Empowering caregivers of orphans and vulnerable children in Swaziland

This paper reports on the health and psychosocial impacts of a programme designed to economically empower female caregivers of orphans and vulnerable children (OVC). The results presented are from a cohort of 492 female caregivers who participated in savings groups and were responsible for 887 OVC. The data showed that the caregivers had increased earnings, borrowed more, repaid their loans, and expanded their businesses. Important health impacts were found for both the caregivers and the OVC. Access to health care increased and the proportion of OVC and caregivers who reported being too ill to take part in daily activities decreased. Results showed that WORTH savings group members increased their financial resources and used them to improve the wellbeing of OVC in their care. The economic and social empowerment had a positive impact on child protection, child wellbeing and health.


Introduction
Pact is an international non-governmental organisation that designs and delivers integrated solutions to complex problems in developing countries. This paper reports the results of an economic strengthening intervention for female caregivers of orphans and vulnerable children (OVC) in Swaziland that was designed to empower them and improve their health and the health of the OVC in their care. These findings are from the Rapid and Effective Action Combating HIV/AIDS (REACH) II project funded by the United States Agency for International Development (USAID) which Pact implemented between 2010 and 2015. The project was designed to mitigate the impact of HIV on OVC households (Pact, 2015). REACH II provided women with standardised information about HIV, child protection, and care of OVC, and equipped them with the skills to discuss these with their families.
For more than 20 years, Pact has implemented its evidence-based WORTH methodology for economic strengthening (Mayoux, 2008). WORTH is based on the premise that dependency is not empowering, so WORTH does not provide seed money, matching grants or subsidised interest rates. Instead, WORTH trains women how to save their own money and to invest it for income generation, in combination with capacity development on community banking, micro-enterprise and marketing skills. Under the REACH II project WORTH members also learnt about child protection, care and support, and HIV and AIDS. REACH II used WORTH to strengthen OVC households' economic security, empower caregivers, and improve the care and health of OVC. Economic strengthening was one component of an integrated project; this paper focuses on the effects of caregiver WORTH participation on targeted OVC outcomes.

Background
Poor women in rural Swaziland are responsible for myriad family responsibilities such as sourcing food and water; caring for children, even those who are not their own; the health of family members; and participating broader community initiatives. These women often have limited access to income, credit, employment opportunities and economic power. According to the most recent Swaziland Demographic Health Survey (CSO/Swaziland & Macro International Inc., 2008), about half the women aged 15-49 years are employed compared to 86% of men; 47.9% of households surveyed were headed by a female.
A comparison of orphan data between 2010 and 2014 shows a slight decrease in the proportion of children who lost one or both parents from 23.6% (CSO/Swaziland & UNICEF, 2011) to 20.4% (CSO/Swaziland & UNICEF, 2016). However, the proportion of vulnerable children in this time frame doubled from 29.5% (CSO/Swaziland & UNICEF, 2011) to 60.1% (CSO/Swaziland & UNICEF, 2016). The OVC needs include food, school accessories (uniforms, shoes, and bus fare), clothes, proper shelter and improved access to health care and legal protection. Gender-based violence is rife in Swaziland and statistics show that 48% of women have experienced some form of sexual violence in their lifetime (UNICEF, 2007).
The WORTH intervention was implemented between August 2013 and February 2015 in three tinkhundla (analogous to districts) of the Lubombo (Lubulini, Nkilongo) and Shiselwini (Somntongo) regions of southern Swaziland (Figure 1).
WORTH was introduced as part of an integrated approach to improve the wellbeing of OVC, targeting female caregivers with household economic strengthening and women's empowerment interventions. This component of the REACH II project relied on empowerment workers (EWs) who recruited female OVC caregivers to join the WORTH savings and lending groups. The women participated in activities related to income generation, group savings, prevention of genderbased violence, HIV prevention, and the care and support for the OVC in their households. At weekly meetings, the women saved and lent each other money and participated in EW-facilitated discussions on the care and support of children, HIV, establishment of small businesses, profit tracking and marketing. The EWs also monitored the group activities and advised on the establishment and expansion of small businesses. Six EWs were employed by Pact who trained them on the management and support of the groups. The training included the use of mobile technology to report weekly on group activities (total group savings, loans, purposes of loans, number and type of small business established, and number and type of discussions facilitated).
As the WORTH groups were formed, the EWs provided training on saving and lending principles, record keeping, small business development and profit tracking. Each group elected a four-member management committee who received further training on group management, and safe money handling. The EWs introduced and facilitated discussions around child protection, OVC care and support, HIV and AIDS, sexual and reproductive health, gender roles, and gender-based violence prevention and response. Pact prepared discussion guides/lesson plans and held monthly training sessions to ensure standardised delivery of session topics. The EWs' reports were reviewed to monitor the uptake of loans and proportions of women establishing small businesses and monitoring their profitability.

Objectives
The theory driving the intervention was that WORTH groups would increase access to savings, loans, and the development of profitable micro-enterprises for OVC caregivers, and simultaneously be a platform to discuss the health, psychosocial support, protection, education, and sexual reproductive health needs of OVC, with both the economic and educational aspects combining to improve OVC outcomes. The increased household economic security would empower the women with funds to help meet the needs of OVCs while the training, learning dialogues, and group discussions would increase their agency for decision making and taking. The programme established at baseline the WORTH group members' savings and the number and type of micro-enterprises they owned, as well as their perceptions of and knowledge about child-care, and gender norms that contribute to gender-based violence. Data were collected on members' demographic, socio-economic status, health, education, and incidence of abuse and psychosocial characteristics of OVC under their care to compare against end-line values.

Study design
The design followed a cohort of women who were active WORTH members for at least 12 months, along with OVCs in their care. The study longitudinally followed their progress at baseline and end-line, collecting data from WORTH members and OVC over age 10 years.

Sampling
Purposive sampling was used to select the women who had participated in the WORTH groups for at least 12 months in the 3 tinkhundla where the WORTH component of the REACH II project was implemented. A total of 36 WORTH groups were formed, but only 34 existed for at least 12 months. Data were therefore collected from the members of these groups. Complete data were collected for 492 women at baseline and 494 at end-line, and from 462 children between the ages of 10 and 17 years.

Data collection methods
Both baseline and end-line data were collected using a questionnaire that was uploaded on mobile devices using the Mobenzi researcher (www.mobenzi.com/researcher) mobile technology platform for data collection. The empowerment workers administered the surveys during group meetings. Each group member was interviewed separately in a secluded area to maintain confidentiality of responses. Responding to questions was voluntary and respondents were allowed to refuse to answer any or all questions. The same data collected at baseline were also collected at end-line. Data were collected on group/individual savings, loans taken and purposes thereof, number of income generating activities initiated and total profit realised, and services provided to OVC. WORTH members were asked whether the project had brought economic change to them, in their families and in their communities, and if and how that change had made a difference to the way they provide care and support to children in their care. To triangulate information on OVC wellbeing, questionnaires were also administered to children older than 10 years at the time of data collection whose caregivers had participated in WORTH for at least 12 months. All data collection questionnaires were translated to Siswati, the local language, to ensure consistency.

Ethical considerations and assurances
The WORTH baseline and end-line data were collected during routine WORTH programmatic data collection in accordance with Pact's ethics policy and guidelines (Pact, 2014) that require confidentiality and respect of individuals' data and adherence to do no harm principles. The data were collected as part of routine monitoring and evaluation activities, not as a research study. However, informed consent was secured from all participants before data collection exercises. Respondents were aware that their participation was voluntary and they could choose to not answer any questions asked of them. All cases of reported abuse were referred for post-abuse care and the project ensured that OVC received those services, and counselling was offered to the caregivers. The aggregated data and summary findings were shared with the empowerment workers and WORTH group members for their learning and use.

Data analysis
The data were analysed using SPSS V.21 software for data analysis. The data were first exported from Mobenzi to Microsoft excel for checking and cleaning. Arc GIS was used to create the map of the geographical coverage of the WORTH intervention. In preparation for data analysis, the team created dummy tables (analysis frame) to determine the focus of the final report and prioritise data for analysis. Key variables included savings, loans uptake, number and type of micro-enterprise, and OVC wellbeing.
The team used SPSS to calculate z-score for selected variables and determine the standard deviations. The data that fell outside of three standard deviation points were considered outliers and removed from the analysis. T-tests (independent samples and paired samples) were conducted for selected variables to determine the significance of the results observed between baseline and end-line. Before the data were subjected to paired samples t-test, outliers were identified and removed by calculating z-scores for each variable and marking any responses that fell off the range as missing information so as to remove skewness of the data. In addition to the t-Test, a Wilcoxon signed-rank test was conducted to determine the significance of results between baseline and end-line. This is the non-parametric test that does not assume normality in the data and can be used to compare two sets of scores that come from the same participants (e.g., pre-and post-intervention scores).

Results
Most WORTH members were over age 18, with just over 10% under that age (as adolescent heads of their households). Table 1 shows the ages and marital status of the caregivers. Figure 2 depicts the respondents' level of education. Overall, most of the women had a primary education with a higher proportion having attended secondary school in Nkilongo than in Lubulini and Somntongo.

Economic outcomes
One of the pillars of WORTH is to create a savings culture where members can save with a particular focus, for example, for children's education, health care, or buying economic assets. In the context of AIDS-affected families in Swaziland, savings are a means of contingency planning to ensure that OVC families are resilient in the event of life stressors. WORTH groups establish two types of savings for members, mandatory (to generate group funds for loans) and voluntary savings (for upcoming large expenses or for contingencies). All members participating had to contribute a mandatory weekly savings of 2 to 10 Swaziland emalangeni (15 to 77 US cents) per week. The exchange rate was E12.95 to US$1.
The data show that almost all WORTH members (99.2%) had E100 or less in savings at baseline. However at end-line, the number of women with E100 or less in savings decreased to 64.9% at end-line due to increased levels of savings. After at least 1 year of participation in WORTH, 14% of members had saved between E101 and E500, 7% had saved E501-1 000, and 2% had more than E1 000 in voluntary savings (Figure 3).

Loans and borrowing
Almost all (99%) the WORTH members reported taking out a loan from the group with nearly half of these (48%) borrowing less than E500 (US$38.50). Details on the loans and repayment by constituency are shown in Table 2. Repayment of the loans was done on time or with small amounts in arrears.
The loans were primarily used to set up small businesses although this was not the case in Nkilongo where loans were taken for other purposes (Figure 4). These included paying for food, health care, education, other business costs, funeral expenses and birth registration.

Establishment and expansion of micro-enterprises
Approximately one-third of the WORTH members who did not already have a small business started one, and among those that already had one, 41.2% either expanded it and/or started one or more additional businesses.
WORTH members improved their economic wellbeing over the course of the intervention, with a statistically significant increase in economic wellbeing at end-line (median monthly income = E450) compared to baseline (p = 0.001). In other words a woman could begin to budget her own income -an experience that some women did not have. A typical budget for a rural woman earning E450.00 (US$35) could include the daily basics like bread, sugar,  potatoes and meat for the family to eat. With this amount she could also afford to pay for transport to take a child to the clinic and to pay for clinic consultation fees, and start saving to buy school shoes and uniform. Figure 5 provides a summary of revenues from WORTH micro-enterprises in the 6 months preceding the end-line survey.
The results show that WORTH members were able to improve household cooking practices, water source, use of electricity, and ownership of household assets in a short time frame (12 to 18 months).

Cooking energy source
The main source of cooking energy at both baseline and end-line was firewood. However, there was a slight reduction in the number of members who reported the use of firewood from 79.5% in the baseline to 74.9% in the end-line. The proportion of respondents who reported cooking with electricity increased slightly from 14.9% at baseline to 18.6% at end-line.

Water source
There was a change in the source of water with an increase in the proportion of communal tap users from 27.6% to 33.2%, and a small increase in home tap users (from 7.8% to 8.3%). The baseline recorded 9.1% of respondents had a home tank as a water source compared to 13.8% at end-line. Members who reported using rivers as source of water reduced from 14.7% at baseline to 9.3% at end-line.

Homestead improvements
When asked about improvements made in the homestead in the previous 3 months, the results showed that there was an increase in the number of people who renovated or connected their houses to electricity from 15.5% at baseline to 18.6% at end-line. Moreover, 36.4% of members bought agricultural inputs implements and 19.8% reported buying household assets.

WORTH groups' social and health outcomes
The goal of the WORTH intervention is to improve the quality of life for children cared for by WORTH members through the economic and social gains of their caregivers. In addition to the benefits described above, the children benefitted in terms of child protection, education, psychosocial support, and health. The data reported in this paper were collected from children (10-17 years old) and do not include data collected from caregivers on behalf of children in their care.

Health
Children whose caregivers participated in WORTH were asked whether they have been too sick (bed ridden for at least 3 days) to participate in their daily activities since the beginning of the school term (previous 3 months) and if so whether they sought care at a healthcare facility. The percentage of children reporting illness dropped from 14.2% (52/367) to 5% (15/303) between baseline and end-line. An independent samples t-test showed that the difference from baseline to end-line was statistically significant (p < 0.001).
The rate at which these children utilised health care when ill remained the same (baseline = 78.8% and end-line 73.3%). The significant decrease in the number of children who reported being too sick to participate in daily activities suggests that caregiver participation in WORTH had a positive outcome on reducing child morbidity ( Figure 6).

Psychosocial support
The study found significant improvements in psychosocial wellbeing from baseline to end-line. The proportion of children who reported always feeling confident increased from 46.8% to 61.4%. Similarly, children reporting that there is an adult who gives them attention rose from 47.2% (baseline) to 68.4% (end-line). Figure 7 shows the results for children's reported confidence and having an adult who gives attention. Stigma and discrimination are an unfortunate reality for OVC in Swaziland. The WORTH intervention therefore included inputs to improve the psychosocial wellbeing among OVC in WORTH members' care; the results are summarised in Table 3.
An independent samples t-test with equal variances not assumed was conducted for the four indicators in Table 3 to compare baseline and end-line. The results show that there was statistically significant reduction in children who felt marginalised and discriminated against (felt that they are treated with less importance, children reporting people gossip about them, people acting afraid of them, and people saying offensive things to them). These results suggest that caregiver participation in WORTH helped reduce children's feelings of being stigmatised and discriminated against, improving their psychosocial wellbeing.

Child protection
Results found a significant reduction in the number of children who reported abuse in the last 6 months. Only 6 children (1.2%) reported abuse in the previous 6 months at end-line, compared to 31 (6.0%) at baseline. A paired samples t-test was conducted to compare these results and for children who reported experiencing abuse in the last 6 months at baseline (p < 0.00) and shows a statistically significant reduction in the number of children reporting abuse at end-line as compared to baseline (0.063). As noted above, all children reporting abuse were referred to and received post-abuse services and caregivers were offered counselling. We found a very slight increase in this short time frame of the number of children who possessed birth registration certificates (from 96.3% to 96.9%) and a slight decrease in national ID from 3.7% to 3.1% at baseline and end-line (Table 4).

Education
We did not find a significant difference between children's school enrolment at baseline and end-line suggesting that participation in WORTH did not have a significant effect on this measure during the period of implementation. However, there was a significant reduction in the number of children reporting absenteeism from school (n = 180 at baseline and n = 75 at end-line), suggesting that caregiver participation in WORTH had a positive effect on absenteeism for the OVC in their care (Figure 8).

Sexual and reproductive health (SRH)
WORTH members were asked about their ability to negotiate safer sex and their ability/confidence to talk to their children about SRH. The proportion of caregivers who "strongly disagreed" that they could talk with their children about SRH decreased from 5.2% at baseline to 2.0% at end-line. This was accompanied by a corresponding increase in the proportion who "strongly agreed" that they could talk with their children about SRH from 13.3% at baseline to 30.8% at end-line. Figure 9 summarises these results for SRH at baseline and end-line.
The percentage of women who strongly agreed that they were better off than their peers in the community increased sharply from 2.6% at baseline to 21.7% at end-line. Similarly, the proportion of women who strongly agreed that they were better off now also increased from 4.2% at baseline to 36.0% at end-line.

Profitability analysis
The empowerment workers collected data about the WORTH members and profits from their micro-enterprises to target support to members whose businesses were losing money. The profitability analysis used the income generated from the business and expenditures on inputs to the business to determine whether the caregivers were making a profit or losing money. The initial analysis determined that 8% of WORTH members were incurring a loss in their businesses and the project team provided tailored support to those specific caregivers. This resulted in an improvement  one year later when 98% of businesses were found to be making a profit.
To better understand the profitability of WORTH related micro-enterprises, the data were further analysed to determine a profit range for micro-enterprises that were making a profit. The results show that just over half (54%) of WORTH members made a monthly profit of E101-500 while 25% had profits ranging between E500 and E1 000 per month. Another 13% of WORTH members made a monthly profit ranging from E1 to E100, while 8% of members made a profit of more than E1 000 per month ( Figure 10).

Increased proportions of OVC receiving protection and psychosocial support services
WORTH membership contributed to improved child protection, decreased child morbidity and higher school attendance. Stigma reduction is one of the major milestones of WORTH in improving child protection, with children reporting feeling less stigmatised, that they are treated with importance, and that they have an adult who gives them attention. These results highlight the potential for WORTH to produce positive outcomes in child protection while also serving as a vehicle for increasing household income and ability to meet the basic family needs.

Increased number of micro-enterprises and savings among WORTH members
Positive outcomes were observed in the establishment of micro-enterprises among WORTH members. One-third (33.4%) of WORTH members who did not have a business at the beginning of the programme were able to establish one, and 41% of the sample were able to expand from one business to two (or more). This is an especially important achievement because maintaining a profitable business requires a level of business acumen beyond the ability to establish a business. The fact that WORTH members were able to expand from one to two or more businesses means that the investment in WORTH support through empowerment workers and project team empowered WORTH women to expand and diversify income generating activities. The increases in earnings from the micro-enterprises and savings and improved profitability combine to result in increased expenditure on basic services to children.

Increased proportion of women who have positive and empowering perceptions on gender and SRH issues
Women who participated in WORTH were able to negotiate safer sex which is key to reducing their vulnerability to HIV infection. They are better equipped to talk with the children in their care about SRH, a critical component in addressing HIV vulnerabilities for young people. This is an example of how complementary efforts across families, development programmes and government services combine to ensure that adolescents are receiving consistent accurate information and services that prevent them from the risks of HIV infection. Women who participated in WORTH reported feeling better off compared to their peers and compared to their lives before joining WORTH. This is an indication of the value that women assign to their participation in the WORTH intervention.

Conclusion
The intervention led to positive results for most women who participated in WORTH groups. WORTH members were able to both increase their income and use it to improve the wellbeing of children in their care, thereby demonstrating both economic and social empowerment. The notable progress towards child protection outcomes in addition to improved household economic status indicates that the WORTH intervention met its objectives.
Based on the results and discussion presented in this paper, three findings and modifications for future economic empowerment interventions in HIV and AIDS programming emerged most vividly. Pact has adapted its programming to incorporate them, in Swaziland and in other countries.
• The WORTH intervention showed improvements in parenting behaviours and skills; adding a specific parenting skills component to WORTH groups would yield even stronger results. • WORTH was an effective strategy to increase access to health care for children; therefore future WORTH programming should explore how to incorporate families affected by HIV to relieve the burden of care and health expenditure in these families. • The WORTH intervention shows positive trends towards improving school attendance for children; future programmes should also focus on progression to the next level of education to determine WORTH's contribution to improving children's access to education.