Group-based microfinance for collective empowerment: systematic review of health impacts

Objective To assess the impact, on health-related outcomes, of group microfinance schemes based on collective empowerment. Methods We searched the databases Social Sciences Citation Index, Embase, MEDLINE, MEDLINE In-Process, PsycINFO, Social Policy & Practice and Conference Proceedings Citation Index for articles published between 1 January 1980 and 29 February 2016. Articles reporting on health impacts associated with group-based microfinance were included in a narrative synthesis. Findings We identified one cluster-randomized control trial and 22 quasi-experimental studies. All of the included interventions targeted poor women living in low- or middle-income countries. Some included a health-promotion component. The results of the higher quality studies indicated an association between membership of a microfinance scheme and improvements in the health of women and their children. The observed improvements included reduced maternal and infant mortality, better sexual health and, in some cases, lower levels of interpersonal violence. According to the results of the few studies in which changes in empowerment were measured, membership of the relatively large and well-established microfinance schemes generally led to increased empowerment but this did not necessarily translate into improved health outcomes.


Introduction
Microfinance initiatives have become popular, particularly in low-and middle-income settings, as a means of promoting rural development, 1 increasing the bargaining power of women and improving household welfare. 2 Such has been the enthusiasm for these schemes that, in 2006, the Nobel Peace Prize was awarded jointly to Muhammed Yunus and Grameen Bank -a microfinance scheme in Bangladesh.
The potential of microfinance to improve health is now being recognized. [3][4][5] The impacts of microfinance initiatives need to be considered in current theory debates about the role that control over destiny plays as a fundamental social determinant of health. 6 Poor control over destiny, which is a characteristic of women in some societies, can be damaging to population health. In general, population and child health improve and the life expectancies of both men and women increase as the participation of women in decisionmaking increases. 6 Group-based microfinance schemes attempt to harness the collective power of mutual support -with members pooling their savings and making small loans to each other so that they can set up small businesses. Most aim to improve the economic power of -and employment opportunities for -women in their immediate community, and many aim to confront engrained discriminatory attitudes to women. 2 Some aim to facilitate the attendance of girls at school and change attitudes to the paid employment of women outside their homes.
The members -who are mostly women -form groups for saving and credit, and are offered literacy classes, legal, social and empowerment training and technical and marketing support (Box 1).
It has been argued that the enthusiasm for microfinance has outstripped the evidence of its effectiveness 7 and that microfinance schemes have the potential to do harm. Schemes can suffer from so-called mission drift and end up favouring those who are more creditworthy while excluding the ultra-poor. [8][9][10][11][12] In some settings, the imposition of a business model on poor female members of a microfinance scheme may lead to increased debt, repayment stress and exploitation. 13,14 The result may be an exacerbation of inequalities rather than a reduction.
We conducted a systematic review of group-based microfinance based on collective empowerment that covered all health conditions and all countries and assessed the impact on health. We addressed three questions: (i) what impact do group-based microfinance schemes

Methods
We reviewed evaluations of group-based microfinance in any country, using published systematic review methods, 15 and assessed the quality of each relevant study using procedures tailored to social interventions in community contexts. 16

Search strategy
We We checked the reference lists of relevant articles and contacted policy-makers and academics for publications in press and in the grey literature.

Inclusion and exclusion criteria
A report was only included if it described an experimental or quasi-experimental evaluation of a group-based microfinance scheme that: (i) employed collective empowerment strategies; (ii) was targeted at a group with some form of disadvantage; and (iii) was delivered among a free-living population in a community setting. To be included, a report also had to disaggregate data by some measure of socioeconomic status and describe at least one healthrelated outcome. We also included qualitative reports that related to an included study. No country or language restrictions were applied.
We excluded reports of individual loan schemes that focused solely on poverty alleviation but did not promote group solidarity and empowerment, and reports on schemes that included restrictions on how loans could be used.

Screening and selection
Titles and abstracts were screened before potentially eligible reports were retrieved in full text and assessed, independently, by two reviewers. Reasons for exclusion were recorded.
Disagreements were resolved by discussion or by recourse to a third reviewer.
Publication: Bulletin of the World Health Organization; Type: Systematic reviews Article ID: BLT.15.168252

Study data
A single reviewer extracted data from each included report and applied a modified version of the quality assessment tool developed by Lorenc et al. 16 Qualitative studies were assessed using the criteria of Mays and Pope. 17 A second reviewer checked extractions and appraisals for accuracy and completeness. A narrative synthesis was performed. 18,19 Differential impacts were identified -particularly in relation to ethnicity, sex and socioeconomic status. Reporting was based on the PRISMA-Equity 2012 extension guidelines. 20

Results
From 4050 articles, only 31 reports -covering 23 studies (Table 1) -met our inclusion criteria ( Fig. 1). The included studies comprised one cluster-randomized controlled trial and 22 quasi-experimental studies that took advantage of naturally occurring comparisons and pre-existing data -e.g. from demographic surveillance systems and health surveys. All of the interventions targeted poor women living in low-or middle-income countries. Most were based in Bangladesh and many focused on women in rural communities. Although we identified some studies of microfinance schemes in central and south America, all but one were excluded because they did not meet the inclusion criteria.
Assessment of the included studies revealed that even the higher quality studies were potentially at risk from several forms of selection bias.

Mortality and morbidity
Two higher quality longitudinal studies revealed that membership of the BRAC initiative, in Bangladesh, was associated with relatively low infant and child mortality compared with non-membership. 21,22 The decline observed in the risk of infant death over a period of 10 years was greatest (53%) for infants of mothers who joined the BRAC scheme, followed by the infants of rich non-members (41%) and then the infants of poor non-members (31%). 21 The risk of death for the infants of poor BRAC members declined to the level recorded for the infants of rich non-members. There was no association between BRAC membership and survival of children aged 1-5 years. In a further study by the same authors, however, the survival of children aged 1-5 years from poor households was found to be significantly improved if their mothers were BRAC members. 22 Two lower quality studies found that BRAC membership was associated with lower child mortality 23 or lower maternal morbidity. 24 A third study, based in Peru, found no association between length of membership in a group-based microfinance scheme and maternal depression 25 29 Qualitative data indicated that the intervention had led to a greater acceptance of intra-household communication about HIV and sexuality and increased confidence and skills that, in turn, appeared to have supported the introduction of condoms in sexual relationships. 29 Findings on the impacts on women's sexual health assessed in two other higher quality studies were equivocal. In Bangladesh, women in villages participating in the Grameen Bank scheme were more empowered (P < 0.01) and more likely to use contraceptives than women in villages without the microfinance scheme (59% versus 43%; P < 0.01). 31 In contrast, no significant association was found between BRAC membership and contraceptive use. Ethnographic data indicated that the Grameen Bank scheme may have increased contraceptive use partly by strengthening women's economic roles and empowerment and partly by directly promoting family planning and influencing community norms. 32 Members of microfinance schemes showed relatively high scores for economic security, contribution to family support, freedom and mobility and freedom from domination. 32 A further longitudinal controlled study in India, evaluated the three-year impact of Pragati -a multicomponent microfinance and health-promotion intervention for female sex workers. In this study, the incidence of sexually transmitted infections decreased and the frequency of condom use at last paid sex increased as microfinance exposure increased over time. 33 Three lower quality studies also assessed women's sexual health. One showed associations between microfinance membership in Bangladesh and higher contraceptive use. 34,35 The other two, in Ethiopia and Peru, found no association between membership and women's health. 25,36 Violence against women Impacts on interpersonal violence against Bangladeshi or South African women were reported in seven evaluations. The highest quality study found that, after two years, levels of such violence decreased in all four study villages covered by the South African Microfinance for AIDS and Gender Equity intervention but stayed the same or increased in the four control villages. 30 Women's membership in the intervention was associated with a reduced risk of exposure to interpersonal violence (aRR: 0.45; 95% CI: 0.23-0.91). 30 Improvements in all nine of the investigated indicators of women's empowerment were observed. 30 Women members had a greater say over household decision-making and felt more able to challenge the acceptability of violence, to expect and receive better treatment from their partners, to leave abusive relationships and to raise public awareness of interpersonal violence in their village. 30 Another study, also assessed as higher quality, measured violence within spousal relationships in Bangladesh. In this study, women who participated in the BRAC or Grameen Bank schemes -and women who were non-members but lived in Grameen Bank villageswere found to be less likely to be beaten by their husbands than women in control villages. 37 In this study, the role of empowerment was not clear and the effect of women's contribution to family support on violence was not significant. 37 Ethnographic data indicated that, in relatively rich households, women's membership of a microfinance scheme may have led to an initial increase in violence as the women's roles and status were redefined and they had Five lower quality studies, all in Bangladesh, gave mixed results. One study found that microfinance participation was associated with a reduction in the likelihood of interpersonal violence against women of 6.8%. 38 In contrast, after controlling for confounders, two studies found no statistically significant association between microcredit participation and current experience of such violence. 24,39 A further lower quality study in Bangladesh found that the better educated women experienced increased exposure to interpersonal violence following membership of a microfinance scheme. This study was poorly adjusted for bias, however. 40 Another study in Bangladesh, that used propensity score matching to construct an appropriate comparison group of non-members, revealed that levels of interpersonal violence did not differ significantly between members and non-members. 42

Nutrition
Impacts on nutrition were reported in seven evaluations, in Bangladesh, Ethiopia, Ghana, India and Peru. The findings were inconclusive. Some studies showed that scheme membership brought nutritional benefits -mainly for the infants and children of membersand others revealed no significant effects.
Of the three higher quality studies, all from Bangladesh, one found that women from villages with any microfinance scheme showed similar increases in their body mass index to women from neighbouring villages without microfinance. 43 In another study, the prevalence of stunting was found to be higher (84.6%) among children of poor non-members than among the children of BRAC members (67.3%) or rich non-members (69.4%). 44 Weight-for-height z-scores of children aged 24-35 months from BRAC households were significantly higher (P < 0.05) than those of their counterparts from control households. 44 The final higher quality study found no significant differences between BRAC households and non-member households in terms of three other indicators of nutritional status in children and women. 45 We included four lower quality studies relating to nutrition in our systematic review:

Well-being and health-care use
One higher quality study evaluated the Indian Self Help Groups scheme and found that membership was associated with significant reductions in emotional stress and significant increases in the use of health care. 49 A beneficial spillover effect was also noted for nonparticipants who lived in a household with a member. No associations were found between participation and self-assessed health or exposure to health risks. This study excluded a socially marginalized group of women -i.e. Paniya women -because they were considered "prone to underestimate their health". Women members used loans to help cover their health expenditures. 50 Two lower quality studies in Bangladesh revealed associations between microfinance membership and increases in emotional stress -but only for non-members in households that received loans 51 -and use of maternal delivery care. 41 A final study from Peru found that length of participation in a microfinance scheme had no association with women's access to cancer screening or their number of sick days. 25

Discussion
Although we searched for evaluations of group-based microfinance schemes that covered any disadvantaged group in any country, all identified studies that met our inclusion criteria were concerned with the empowerment of poor women in low-or middle-income countries, mainly in Asia.
There is clear evidence of improvements in some important maternal and child health outcomes associated with membership of the long-established BRAC and Grameen Bank microfinance schemes in Bangladesh, including better child survival. 21,22 and use of contraceptives. 31 Results from studies, in a range of countries, on nutritional status and the general health of women who were members of schemes were equivocal. Membership of a microfinance scheme specifically for female sex workers in India was associated with decreases in sexually transmitted infections and increases in condom use during paid sex. 33 A complex picture emerges for the impact of microfinance on interpersonal violence. The evidence indicates that, while microfinance may eventually lead to a reduction in such violence, an initial increase may occur as gender norms are challenged. The most robust study, a cluster-randomized controlled trial, showed that microfinance schemes can reduce the risk of physical or sexual violence by an intimate partner. 30 The few included studies that measured aspects of empowerment generally found improvements in empowerment associated with membership of the major schemes 30-32,46,48though these did not necessarily translate into improved health outcomes. 32 Others have warned that the provision of credit to women does not guarantee their control over the credit's use and may lead to excess anxiety over the pressure to pay back loans -diminishing, rather than increasing, any sense of empowerment. 30 The studies with a strong qualitative component provided some of the most convincing evidence of the role of empowerment in the creation of the beneficial effects of microfinance schemes. In the study of the South African Microfinance for AIDS and Gender Equity intervention, participants revealed how reductions in violence resulted from a range of responses -some linked to increasing confidence and empowerment of the women in handling potential flash points. 30 In Bangladesh, microfinance schemes can empower women by strengthening their economic roles, increasing their say over household decision-making and changing community norms. 32 We found no relevant studies that assessed the impact of microfinance schemes on ethnic inequalities in health. Some evidence did emerge on how microfinance schemes might help tackle socioeconomic and gender inequalities in health. Most notably, evaluations in Bangladesh indicated that the BRAC microfinance scheme may help to narrow the inequalities in health between boys and girls and the rich and poor. 35,44 Such schemes may work not only through improvements in the economic status of the mothers who become members but also through cultural changes in the way girls are valued and nurtured, leading to additional gains for poor girls in relation to poor boys.
The potential for a microfinance scheme to have adverse health impacts was largely unexplored in the evaluations included in our review. Although there have been indications of increased violence between intimate partners as the result of the female empowerment promoted by microfinance, the most robust relevant studies have shown overall reductions in such violence, at least in the long term. 30,37 The potential negative health impacts of microfinance schemes as a result of the debt stress associated with the repayment of loans have yet to be investigated in detail. [11][12][13][14]52 Research into the positive and negative impacts of microfinance schemes may be particularly challenging, not least because of the potential for selection bias of various forms. hard to disentangle the role of the empowerment strategy from that of the poverty-reduction component. Measurement of differential impacts by socioeconomic status is also rare but studies in Bangladesh have shown how this could be done and revealed its potential value.
In conclusion, group-based microfinance schemes represent perhaps the largest experiment in collective empowerment in the world to date. These schemes -and their potential impacts on both health and poverty -deserve close scrutiny. In terms of improvements in selected health outcomes, the evidence coming from the larger, longestablished schemes is encouraging. Many questions remain, however, including the scale of the potential for microfinance schemes to do harm. These questions need to be addressed by appropriately designed evaluations that incorporate community-wide assessments of all potential impacts.

Box 1. Microfinance schemes based on collective empowerment
Roughly 5 million poor rural women in Bangladesh are involved in microcredit programmes, most of them associated with the BRAC or Grameen Bank microfinance schemes.
The BRAC scheme is designed for women living in poor and landless households. It involves the formation of women's groups for saving and credit, training and skills development, functional literacy -including legal and social awareness -and technical and marketing support. Money saved by a group is used to make loans to group members to support income-generation activities such as cottage industries and goat rearing. Sometimes these elements are combined with so-called bolt-on public health components such as the promotion of maternal and child health or family planning.
The main aims of the scheme are to reduce women's economic dependence on men, strengthen their positions within their families, draw them into the public sphere and expose them to new ideas and education. The theory is that the scheme may influence health in many different ways -e.g. it may increase demand for family planning services and reduce the social costs of fertility regulation, leading to fewer, healthier children and better maternal health. It may also lead to improvements in the care and nutrition of children and so reduce child mortality in general and, particularly, the high rates recorded among girls.