to seek help from nurses or other health workers?
Jombu: It is I, as the husband, I am the one who decides for family members to seek medical help elsewhere. I will ask my wife to take loan [from the VSLA] if we don't have enough money.

JTK:
Who pays for the related cost of health services?
Thomas: I am the one who pays. If my wife takes out a loan from the scheme [VSLA], she hands over the money to me and I negotiate with the nurses for the payment.
(FGD 101, M) However, many women disclosed that in the absence of men taking a decisionmaking role, they would make decisions themselves regarding whether to seek necessary health care. This occurred when men avoided the health clinic visits with their wives, in order to avoid incurring the costs, despite wanting to be responsible for decision-making surrounding the visit.
Even though the free health care targets the under-fives, lactating and pregnant women, the husbands don't also accompany their wives since they will be expected to meet at least the costs for baby wrappers and napkins and so they will prefer not to accompany their wives. In cases where the pregnant woman is economically empowered, the husbands will be more than willing to attend since they will incur no extra cost. (KII 202,Pujehun District Council) In cases where women paid for health care treatment for children themselves, they would ask for reimbursement from their husbands. For health care treatment of adult women, or for other family members, the decision was less clear. Many relied on the VSLA scheme to cover adult health expenditures and repaid the loan gradually. Before the VSLA scheme was implemented, women mentioned that they used more expensive loans from other traders or borrowed money from family members.
Decision-making about family planning and contraceptive use appears to be mostly made jointly by husbands and wives. Nurses were also reported to make decisions about the types of contraceptive use. Injections and implants were more commonly reported as contraceptive options than condom use, and traditional methods of tying string about the waist were also mentioned. Many women reported unpleasant side effects of contraceptive use.
Despite general support for women's economic empowerment, in terms of income generation and making decisions for household-related purchases, most men in the FGDs believed that they were ultimately the household head, and that they held decision-making authority over their wives. They were clear that women needed to consult their husbands. Transgression of these gender roles could increase household tension.
Even if the woman is generating money, if she does not respect her husband that is not good. It is the husband that has allowed her to work, to look for money. So we need to cooperate. If my wife is generating income, is like a trap which can catch an animal, it is I who determines it. My woman should not disrespect me even though she is generating income for the home. If she does so, that will be bad. (Jibao, FGD 101 M)

Discussion
Findings from this study contribute to the growing body of research on women's economic empowerment, gendered notions of power and gender roles related to health decision-making in rural Sierra Leone. In particular, they reveal the benefits of economic empowerment of women and the complexities surrounding the intersection with male authority patterns in this socio-cultural context of rural Sierra Leone. Three main findings emerge from this research.
First, women were frequently involved in various income-generation activities. Almost all of the women interviewed were working in some capacity and enjoyed the financial independence and accompanying confidence boost. Several women expressed feelings of happiness and fulfilment when they earned money, and the additional income was reported to reduce relationship tensions. Primary reasons given by women when asked about their understandings of power and their reasons for feeling powerful were earning money and being financially self-reliant. Furthermore, being financially independent contributed to women feeling economically empowered as they were in a better position to be able to fulfil their responsibilities, such as looking after children's needs, contributing to funeral costs and being able to afford desired items. This corresponds to assertions from feminist theory that enhanced female autonomy and power improves women's bargaining power due to its association with women's greater confidence, financial independence, and a lower likelihood of tolerating male domination and control (Buller et al. 2018). However, the evidence regarding this is still limited and these relationships are complex, as described in several studies (Mishra 2014;Schuler and Nazneen 2018). In addition, several of the women and men mentioned cooperation and stronger social relationships, both within the household and with other community members, as a way of feeling powerful. This aligns with research findings from Southern Africa that strong social capital and relationships contributes to increased household harmony and raises women's self-confidence (Brody et al. 2015;Ranganathan et al. 2019).
Second, women were considered by both men and women as custodians of family finances; they were more likely than their husbands to keep money safely and also to use the money for family-related expenses. This aligns with research from other settings in sub-Saharan Africa where women are considered 'household managers'considered to economise and to have a financial interest in the entire family (Wasilkowska 2012). In addition, this study found that the VSLA scheme has been particularly helpful in supporting women and their families to pay for health-related expenses leading to greater harmony in households, as women have less need to ask men for money. Additionally, VSLAs are particularly relevant in this context where formal banking mechanisms are not easily accessible, as members save and borrow and have full control over their finances. There is evidence of the positive aspects of the VSLA scheme and the effect on gender relations from other East African settings (Nuwakora 2014).
Third, decision-making relating to health was considered to be within the man's domain. The cultural norm as described by the participants was that men were responsible for the healthcare costs related to pregnancy and childbirth, and for their children.
However, it appears that whilst men want to be in control of health decisions, they were not necessarily able to follow through on actual transactions, either because they could not afford to do so or because they do not prioritise these payments. Men expected women to take out a loan, or to use their own savings, if they themselves could not afford a health care service. Whilst the FHCI provides free services to pregnant and breast-feeding women, and to children under five, health care costs remain for older children, and other household members. Even for those entitled to free health care, there are still expenses associated with transport, clothing or wrappers for babies. In a similar study in another district in Sierra Leone, the participants described these associated health care costs (Treacy, Bolkan, and Sagbakken 2018, 17). In this study, women could express a preference or make the final decision about contraceptive use. In fact, in the absence of men, women would make health-seeking decisions and pay for care, as needed.
There was an overall recognition amongst respondents that despite encouraging women's economic empowerment and having women in charge of household finances, men still assert greater power in the household, and expect women to respect and defer to them. This was clear from the FGDs, with women seeking their husband's permission even after earning more money. Despite being engaged in incremental processes of changing gender relations and achieving a greater sense of empowerment, the women nonetheless appeared averse to upsetting the persistent gender norms that define husband-wife relationships. Implicitly they appear to acknowledge that their own empowerment is most beneficial when it also enhances family and community cohesiveness. This aligns with a study in Rwanda that showed that traditional gender roles still prevailed while women achieved a greater sense of empowerment, with men claiming authority in decision-making (Stern, Heise, and McLean 2017).
This study underscores the potential value of publicising new models of shared household economic and decision-making roles for the lives of women, men and children. Critically, any effort to shift attitudes towards gender roles to support women's empowerment and gender equality will require gaining the support of men and combining complementary programmes on gender training with economic interventions. Furthermore, there is value in external agencies consulting local (often, rural) women, and developing assistance projects with local women's input.
From a programmatic and policy perspective, the high costs of accessing health care will continue to impact on women's economic empowerment particularly given, as shown in this study, that health care costs are often covered by women and are often the cause of debt and hardship. Whilst the FHCI has seen improvement in health care access, fees remain for children over five years and adults. This emphasises the positive benefits of extending the FHCI to additional groups, or additional levels of support for accessing health care services. The VSLAs emerged as a mechanism for mitigating some of the risks associated with ill health. But, the schemes are small-scale and could struggle to cope with increases in demand. The VSLA project established by Christian Aid and RADA-SL is still being fully established, and the evidence is yet to emerge regarding its longer-term sustainability, although the data presented here are promising. In addition, when VSLA funds are used for health care, this restricts the amount of money available for small business loans that offer longer-term financial security and independence and may slow the progress of wider economic development.

Strengths and limitations
In terms of the study's strengths, this study was undertaken in the communities of Pujehan district in Sierra Leone, one of Sierra Leone's 16 districts. Fourteen of these communities are rural with considerable similarities in levels of poverty and poor infrastructure. Pujehun is more rural than the other districts with a small capital town and some areas cut off by river floods in the rainy season. It mainly relies on agriculture, but the district also has economic opportunities through fishing, mining and quarries. However, the region is also broadly representative of rural Sierra Leone and our findings may have relevance to other parts of rural Sierra Leone (Tracy 2010). A further strength of this study is that we conducted FGDs with both men and women (separately), to gain the perspectives of both on gender-related issues.
The study's main limitations result from the attempt to undertake systematic research with a small budget in a challenging rural context. The communication and connectivity difficulties with, and within, Sierra Leone, meant that the study processes had to be adapted in-situ. Travel times and challenges with accessing the more remote communities were considerable. A particular limitation emerged around language; the dispersed and rural context means that Mende is a highly heterogeneous language with distinct local dialects. This meant that the initial training had to be refocussed to include increased language sensitisation provided by RADA-SL for the field research teams. This was necessary and beneficial to the project but resulted in limited time spent preparing the participatory activities for the FGDs. The resulting datawhilst reported as useful for stimulating discussion and producing some valuable insightswas not directly comparable between the different facilitator teams.

Conclusion
This study of women's economic empowerment and health decision-making in rural Sierra Leone found that female respondents frequently described power as relating to women's income generation and financial independence, as well as in terms of women being listened to generally in their social relationships and by their husbands. Whilst women's financial independence was reported to ease marital relationship tensions and supported their ability to undertake responsibilities, men still remained the authority figures in households, often in regard to health care decision-making. The study findings suggest that economic strengthening models, such as the VSLA, can have an important role in supporting women's economic empowerment, especially as they relate to health-related decision-making and addressing unequal gender norms. However, involving men, and coupling economic interventions with complementary programmes to further tackle social empowerment could result in a truly gender transformative programme.