Article Text

Randomised controlled trial of a livestock productive asset transfer programme to improve economic and health outcomes and reduce intimate partner violence in a postconflict setting
1. Nancy Glass1,
2. Nancy A Perrin1,2,
3. Anjalee Kohli1,
4. Jacquelyn Campbell1,
5. Mitima Mpanano Remy3
1. 1Johns Hopkins University School of Nursing, Baltimore, Maryland, USA
2. 2Center for Health Research, Kaiser Permanente Northwest, Portland, Oregon, USA
3. 3Programme d'Appui aux Initiatives Economiques (PAIDEK), Bukavu, Democratic Republic of Congo
1. Correspondence to Dr Nancy Glass; nglass1{at}jhu.edu

## Abstract

Background Diverse economic empowerment programmes (eg, microcredit, village-led savings and loan, cash and productive asset transfers) for the poor have demonstrated mixed results as vehicles for improved economic stability, health and women's empowerment. However, limited rigorous evaluations exist on the impact of financial and non-financial outcomes of these programmes, especially in conflict-affected areas.

Methods The team evaluated the effectiveness of an innovative livestock productive asset transfer intervention—Pigs for Peace (PFP)—on economic, health and women's empowerment outcomes with participants in households in 10 villages in Eastern Democratic Republic of Congo. Residual change analysis was used to examine the amount of change from baseline to 18 months between the intervention and delayed control groups, controlling for baseline scores.

Findings The majority of the 833 household participants were women (84%), 25 years of age or older, married, had on average 3 children and had never attended school. At 18 months postbaseline, the number of participants in the PFP households having outstanding credit/loans was 24.7% lower than households in the control group (p=0.028), and they had an 8.2% greater improvement in subjective health (p=0.026), a 57.1% greater reduction in symptoms of anxiety (p=0.020) and a 5.7% greater improvement in symptoms of post-traumatic stress disorder (p<−0.001). At 18 months postbaseline, partnered women and men reported a reduction in experience and perpetration of all forms of intimate partner violence, although not statistically significant between groups.

Interpretation The findings support scalability of a livestock productive asset transfer programme in rural and conflict-affected settings where residents have extremely limited access to financial institutions or credit programmes, health or social services and where social norms that sustain gender inequality are strong.

Trial registration number NCT02008708.

This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

### Key questions

• Rigorous evaluations of microcredit programmes are limited and those that do exist provide mixed evidence on effectiveness for improving economic, health and women's empowerment outcomes.

• Productive asset transfer programmes (eg, livestock transfer) with comprehensive services can increase consumption and income for participating households.

• Limited evidence exists to guide development and evaluation of microcredit or productive asset transfer programmes in conflict-affected settings.

#### What are the new findings?

• An innovative productive asset transfer programme, Pigs for Peace (PFP), increased economic stability, improved subjective health and mental health in conflict-affected villages. Partnered men and women reported a reduction in perpetration and victimisation in all forms of intimate partner violence, although not significantly different from the control group.

• PFP has the potential to contribute to the achievement of the Sustainable Development Goals, through reducing poverty, ensuring health and achieving gender equity.

• PFP demonstrates the importance of partnerships with local expertise to transition from humanitarian ‘granting’ to household ‘investing’ for development.

#### Recommendations for policy

• Pigs for Peace (PFP) is a promising programme for improving economic stability, health and women's empowerment with hard-to-reach and underserved communities.

• Collaboration with established gender and health programmes could further develop the PFP programme and advance outcomes for participating households.

• Additional research is needed on adapting the PFP programme with diverse populations to insure the acceptability and scalability.

## Introduction

### Ethics statement

The Institutional Review Board of the Johns Hopkins Medical Institute (JHMI) approved the study on 18 November 2010 (NA_00044037) and a committee of respected Congolese educators at the Universite Catholique at Bukavu (UCB) and community members reviewed and approved the study. Interviews were initiated only after receiving oral, voluntary informed consent from the participant. Oral consent was approved during ethics committee review as the majority of our participants had never attended school, so written consent was perceived as a significant challenge and potential barrier to participation. Inclusion of eligible minors (16–17 years old) was also approved during the ethics committee review. Study identification codes and names were recorded during one-on-one interviews; all data recorded through the tablet-based program were encrypted and uploaded to a password-protected and HIPAA-certified server managed by the study team. Once uploaded, data were automatically erased from the tablet-based program. Names were centrally removed and stored in a separate file on a password-protected study computer.

### Study questionnaire

#### Demographics and household wealth

Our questionnaire was developed using validated items from previous studies, including the Intervention with Microfinance for AIDS and Gender Equity (IMAGE) study team in South Africa16 and the WHO Domestic Violence and Health (2005) study.36 We collected current demographic information from the participant on his/her marital status, educational level, regular work (yes/no), perceived household wealth in comparison to other households in the village (ie, 1=worse than others, 2=same as others, 3=better than others), dwelling details (yes/no for durable housing defined as roof made of tin, walls of wood/brick) and household savings (yes/no). We also asked participants to report on other adults and children living in the household by age and gender.

#### Economic stability and livestock/animal assets

Economic stability was measured by the number of cash and non-cash loans a participant had in the 12 months prior to baseline interview and the months prior (∼6 months) to the follow-up time point. This was dichotomised into none versus one or more loans. A total livestock/animal asset score was computed for each participant based on the number and type of animals owned to establish the value of each type of animal. The team surveyed nine livestock/animal vendors in five different village markets in the study area and collected the current price to purchase the most commonly owned livestock/animal assets. The average cost in US dollars are cows $450, pigs$70, goats $50, poultry$10, rabbits $8 and guinea pigs$1. We computed a total livestock/animal asset score for each participant's household by multiplying the average market price for the livestock/animal by the number of household livestock/animals reported at the baseline and 18-month follow-up interview.13 Since these scores were extremely skewed, they were recoded into quintiles based on the baseline distribution and the ordinal quintile scores were used in the analysis.

#### Traumatic events, subjective health and mental health

The exposure to trauma events section of the questionnaire was adapted from the Harvard Trauma Questionnaire (HTQ), a multipart cross-culturally validated instrument that measures traumatic events and PTSD37 that the team had previously used in the study setting.13 ,38 Exposure to trauma was analysed as a continuous variable (0–18 different traumatic events). A 16-item version of section 4 of the HTQ39 was used to identify symptoms consistent with PTSD in the past 7 days. Subjective health was measured with one item, rating health from poor to excellent in past 30 days. The depression and anxiety components of the Hopkins Symptom Checklist (HSCL) were used for reporting the experience of symptoms that bothered or distressed the respondent during the past 1 month.39 An average symptom score for PTSD, depression and anxiety was calculated. The HTQ and HSCL have been used widely in conflict-affected and humanitarian emergencies and both have strong psychometric properties for measuring of traumatic events and symptoms consistent with PTSD and depression in conflict-affected settings.40–42 In this sample, Cronbach's α was 0.86 for anxiety, 0.85 for depression and 0.89 for PTSD.

#### Intimate partner violence

Women were asked about psychological abuse, physical and/or sexual violence perpetrated by their male partners; men were asked about their perpetration of IPV against their female partners. The items asked about the partner included: (1) humiliating, (2) hurting, (3) insulting, (4) pushing, (5) slapping, (6) twisting arm or pulling hair, (7) punching, (8) kicking, dragging or beating, (9) choking or burning, (10) threatening or attacking with a weapon, (11) forcing to have sexual intercourse and (12) forcing to perform other sexual acts. Binary variables were created indicating any experience/perpetration of each of the following types of IPV: psychological abuse (items 1–3), physical violence (items 4–10) and sexual violence (items 11 and 12).

Statistical power: We used data from Roberts et al43 study of the reliability and validity of the SF-8 with a conflict-affected population in northern Uganda as the basis for our power analyses. Since we did not have an estimate of the intraclass correlation (ICC) (individuals nested within villages), we varied the ICC from 0.001 to 0.010. For a sample size of 300 per group and assuming no change from baseline to 18 months in the control group and a 10% improvement in the intervention group, a α level of 0.05, the power is 0.94, 0.89 and 0.83 for ICCs of 0.001, 0.005 and 0.010, respectively.

### Statistical analysis

For the analyses, the intervention group (N=309) was compared with the first delayed control group (N=296) and the second delayed control group (N=228). All analyses used a generalised estimating equation approach to control for the clustering of participants within villages. The intervention and control groups were compared on baseline variables using a Gaussian distribution and identity link function for continuous variables. Those with and without missing data were compared using the same approach to determine if there were baseline factors related to missingness. The analyses were based on intention to treat with all participants included in the main analyses. Multiple imputation (with 10 imputed data sets) was used to replace missing data based on recommendations of Schafer.44 As a sensitivity analyses, we compared the results from multiple imputations with a completers only analysis.45 The main GEE model was a residualised change regression to examine the difference in continuous variables in the amount of change from baseline to 18 months between the intervention and control groups, controlling for baseline scores on the outcome. Change score on the outcome was the dependent variable predicted from a dummy variable for group (intervention vs control) and baseline score on the outcome. The residualised change model was selected over ANCOVA model because the former accounts for differences in the outcome at baseline and the latter assumes differences between the groups at baseline. The analysis accounts for any differences at baseline because the differences may affect the degree of change.

### Role of the funding source

The funding source did not play a role in the design of the study, data collection, analysis, interpretation or writing of the results. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

## Results

### Sample description

The majority of the 833 household participants were women, 25 years or older, married, had on average 3–4 children in the home and had never attended school. Most participants reported living in homes with non-durable walls (N=740, 89%) and roofs (N=487, 59%). Very few (N=38, 5%) had savings; the vast majority (N=680, 82%) described their household wealth as being the same or worse off than most people in the village. The intervention and control groups were not significantly different on any of these variables at baseline (table 1). The two groups were not significantly different on any of the outcome variables at baseline, with the exception of anxiety. The control group had significantly higher anxiety than the intervention group. Lost to 18-month follow-up was not significantly different between the groups; 13.0% (N=68) of the control group and 16.2% (N=50) of the intervention group did not complete the interview (figure 1). Men (N=36, 27.7%) were more likely than women (N=82, 11.7%, p<0.00001) to not complete the 18-month interview. Completers and non-completers did not differ on age, marital status, schooling, perceived wealth, subjective health, PTSD, IPV, livestock/animal assets or having loans at baseline. Those who were lost to follow-up had significantly lower anxiety (p<0.001) and depression (p=0.015). In addition, there were no differences between the intervention and control group non-completers on the baseline characteristics.

Table 1

Demographics by condition

Figure 1

Trial profile.

Table 2 summarises the baseline and 18-month outcomes by group.

Table 2

Baseline and 18-month means (SD), parameter estimate from the residualised change analyses and associated effect sizes

Economic status. At 18 months postbaseline, the participants in PFP reported significantly greater increase in household livestock/animal assets than the control group (p=0.00004), controlling for assets at baseline. Participants in PFP were significantly less likely to have one or more loans (received as cash loan or inkind) (p=0.028) than control group participants, controlling for whether or not they had one or more loans at baseline. The same pattern of results was found with completers only. Participants reported that loans were typically received from family members or friends, health centres or small businesses in the village, only 1% of participants reported having credit with a traditional microfinance organisation.

Physical and mental health. PFP participants had significantly greater improvement in subjective health (p=0.035), controlling for their baseline subjective health. The intervention group also had greater improvement in symptoms of anxiety (p=0.023) and post-traumatic stress (p=0.0004), but did not differ on change in symptoms of depression (p=0.089). The same pattern of results was found with completers only.

Intimate partner violence. Among men and women who were partnered at baseline and 18 months (N=311 control, N=162 intervention), the groups differed on experiencing/perpetrating psychological abuse (35.1% control, 27.2% intervention; p=0.080) at 18 months although not statistically significant. Further, partnered women and men in the intervention and delayed control groups reported a decrease in experienced/perpetrated physical and sexual violence, the groups did not differ significantly on physical (p=0.340) or sexual violence (p=0.503) at 18 months. Importantly, the study was powered for the main outcomes using the entire sample (N=833); therefore, the analyses for IPV among those married at baseline and 18 months (N=473) are underpowered.

## Discussion

The study findings confirm the hypotheses that participants in PFP households would have increased economic stability and improved subjective health and mental health compared with participants in delayed control households in rural, conflict-affected villages. In rural DRC, like rural communities globally, animal husbandry continues to be one of the few opportunities for economic stability as livestock are productive assets to accumulate to rebuild household wealth and social status.29 ,35 Livestock is a visible symbol of wealth, productivity and social status to the extended family and larger community. Livestock possession and productivity influences the owners' positive perception of self and household wealth.13 ,28 ,29 The local implementing partner, PAIDEK, was essential in identifying the productive asset that would result in improved economic stability and engaging men and women in the programme, as cooperation and shared decision-making was viewed as critical to success.29 ,46 ,47 For example, cows and goats were not selected as the productive asset because women cannot sell cows or goats without consent from the husband or male member of the household, as these animals are tied to the dowry system. As is tradition among the Shi people, the majority tribe in the study area, the future husband's family provides one cow to the future wives family. In recent years, with the loss of livestock wealth in rural areas, goats are often used for the dowry. Further, important in selecting the productive asset is that the vast majority of residents in the target area are either Catholic or Protestant and pork/pork products are produced and regularly consumed. Thus, the pig was the productive asset that represented a gender-neutral intervention to bring husbands and wives and other family members together in income-generating activities to improve economic stability for the household.

The findings demonstrated improved subjective health and a reduction in symptoms associated with poor mental health. Women and men in this area of Eastern DRC have experienced significant trauma over a prolonged period, resulting in symptoms of PTSD, depression and anxiety that can negatively impact productivity and family relationships.23 ,48–50 PFP may have reduced these negative mental health symptoms by limiting stress through increased livestock/animals assets and less cash or inkind credit with family, friends and others.13 There have been other successful and innovative efforts to address unmet mental health needs through skilled healthcare. For example, Bass et al48 conducted a study with female sexual violence survivors in Eastern DRC to examine the effectiveness of an adaptation of group cognitive processing therapy (CPT) provided by community-based psychosocial assistants supervised by psychosocial staff at an international NGO and US-based clinical experts. The findings indicate that psychosocial assistants with appropriate training and supervision can implement psychotherapeutic treatments such as CPT and improve mental health for women. Our findings build on this work by demonstrating positive mental health outcomes, reduced symptoms of PTSD and anxiety for men and women participants. This is an important finding as the intervention is effective with male and female participants that had experienced multiple and diverse forms of traumatic events, beyond sexual violence. Further, it provides an example of a potentially sustainable economic programme led by village associations that has the added benefit of reducing mental health symptoms in settings that have extremely limited infrastructure and capacity to provide mental healthcare. In Eastern DRC, as in many low-resource countries, there is a lack of government-funded health centres with a workforce that has training in mental healthcare. It is estimated that DRC has 0.07 psychiatrists working in the mental health sector per 100 000 population.51

Although PFP does not include a women's empowerment component, our staff emphasised the importance of communication and shared decision-making between husbands and wives in the programme. Analyses of IPV are based on those who were partnered at baseline and 18 months (56.8% of the total sample) and do not have adequate statistical power; however, the pattern of reduction of IPV is clear. At 18 months postbaseline, fewer participants in the intervention group reported experiencing/perpetrating psychological abuse than the control group (27.2% intervention, 35.1% control; p=0.080). Further, reductions in physical (3.5%) and sexual (10.6%) IPV were reported by partnered women and men in the intervention group; however, the reduction was not significantly different between intervention and control group participants. Research has suggested that in conflict-affected populations, men express a need to recover their authority and role as head of household, despite the severe economic and health stress on the family.34 ,52 With the DRC national prevalence of past year frequent (sometimes or often) physical, sexual and/or psychological IPV of 43.9%,53 identifying effective approaches to IPV prevention is critical to sustained development. Women's report of decreased IPV by their husband/partner was consistent with men's report of reduced use of IPV over the 18-month period. Indepth qualitative interviews were conducted at ∼6–9 months postbaseline with married/partnered male and female PFP participants that reported IPV perpetration or victimisation at baseline. The indepth interviews added to our understanding of risk factors associated with IPV in participating households. Men and women described financial stress including lack of work for men outside the home, alcohol use, male peer group sanctions use of IPV and social norms that support a husband's role in disciplining his wife as risk factors for husband's use of IPV.54 Future PFP programming will include primary prevention of IPV through engaging men and women in changing social norms that sustain gender inequality and reducing the multiple risk factors identified in our work and others that may provide additional reductions in IPV and enhance women's empowerment outcomes.

PFP prioritised the focus on transitioning from humanitarian ‘granting’ to household ‘investing’ for sustainable development. The success of PFP provides support for the importance of indigenous expertise in sustainable development programmes to improve economic stability, subjective health and mental health and reduce IPV in rural households. Our implementing partner had the expertise and access to engage traditional and administrative village leaders in productive and culturally appropriate economic activities that supported the participation of men and women.27 Partnerships have the benefit of also building local capacity to provide economic and other development initiatives, which is a critical step to ending a dependence on humanitarian aid and progress to sustainable development that will advance wealth, health and gender equality. It is certain that credit and productive asset transfer programmes alone will not solve the multiple challenges facing families in conflict-affected settings. However, a collaboratively developed and culturally relevant economic development programme that has the benefit of improved health and women's empowerment has potential for advancing SDG.

This study has limitations. The study was conducted in 10 conflict-affected rural villages in 1 province in Eastern DRC. Therefore, the experiences of the male and female participants are not generalisable to all rural households experiencing conflict. Participants reported exposure to multiple traumatic events within the past 10 years, representing at least two periods of conflict; therefore, some of the reported traumatic experiences were likely in the recent past and others several years prior to the baseline interview, so recall bias is an issue. Further, given the limitations of resources, our PFP programme staff was also trained as research assistants and participated in data collection and supervised interns that conducted interviews with participants across the 18 months. We also acknowledge the potential for contamination between the groups, as the delayed control group households were in the same villages as the intervention households.

## Conclusions

PFP has important implications for achieving the SDGs with positive findings that intersect areas critical to sustainable development—economic stability, improved subjective health and mental health and reduced violence against women. PFP has potential for scalability, given that it was successful in a challenging rural and conflict-affected setting where residents have extremely limited access to financial institutions or credit programmes, health or social services and where social norms that sustain gender inequality are strong.

## Acknowledgments

The authors thank all the members of PFP for their participation and willingness to spend their valuable time in responding to our study. The authors also thank our Congolese team in the field, Luhazi Banywesize, Alfred Backikenge Mirindi, Jean Heri Banywesize, Clovis Murhula Mitima, Arsène Kajabika Binkurhorhwa, Eric Mitima Ntwali, Gisele Ntakwinja Mushengezi, Gracia Kindja, Sora and Nadine Mwinja Bufole. Their expertise, desire and commitment to rebuilding their communities are what made the study possible.

## Footnotes

• Handling editor Seye Abimbola.

• Contributors NG, NAP and MMR designed the study, participated in development of data collection measures and analysis, led data interpretation, drafted the manuscript and obtained funding. AK managed the data entry, data collection for baseline and 18 months, and participated in the interpretation of the data and writing of the manuscript. JC participated in the design of the study, data interpretation and writing of the manuscript.

• Funding National Institues of Health/National Institute of Minority Health and Health Disparities, R01 MD006075.

• Competing interests Pigs for Peace was cofounded and developed by MMR of PAIDEK microfinance, DRC, and NG, Johns Hopkins School of Nursing, USA. MMR supervised the Congolese-based implementation of PFP microfinance. The Hopkins team collaborated with Congolese team on all components of the evaluation, but was not directly involved in data collection. No other authors have any conflict of interest to declare.

• Ethics approval Johns Hopkins Medical Institutions.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data sharing statement All data supporting this study and supplementary information are available by contacting the corresponding author.