Vulnerable young women and frontline service providers identify options to improve the HIV-sensitivity of social protection programmes in Botswana: A modified Policy Delphi approach

ABSTRACT

Poverty, lack of education and gender inequality make unemployed and out-of-school young women extremely vulnerable to HIV infection. Promotive social protection programmes aim to increase livelihood and capabilities and could empower this priority population to act on HIV prevention choices. In Botswana, they rarely benefit from such programmes.

A modified Policy Delphi engaged a panel of 22 unemployed and out-of-school young women and eight frontline service providers to consider alternative policy and practice options, and tailor available programmes to their own needs and social situation. The panel assessed the desirability and feasibility of improvement proposals and, in a second round, ranked them for relative importance.

Nearly all 40 improvement proposals were considered very desirable and definitely, or possibly, feasible, and panellists prioritised a wide range of proposals. Frontline service providers stressed foundational skills, like life skills and second chance education. Young women preferred options with more immediate benefits. Overall, panellists perceived positive role models for programme delivery, access to land and water, job skills training, and stipends as most important to empower HIV-vulnerable young women. Results suggest ample policy space to make existing social protection programmes in Botswana more inclusive of unemployed and out-of-school young women, hence more HIV-sensitive.

KEYWORDS:

• HIV infection
• poverty
• empowerment
• women
• public policy

Introduction

Despite sustained massive investments to gain epidemiological control over HIV, Botswana’s HIV prevalence remains one of the highest in the world, with up to three times as many young women (15–30 years) affected as young men (Ministry of Health et al., 2022). Poverty, gender inequality and lack of education are structural factors that increase young women’s vulnerability to HIV while reducing their ability to act on HIV prevention choices (Andersson, 2006; STRIVE, 2019). Severe food insecurity (proxy for extreme poverty) is associated with increased sexual risk taking (Weiser et al., 2007) and a twofold increased risk of recent HIV infection (Low et al., 2022). Unemployment predicts women’s disproportionate HIV burden (Austin et al., 2017), but educational attainment (De Neve et al., 2015) and school attendance (Stoner et al., 2017) protect against HIV.

Social protection could address such HIV-vulnerabilities through traditional programmes that transfer resources to prevent or protect against poverty, but also through promotive social protection: programmes that aim to enhance income and capabilities (Devereux & Sabates-Wheeler, 2004). There is no formal definition for HIV-sensitive social protection, but ‘HIV-sensitivity’ implies that social protection programs are (1) inclusive of HIV-vulnerable groups, and HIV-vulnerable groups are (2) meaningfully involved in the (re-)design of programmes to address the multidimensionality of their vulnerability (UNAIDS, 2017). To end AIDS by 2030, UNAIDS Fast-Track Commitment 6 recommended that 75% of persons at risk of, living with, or affected by, HIV infection should benefit from HIV-sensitive social protection by 2020 (UNAIDS, 2016), including from socioeconomic, or promotive programmes (UNAIDS, 2018).

Botswana offers no universal basic income apart from social pensions, but young women might benefit from a range of promotive programs that aim to alleviate poverty while contributing to economic growth (WorldBank, 2022a). These include microenterprise development, microloans, income transfers through public works (Ipelegeng), productive asset transfers (transfers of small stock or chickens), apprenticeships (Tirelo Sechaba) and second chance education. Existing programmes are not HIV-sensitive, as the HIV-vulnerable group of unemployed and out-of-school young women rarely benefit from them (Cockcroft et al., 2018b). Barriers to perceived programme benefit were found at every socioecological level. At a personal level, young women lacked life and job skills. At interpersonal level, care responsibilities, negative peer influence, and lack of support from family and boyfriends prevented success. At community and structural levels, barriers included discriminatory gender norms, poverty, and lack of coordination (van der Wal et al., under review).¹

A systematic review of regional best practices identified which programme elements could help improve the HIV-sensitivity of promotive programmes. It found microgrants, savings, and life – and job skills effective for HIV- and socioeconomic outcome categories, especially in comprehensive packages with supportive mentors and social safe space. Microenterprise development contributed positive socioeconomic but mixed HIV-related outcomes and microcredit was not recommended for this target group (van der Wal et al., 2021). By involving unemployed and out-of-school young women and frontline service providers in the assessment of alternative policy and practice options, this study aims to identify how promotive social protection programmes in Botswana could become more HIV-sensitive.

Methods

We used a modified Policy Delphi to investigate how promotive social protection programmes in Botswana could be made more inclusive of HIV-vulnerable young women. We identified policy and practice proposals that unemployed and out-of-school young women and frontline service providers supported and prioritised, and examined how their views differed.

Policy Delphi

A Policy Delphi offers HIV-vulnerable groups the opportunity to be meaningfully involved in the assessment of programmes or policies. The Policy Delphi is a decision-facilitation tool that generates and assesses divergent, even opposing, viewpoints to inform policy and practice with new or alternative options (Linstone & Turoff, 1975). It accommodates the complexity of social policy (De Loë et al., 2016) as it taps into knowledge from stakeholders purposively selected for their closeness to the issue: those with lived experience, affected by decisions, and/or familiar with the context (Baumann et al., 1982). As such, it permits greater inclusion of viewpoints at the lower end of the power continuum like frontline workers (Hasson et al., 2000) and lay persons (Rowe & Wright, 2011).

The approach engages a panel in an anonymous, multi-round, structured dialogue working toward stability of informed opinion rather than consensus (Belton et al., 2019). There is no standardised protocol, but a generally accepted approach involves an exploratory round of qualitative research or a systematic review to gather ideas (Hasson et al., 2000), followed by two or more rounds to evaluate ideas (De Loë et al., 2016). For resolution of policy issues, panellists assess the desirability, feasibility, and relative importance of ideas, while providing underlying rationales to explore dissensus (Linstone & Turoff, 1975). The iterative process involves returning analysed and aggregated responses from earlier rounds to panellists for continued and more refined assessment (De Loë et al., 2016).

By focusing on solutions, the Policy Delphi considers diverse opinions in a non-adversarial manner (Hasson et al., 2000). This may contribute more nuanced and realistic solutions while supporting social learning (Kezar & Maxey, 2016). The Policy Delphi does not rely on in-person group meetings (Linstone & Turoff, 1975), reducing dominant voice and groupthink while leveraging positive attributes of group interaction like creative synthesis and diversity of opinions (Hsu & Sandford, 2007). Anonymity levels power differentials in a group, as stakeholders contribute opinions that are assessed on merit, hence are considered more equally without fear of embarrassment or retribution (Rowe & Wright, 1999). Counterbalancing these advantages, the process is researcher-driven and isolation of panellists by their anonymity may reduce collaborative learning.

Setting

Botswana is an upper middle-income country, but poverty (56.6%) and unemployment (24.5%) rates are high, disproportionately affecting youth, in particular young women (StatisticsBotswana, 2021; WorldBank, 2022b). This study contributes to the Inter-ministerial Structural Intervention Trial of HIV prevention (INSTRUCT-ISRCTN54878784), which aims to leverage promotive social protection programmes for HIV prevention (Cockcroft et al., 2018a). We conducted this research with panellists from Moshupa sub-district, in particular Moshupa town and eight rural villages, of which four have relatively good geographical connectivity; four were more remote.

Panellists

A Policy Delphi engages 10–50 panellists purposively recruited to reflect divergent views (Linstone & Turoff, 1975). We purposively selected individuals with a direct stake in, experience with, or knowledge of, promotive social protection programmes in Moshupa sub-district. We invited two stakeholder groups: young women who had attended two-day information and empowerment workshops (2015–2017) (Cockcroft et al., 2018a) and frontline service providers (officers) whose current or past duty station was Moshupa sub-district. When initially recruited, the young women had been 18–30 years old, unemployed and out-of-school. Officers included programme officers delivering promotive social protection programmes and technical officers who allocate land and water, as these were prerequisites for application to several programmes.

Data collection

RW collected data in June–October 2021 (Round 1-R1) and in March–June 2022 (Round 2-R2), mostly in person (80%) or by phone. RW collected data with officers who were all fluent in English. MK, a Motswana woman assistant, assisted data collection with young women through translation and follow up. She conducted telephone interviews independently with young women who were unavailable for in-person interviews.

ROUND 1: Candidate proposals (n = 33) for the survey questionnaire were generated in an empirical study of barriers and solutions to promotive social protection programmes in Botswana (van der Wal et al., under review), and a systematic review of HIV-sensitive social protection programmes in East and Southern Africa (van der Wal et al., 2021), both papers focusing on unemployed and out-of-school young women (Appendix 1). After translation into Setswana and back-translation into English, we piloted and adjusted the questionnaire. We grouped proposals into four policy and practice improvement categories: (1) programme; (2) training; (3) outreach; (4) programme delivery and coordination (Appendix 2).

To ensure panellists shared the same baseline knowledge, a 10-minute YouTube video (https://www.youtube.com/watch?v=kY5idSBnLrY) summarised exploratory findings in English and in Setswana (https://www.youtube.com/watch?v=Kjhoru0_s-0). We used email or WhatsApp to share: (1) information about the Policy Delphi (Appendix 3); (2) questionnaires; (3) YouTube videos. Young women with Smartphones preferred using WhatsApp, on which data transfer (1GB) permitted watching the video multiple times. Young women also received paper handouts of the PowerPoint slides used in the videos. Appendix 4 provides detailed data collection and dissemination information.

All R1-data were collected from individual panellists, who watched the videos on an iPad. After respondent validation to confirm exploratory findings, we read out loud each proposal. We asked panellists to rate proposals for desirability and feasibility with 5-point Likert scales (Table 1), to provide underlying rationales and additional improvement proposals. We conducted telephone interviews with panellists who had moved out of Moshupa sub-district or were unavailable for in-person interviews. R1-data collection lasted 1.5–3 h, depending on the need for translation. In-person interviews took place at panellists’ homes or offices. RW aggregated R1-responses per proposal (long summary) and per improvement category (short summary) (Appendices 5 and 6).

Table 1. Definitions for desirability and feasibility.

Desirability – is this proposal going to be beneficial?
4	Very desirable	Extremely beneficial proposal
3	Desirable	Beneficial. This proposal will have a positive effect with little negative effects
2	Undesirable	Will have negative effects
1	Very undesirable	Will have major negative effects. Extremely harmful. Not justifiable
0	No opinion	I do not know what effect this proposal will have
Feasibility – could this proposal be implemented here in … . and with this group of people?
4	Definitely feasible	This proposal can be implemented and is acceptable to the public and policymakers
3	Possibly feasible	This proposal could be implemented but political or public support is not clear
2	Possibly unfeasible	There are many unanswered questions, so this proposal might not work
1	Definitely unfeasible	This proposal cannot be implemented
0	No opinion	I do not know whether it is possible to implement this proposal

Adapted from Linstone & Turoff (1975).

ROUND 2 focused on respondent validation for R1-results, social learning, and ranking for relative importance. R2-panellists commented on summarised findings from R1, ranked two or three priority proposals per improvement category, and provided rationales. RW conducted individual interviews with officers. We organised workshops for young women to leverage collaborative learning. After contributing individual rankings, young women discussed their priorities together, and agreed on group rankings. Workshops lasted 0.5–2 days (Appendix 4), and mostly took place at young women’s homes. We provided refreshments, meals, airtime (USD5), and reimbursed transportation costs. R2-interviews with officers lasted around 1.5 h.

Analysis

RW analysed quantitative data with descriptive statistics. Median and interquartile range described the degree of support and spread in the distribution of desirability and feasibility of improvement proposals; means compared importance ranking between stakeholders. We triangulated quantitative data with qualitative descriptions of rationales. RW thematically analysed qualitative data with Framework, a matrix-based applied qualitative research method (Ritchie & Spencer, 2002). Starting from improvement proposals, the deductive-inductive approach involved line-by-line coding and aggregation into themes and main categories while writing analytical memos. Descriptive statistics and qualitative text were analysed in Microsoft Excel. Respondent validation for R1 occurred in R2; for R2, we received validation after reading back what panellists reported.

Findings

Overview

We invited 25 young women and 12 officers to participate in the Policy Delphi, of which 30 accepted (22 young women, 8 officers). Response rates were 80% overall, R1: 73%, R2: 76%. In R2, we lost one officer and one young woman but gained three young women who had been unable to participate in R1; 24 panellists participated in both rounds (Table 2).

Table 2. Panellist characteristics and programme & service descriptions.

Panel demographics	Number of panellists (n = 30)
	Round 1 (n = 27)		Round 2 (n = 28)
	YW	Officers	YW	Officers
Gender
Female	19	4	21	4
Male		4		3
Employment status
Unemployed	6		7
Location
Within Moshupa sub-district	18	5	17	5
Outside Moshupa sub-district	1	3	4	3
Promotive social protection programmes/services in Botswana
BOCODOL: second chance education for persons without a Junior 3 diploma		1		1
ISPAAD: agricultural inputs (seeds, fertilizer, tools) and subsidies (ploughing, planting); horticulture (part loan/personal capital)		2		2
LIMD: productive asset transfers (goats and chickens) for destitute people*		1		1
S&CD: APP programme: microgrants, microenterprise development, productive asset transfers (goats and chickens) for destitute people*; Ipelegeng: income transfers through public works		1		1
Youth Development Fund (YDF): part loan/grant, productive asset transfers (goats and chickens), youth apprenticeships (Tirelo Sechaba)		1		1
Sub-Land Board: approves land allocation		1
Water Utilities Corporation: approves water allocation		1		1
Ever benefitted from a promotive social protection programme	10	1	10	1
Total n panellists per round	19	8	21	7

YW: young women; S&CD: Social and Community Development; APP: Alternative Package Programme; *destitute people: individual income < BWP120 (US$11); household incomes < BWP150 (US$14).

In R1, panellists added six proposals (in italic) (Table 3). In R2, another added proposal brought the total to 40 proposals. In R1, panellists rated nearly all proposals as ‘very desirable’, two third as ‘definitely feasible’, and one third as ‘possibly feasible’. Panellists offered implementation suggestions while providing rationales for feasibility. In R2, stakeholder agreement of perceived relative importance was also relatively high. Nuances were in officers stressing foundational skills, while young women preferred proposals with more immediate benefits.

Table 3. Desirability & feasibility of improvement proposals by stakeholder group – median and interquartile range.

#	Proposals to improve programmes	Young women Me (IQR)		Officers Me (IQR)
		Desirability	Feasibility	Desirability	Feasibility
1.1	Safe savings	4 (0)	4 (1)	4 (0.25)	3.5 (1)
1.2	Microgrants rather than loans	4 (0)	2.5 (0.5)		3 (0.25)
1.3	Mentors for psychosocial and business support	4 (0)	4 (0.25)	4 (0)	3.5 (1)
1.4	Weekly meetings for professional and life skills training, psychosocial support, and a social network	4 (0)	4 (1)	4 (1)	3 (0.25)
1.5	Childcare	4 (0)	4 (0)	3 (1)	2.5 (1)
1.6	Create local demand for microenterprise products and services	4 (0)	4 (1)	4 (0)	4 (0)
1.7	Help bring products to markets	4 (0)	3 (0.5)	4 (1)	4 (1)
1.8	Establish private sector links through apprenticeships	4 (0)	2 (0)	4 (1)	3 (1.125)
1.9	Establish professional networks YW	4 (0)	3 (0.5)	4 (0.25)	3.5 (1)
1.10	Offer support for at least one year	4 (0)	-	4 (0.25)	3 (1)
1.11	Offer combination packages: cash grants, skills training, safe saving options, HIV/health/gender training, weekly meetings, and mentors	4 (0)	4 (0.5)	4 (1)	3 (1)
1.12	Ask people themselves what they want	-	-	-	-
#	Proposals to improve training	Young women Me (IQR)		Officers Me (IQR)
		Desirability	Feasibility	Desirability	Feasibility
2.1	Life skills training in all programmes	4 (0)	4 (1)	4 (0.25)	3.5 (1)
2.2	Business and financial skills training	4 (0)	-	4 (0)	4 (0.25)
2.3	Vocational job training	4 (0)	4 (0)	4 (0)	4 (0)
2.4	Include training component in Ipelegeng	3 (2)	-	3 (0.5)	3 (1)
2.5	Training in preferred (local) language	4 (0)	-	4 (0)	4 (0)
2.6	Gender awareness training for YW	4 (0)	4 (0.5)	4 (0)	4 (1)
2.7	Gender awareness training for important persons in their lives (mother, father, brothers, boyfriends)	4 (0)	3 (0.5)	4 (1)	4 (1)
2.8	Provide outlets where YW may confidentially share their life experiences	-	-	-	-
2.9	Ask YW what training programmes they would want and bring that to reality	-	-	-	-
#	Proposal to improve outreach	Young women Me (IQR)		Officers Me (IQR)
		Desirability	Feasibility	Desirability	Feasibility
3.1	Promote second chance education	4 (0)	3.5 (0.5)	4 (0)	4 (0)
3.2	Diversify venues for outreach	4 (0)	4 (0)	4 (0.25)	4 (0.25)
3.3	Diversify means for outreach	4 (0)	3.5 (1.25)	4 (0.25)	4 (0.25)
3.4	Involve YW in outreach	4 (0)	4 (1)	4 (1)	4 (0.25)
3.5	Involve boyfriends in outreach to YW	4 (0)	3.5 (1.5)	3.5 (1)	3.5 (1.5)
3.6	Involve communities in outreach to YW	4 (0)	3.5 (2)	4 (0.25)	4 (1)
3.7	Train YW to help peers to fill out applications	4 (0)	4 (0.25)	4 (0)	4 (0)
3.8	Stipends for peer training and programme dissemination	4 (0)	-	4 (0.25)	3 (1)
3.9	Develop a TV programme about YW. A programme that travels all over the country and documents our issues	-	-	-	-
3.10	Use boot camps for several weeks that offer all the different kinds of training	-	-	-	-
#	Proposals to improve programme delivery & coordination	Young women Me (IQR)		Officers Me (IQR)
		Desirability	Feasibility	Desirability	Feasibility
4.1	Performance indicators for programme officers	4 (0)	4 (0)	4 (0)	4 (0)
4.2	Positive role models in programme delivery	4 (0)	4 (0)	4 (0.25)	4 (0)
4.3	Positive role models for agriculture programmes	4 (0)	4 (0.25)	4 (0)	4 (0.25)
4.4	Improve coordination between programmes	4 (0)	4 (0)	4 (0)	3.5 (1)
4.5	Improve coordination between programmes and land and water boards	4 (0)	-	4 (0)	4 (0.25)
4.6	Improve access to land and water by YW	4 (0)	-	4 (0)	4 (1.25)
4.7	Develop national strategy for HIV-sensitive social protection	4 (0)	-	4 (0.25)	3 (1)
4.8	Unemployed and out-of-school YW should be offered leadership roles, like champions.	-	-	-	-

Me: median; IQR: interquartile range; YW: young women; italic: proposals proposed by panellists in Rounds 1 and 2; Ipelegeng: public works programme for income transfers.

Round 1: Perceived desirability & feasibility per improvement category

1. Programme improvement

Proposals centred around (1) improving access to programmes (childcare); (2) improving programmes themselves (grants rather than loans; add a savings component); (3) improving young women’s capabilities (mentors, social networks, private sector apprenticeships, one year programme support); and (4) leveraging local context (by establishing local demand, a professional network, and help to bring products to market).

Including a savings component in programmes was very desirable and, due to a well-established cell phone banking system, definitely feasible. Changing microloans into microgrants was also very desirable. ‘It is not easy … to gain money from microloans. Repayment is too difficult. We end up not paying and then they go to court to repossess the business and materials. A microgrant would be much better.’ (YW17). As Youth Development Fund (YDF) loans had been grants in the past, and youth rarely repaid loans, officers believed the proposal was feasible.

Mentorships were very desirable and, incorporated into programmes or outsourced to specialised parastatals or government psychologists, definitely feasible. Panellists preferred regular meetings focused on professional rather than social support, and monthly, rather than weekly meetings. Panellists strongly supported comprehensive support for at least one year but many preferred prolonging it to three years.

Many panellists interpreted ‘help to bring products to market’ as government assistance with transport. Some considered marketing, market development and negotiation training. Establishing market days, virtual markets, and social media networks could help develop local demand and professional networks. ‘The reality is that young women have a challenge to transport their products to markets. Lacking marketing or negotiation skills, they fail to supply. With a little help they could go far’. (PO10)

Although desirable, panellists doubted private sector apprenticeships were feasible. Rural areas often lack a private sector and panellists feared exploitation. Young women suggested clear terms of reference could help prevent the latter. Young women embraced government-sponsored childcare, but officers doubted the government would fund it and some felt childcare was a family responsibility.

B.	Improved training

Proposals to improve training focused on (1) skills development (life and job skills), (2) gender awareness training, and (3) access to training (Setswana language and Ipelegeng – the easiest accessible programme for vulnerable young women). Panellists perceived skills development as very desirable and feasible if offered by programmes or specialised parastatals. Training in the local language would make trainings more accessible. Offering training in Ipelegeng received mixed reactions, but most panellists felt it could be an opportunity to provide youth with relevant skills for Botswana’s future. Officers suggested redesigning Ipelegeng as apprenticeships with vocational training. ‘Maybe Ipelegeng should be improved. Not just cut trees [roadside maintenance], but a separate programme for the young ones that focuses on improving their future. It could be a chance for the government to train them in different jobs.’ (PO12)

Panellists felt gender awareness training was very desirable for young women and others in their social environment, especially boyfriends. Gender training could contribute to improved gender equality and a better understanding of gender violence and its consequences. It could increase self-esteem and empowerment of young women and benefit the larger community. Panellists considered gender training definitely feasible if offered to young women, but many doubted men would be interested in, or attend, such training. ‘Because men would never show up. They know. They are always the cause of the issues. They would not want to learn about these issues. They want to be heard, not listen.’ (YW22). To sensitise men about gender issues, panellists stressed using social media and informal, confidential, and non-confrontational peer approaches.

C.	Improved outreach

Proposals to improve outreach focused on (1) diversifying places and ways of information dissemination, (2) leveraging social networks (of young women, their boyfriends, and communities), and (3) paid peer assistance. Panellists felt all outreach proposals were very desirable and possibly feasible due to existing platforms and perceived cost-effectiveness. Panellists highlighted the potential of leveraging social media to provide youth with easy access to up-to-date information. Despite challenges like lack of connectivity or affordability, young women expressed an overriding wish to use social media. They offered solutions like posting community messages and proactively informing peers without Smartphones or network. ‘Nowadays [Covid-19], we’re not allowed to gather. On social media we would get information directly. We’re connected. Girls nowadays like Facebook and WhatsApp too much. It is our thing. When our parents give us phones, we use it immediately’. (YW13)

Panellists perceived young women as convincing and relatable messengers for peers, but several doubted the ability or goodwill of boyfriends or community members to inform and empower young women. Panellists strongly supported making outreach effective through paid peer approaches. Such peers would have to be trained, well-informed role models with a proven history of success. Outreach-associated expenses (copies and transport) and young women’s economic situation justified stipends. Officers viewed such approaches as cost-effective and sustainable, especially for rural areas. With someone trained locally, the information remained accessible, and cost less than programme officers travelling to rural areas. Young women expected stipends would boost their self-esteem, motivation, and possibly counter envy-driven withholding of information by unpaid peers. ‘This could create jobs, reduce rural-to-urban migration. It is not that we want to go to urban areas. We just go there because we have nothing to feed our families. We would be determined to reach many young women’. (YW4)

D.	Programme delivery and coordination

Panellists agreed all programme delivery and coordination proposals were very desirable and definitely, or possibly, feasible. Holding programme officers to account through performance indicators was considered necessary and easy to implement. Some officers interpreted ‘role models’ as celebrities. Most panellists thought of successful peers, including in agriculture, who could motivate young women to emulate them and persevere. ‘We want to learn from hands-on people who are successful despite some challenges. This can give us courage and strength to soldier on’. (YW5). Role models should be helpful and understanding, offer practical knowledge, share good and bad experiences, and explain how to survive while invested in long-term, but not yet profitable, projects. Several officers felt successful clients had a moral responsibility to help less fortunate peers.

To improve coordination, panellists proposed a one-stop-shop that would include all promotive social protection programmes and land and water boards. Cutting red tape, it could prevent programme duplication, increase efficiency, shorten processes, save human and financial resources, and improve service delivery. The district commissioner or village extension committee could be coordinating bodies at local level. ‘You need to bring all stakeholders together. Avoid duplication, wasting resources. Bring everything under one umbrella.’ (PO1)

Acknowledging access to land and water as structural barriers to economic empowerment, panellists proposed affirmative action for land and water allocation to young women. An officer suggested the government take a critical look at land policies to reallocate idling farmland, as many owners allegedly use only a third of large plots (70–100 hectares) allocated in the past. An officer proposed establishing a farming incubation programme with programmes applying for land to support young women on a rolling basis with comprehensive farming support for three or four years. Once graduated, young women could undertake farming autonomously. ‘We could have big plans … [having land and water] can encourage us to produce more and feed the nation.’ (YW5).

Officers stressed the need for a holistic approach, supported by a national strategy for HIV-sensitive social protection. As this depends on political will, such an initiative should come from the highest level of government, the Cabinet, or the Office of the President.

Round 2: Respondent validation

Panellists validated all R1-findings. Given the high level of agreement in R1, we did not ask panellists to adjust their views on individual proposals but offered all 39 for importance ranking in R2 (De Loe, 1995). Young women were keen to know about officers’ feedback and agreed with nearly all their suggestions with a few exceptions. That officers feared grants might induce laziness angered young women, prompting comments about officers’ responsibilities to assist in a timely manner and provide job skills and mentoring. Young women did not support an officer’s suggestion to organise skills training after working for Ipelegeng from 7am to 1pm, citing domestic and farm responsibilities. Childcare being a family responsibility triggered strong reactions. Family ties were not always strong, family members had their own responsibilities, and some young women were orphans. They perceived childcare as crucial to workforce participation, and suggested the government provide free childcare, or a bridging measure until stable in their livelihood. ‘We might have received grants, mentorship, and transport support, but we’ll fail if we have no childcare … and the government would say we’re lazy and irresponsible, and incapable of receiving grants. We are capable provided the core resources are availed’. (WS3)

Young women also reacted strongly to fear of exploitation in private sector apprenticeships, recounting government exploitation of Tirelo Sechaba apprentices. Apprentices cook, clean, and cut grass, and even teach when teachers are absent. They felt clear terms of reference were required in both public and private sectors. Young women were disappointed that an officer suggested leveraging Tirelo Sechaba apprentices for peer outreach and programme delivery, as they had hoped for another opportunity to earn income without the 30-year age limit.

During workshops with young women, we observed a gradual change in their views of officers. They appreciated officer input and felt better understood. They stood up, applauded, and bowed when learning that officers supported comprehensive packages with cash grants, savings, mentorship and health, HIV, gender, and skills training.

‘Hallelujah! We are really surprised with the feedback from the officers. They like hearing about us. They’re seeing us. They understand the challenges we’re facing. We thought they didn’t care about us. We’re surprised they’re willing to seek solutions to help us.’ (WS4)

Officers who were unsurprised by the rating similarities between young women and themselves said they knew the reality on the ground. Others were impressed with young women’s feedback. ‘[Young women] know their weaknesses … I thought they were a bit ignorant, so I am surprised they are so smart and honest.’ (PO2)

Round 2: Relative importance – across all proposals

Involving positive role models in programme delivery and improving access to land and water were in the Top 3 priority proposals of all stakeholders (Table 4). Officers stressed foundational skills with life skills as their top choice, followed by mentors and second chance education. Workshops mediated young women’s individual rankings: safe savings disappeared while proposals with more immediate benefits like grants, help to bring products to markets, and training in Setswana appeared after young women discussed priorities with peers.

Table 4. Top 3 Priorities across improvement categories

#	Improvement proposal	Overall	WS	YW	Officers
4.2 4.3	Positive role models in programme delivery, including in agricultural programmes	1	1	1	3
4.6	Access to land and water by YW	2	3	3	3
2.2	Business and financial skills training	2	3	3
3.8	Stipends for peer training and programme dissemination	3	2	2
1.2	Microgrants rather than loans		1
2.1	Life skills training in all programmes				1
1.7	Help bring products to markets		2
1.3	Mentors for psychosocial and business support				2
2.5	Training in preferred (local) language		3
1.1	Safe savings			3
3.1	Promote second chance education				3

Top 3 priority ranking ranges from 1 (highest) to 3 (lowest). Overall: mean rankings WS, YW, Officers; WS: young women group ranking in workshops; YW: individual rankings young women; Officers: individual officer ranking.

Round 2: Relative importance per improvement category and ranking differences between stakeholder groups

Table 5 shows the relative importance of proposals overall, for workshops, young women, and officers per improvement category. All proposals, except for gender awareness training for young women received a top three priority ranking by at least one panellist.

Table 5. Importance rankings for improvement proposals by stakeholder group

#	I. Proposals to improve programmes	Overall	workshops	YW	Officers
1.1	Safe savings	0.7	0.4	1.6	-
1.2	Microgrants rather than loans	1.3	2.2	1.1	0.6
1.3	Mentors for psychosocial and business support	1.4	1.4	0.9	1.9
1.4	Weekly meetings for professional and life skills training, psychosocial support, and a social network	0.5	0.4	0.3	0.7
1.5	Childcare	0.3	0.1	0.5	0.1
1.6	Create local demand for microenterprise products and services	0.3	-	-	0.7
1.7	Help bring products to markets	1.0	2.0	1.0	0.1
1.8	Establish private sector links through apprenticeships	-	-	0.1	-
1.9	Establish professional networks YW	0.3	-	-	0.9
1.10	Offer support for at least one year	0.1	-	0.1	0.1
1.11	Offer combination packages: cash grants, skills training, safe saving options, HIV/health/gender training, weekly meetings, and mentors	0.5	0.2	0.3	1.0
1.12	Ask people themselves what they want	0.2	0.6	0.1	-
#	II. Proposals to improve training	Overall	workshops	YW	Officers
2.1	Life skills training in all programmes	1.0	-	0.4	2.7
2.2	Business and financial skills training	1.6	1.8	1.7	1.4
2.3	Vocational job training	1.0	0.8	1.3	0.9
2.4	Include training component in Ipelegeng	0.4	0.6	0.5	-
2.5	Training in preferred (local)	1.0	1.8	0.6	0.6
2.6	Gender awareness training for YW	-	-	-	-
2.7	Gender awareness training for important persons in their lives (mother, father, brothers, boyfriends)	0.3	0.3	0.3	0.1
2.8	Provide outlets where YW may confidentially share their life experiences	0.2	-	0.3	0.1
2.9	Ask YW what training programmes they would want and bring that to reality	0.3	0.6	0.3	0.1
#	III. Proposals to improve outreach	Overall	workshops	YW	Officers
3.1	Promote second chance education	0.9	0.6	0.5	1.6
3.2	Diversify venues for outreach	0.5	0.4	0.6	0.6
3.3	Diversify means for outreach	1.0	1.2	0.9	1.0
3.4	Involve YW in outreach	0.4	-	0.2	0.9
3.5	Involve boyfriends in outreach to YW	0.1	-	0.2	-
3.6	Involve communities in outreach to YW	0.1	-	0.1	0.3
3.7	Train YW to help peers to fill out applications	0.9	1.6	0.9	0.1
3.8	Stipends for peer training and programme dissemination	1.5	2.0	1.7	0.7
3.9	Develop a TV programme about YW. A programme that travels all over the country and documents our issues	0.7	0.6	0.7	0.7
3.10	Use boot camps for several weeks that offer all the different kinds of training	0.2	0.2	0.4	0.1
#	IV. Proposals to improve programme delivery and coordination	Overall	workshops	YW	Officers
4.1	Performance indicators for programme officers	0.6	0.6	0.4	0.9
4.2	Positive role models in programme delivery	1.0	1.6	0.6	0.9
4.3	Positive role models for agriculture programmes	0.8	0.6	1.2	0.7
4.4	Improve coordination between programmes	0.1	-	-	0.4
4.5	Improve coordination between programmes and land and water boards	0.4	-	0.6	0.6
4.6	Improve access to land and water by YW	1.6	1.8	1.6	1.4
4.7	Develop national strategy for HIV-sensitive social protection	0.3	0.4	0.5	-
4.8	Unemployed and out-of-school YW should be offered leadership roles, like champions	0.03	-	0.1	-
4.9	Ensure YW have political representation*	0.3	0.6	-	0.4

Mean importance is based on rank 1st: 3 points; 2nd: 2 points; 3rd:1 point. Overall: mean for combination of workshops, individual young women, officers. Workshops: ranking by groups of young women. YW: ranking by individual young women. Bold: most important proposal in improvement category; italic: proposals proposed by panellists; Bold-italic: most important improvement proposal when combined. *proposal added in 2nd round.

Comparing stakeholder rankings, we found young women and officers did not share any of the top three priority proposals in the programme improvement category. Young women’s first choice was safe savings followed by grants and help to bring products to markets. ‘If we have money in the house, I go buy unnecessary things. If we can save, we can budget and plan for the future’ (YW16). Grants were preferable to loans as profits were small and went to loan repayment. Among officers, these proposals enjoyed some (grants) to no (savings) priority support. Officers’ top three priorities were mentors, comprehensive packages, and establishing professional networks.

Priorities for improved training were similar between young women and officers and focused on life- and job skills training (business, financial, and vocational). Officers were near-unanimous in their preference for life skills: ‘Life skills training is the core issue … It is foundational to life in general but also to business’ (PO3). No officer selected training in Ipelegeng, which was young women’s second choice.

Officers and young women differed but supported each other’s top three proposals for both outreach and programme delivery/coordination categories. For outreach, officers’ first choice was advertising free second chance school options. Young women preferred stipends for peer outreach and help filling out applications. Both groups selected using social media for information dissemination.

Positive role models in programme delivery (general and in agriculture combined), access to land and water, and accountability through performance indicators were panellists’ top three priorities for programme delivery/coordination. ‘Program officers would then work hard because they know … they have to reach a certain target. Like this they will help more young women than now’ (YW15). Young women supported positive role models in agriculture even more than officers, offering creative ideas ‘ … ostrich farming. That kind of farming can diversify farming, the economy of the country and trigger a turnaround of our lives. We could work there … employ and contribute to job creation’ (WS5).

Comparing young women’s individual and group rankings, stipends and access to land and water remained unchanged. We found workshops mediated individual rankings for grants, replacing safe savings (ranked first by individuals) and training in Setswana (not in the individual top three), which shared first place with business and financial skills training. One group added ‘political representation’ as a 40th improvement proposal and ranked it number one.

Discussion

Promotive social protection programmes could reduce the detrimental effects of structural drivers of HIV-vulnerability like poverty, lack of education, and gender inequality, but unemployed and out-of-school young women in Botswana rarely benefit from these programmes. A panel of young women and frontline officers identified proposals that could make policy and practice more inclusive of HIV-vulnerable young women, hence more HIV-sensitive. As barriers to inclusion not only related to programs themselves but were in part due to young women’s skill levels and structural factors, improvement proposals were comprehensive, extending beyond the strict remit of social protection.

Panellists considered nearly all 40 improvement proposals as very desirable, most as definitely feasible, and prioritised a wide range of proposals. It suggests ample policy space to improve the HIV-sensitivity of Botswana’s social protection programmes promoting (youth) social and economic empowerment. This is further supported by the high level of agreement between stakeholders, and the high response and retention rates in our panel. As potential clients or frontline service providers, they know the extensive socioeconomic disadvantage of unemployed and out-of-school young women and are frustrated by the lack of current programmes to address it. Panellist support for comprehensive packages and a holistic approach reflects their desire to achieve structural change and positive impact on young women’s socioeconomic success.

All panellists valued education, training, and skill building. Whereas officers stressed life skills training, young women preferred business, financial and vocational training. Age might explain this difference. Most of the young women, who had been in their early twenties at the beginning of this process, were now in their late twenties. Older age and participation in this research project may have increased their self-confidence and communication skills. Life skills remain important for adolescent girls and younger women, however, as life skills contribute to increased socioeconomic success (Goodman et al., 2016; Pronyk et al., 2006) and HIV prevention through improved sexual negotiation and reduced sexual risk behaviours (Gibbs et al., 2020; Pettifor et al., 2019). Older age may also explain young women’s disappointment with an officer’s proposal to leverage age-limited apprenticeship programmes for peer approaches, their relative lack of interest in returning to school, and interest in farming.

In exploratory findings, young women had not been interested in farming (van der Wal et al., under review), but the prominence of ‘improving access to land and water’ demonstrates agricultural interest. Shortages of agricultural produce due to COVID-19 restrictions (UN Botswana, 2020) and increasing governmental protectionist policies promoting local farm produce in Botswana (Thukwana, 2022) may have contributed to this phenomenon. Considering the strong support for vocational job training, it may also reflect a paradigm shift from academic success and white-collar jobs to self-sufficiency through agriculture. This is relevant for HIV prevention, as access to arable land is associated with reduced food insecurity (Nnaji et al., 2022), a known risk factor for HIV infection (Low et al., 2022; Weiser et al., 2007).

Gender awareness training for young women was considered very desirable but the only proposal failing to garner top three support for relative importance, unlike gender training for boyfriends despite doubt men would attend. Nearly all men and women had enjoyed participation in gender equity training South Africa’s Stepping Stones trial, however (Jewkes et al., 2012). Considering the monumental sociocultural change required for gender equality, tackling poverty may have seemed the lower hanging fruit, or panellists may have viewed gender training for boyfriends more effective given women’s lack of relational power (Leiter et al., 2007; Nkomazana, 2021). It may indicate a desire to tackle gender-based violence, which is common in Botswana (Machisa & van Dorp, 2012), or increase supportive behaviour among important men in the lives of economically active young women, as gender training in Uganda did (Green et al., 2015).

Panellist preference for peer approaches in outreach and programme delivery may have a plausible explanation in Botswana’s foundational development principle – Botho, a concept of mutual respect, responsibility, interdependence, and empowerment through a process of empowering others (Government of Botswana, 2016). While officers referred to beneficiaries helping peers as an obligation to give back, young women were excited with the prospect of learning from positive role models and opportunities to contribute to society. References to job creation, rural-to-urban migration and feeding the nation indicate their concern for the nation’s wellbeing, which concurs with research on young farmers in Botswana (Williams & Hovorka, 2013). Together with proposals that highlighted elements of political, popular (with a television show), or functional (role models) representation, it indicates young women’s desire to be viewed as contributing members of society, worthy of respect (Ingram et al., 2007).

Young women’s active engagement in assessing public policy might also support their social reconstruction as deserving, capable, and powerful clients. Schneider and Ingram (1993) posit that such positive reconstruction might have real world consequences, as powerful, positively constructed people tend to receive more, and more generous benefits, framed as rights and offered proactively. In contrast, social protection programmes for negatively constructed, or ‘undeserving’ groups, tend to be means-tested with stigmatizing labels and parsimonious benefits that beneficiaries often need to claim through inefficient and demeaning processes with little options for recourse (Schneider & Ingram, 1993). Support for improved accountability, coordination, and the idea of a one-stop-shop seem to support this theory while also demonstrating panellists’ aspiration to reimagine programmes and programme delivery with young women at the centre. The additional proposals panellists contributed also focused on young women’s preferences, stressing the importance of including their voice in potential redesign of Botswana’s social protection programmes.

Strengths and limitations

Attrition, a standard Delphi limitation (Hsu & Sandford, 2007), was low, likely due to our long engagement with panellists (since 2017), mostly face-to-face interviews, and regular communication. R1’s individual approach, anonymity, videos, and ample time for understanding and discussing improvement proposals may have facilitated truthful contributions. We maintained anonymity, a Delphi characteristic to reduce power imbalance, between stakeholder groups. For the single stakeholder group of young women, we believed collaborative learning outweighed anonymity and peer discussions may have encouraged more critical reflection. One R2-workshop was dominated by a young woman, possibly introducing groupthink (Hsu & Sandford, 2007), as others aligned their preferences with hers. Other workshops probably benefitted from collaborative learning, with young women shifting proposal choices from personal to more general benefit for others like themselves. We elicited personal opinions, which could be subject to social desirability bias. There may have been selection bias such that officers holding more favourable views of young women and potential solutions were more likely to participate than officers with less favourable views. All panellists received the long summary with feedback per proposal. Panellists also receiving the short summary may have based their priority ranking on less detailed feedback, but we do not expect this to disturb the main message from the study. Reporting majority opinions could lead to groupthink, which we aimed to counter by also reporting minority opinions when aggregating group responses.

Our approach to work with potential clients and frontline service providers may have generated context-sensitive, feasible implementation solutions. Given the homogeneity of ratings, and panellists offering implementation suggestions while considering feasibility, we judged two rounds as sufficient. A third round to reconsider rankings might have increased the stability of panellists’ priorities. R2’s very spread, however, underlines the need for a comprehensive policy review with an HIV-sensitivity lens. Our study focused on one district, from which the exact findings might not be transferable to other districts in Botswana. Our illustration of a contextualising process is transferable, however.

Conclusion

This study investigated the perceived desirability, feasibility, and relative importance of evidence-based proposals that could improve the HIV-sensitivity of promotive social protection programmes in Botswana. By reducing social and economic disadvantage of unemployed and out-of-school young women, available programmes might reduce their HIV-vulnerability. Young women and frontline officers contributed comprehensive improvement proposals tailored to their needs and local context, offering decision-makers a wide spectrum of realistic redesign options with supporting evidence for acceptability and feasibility.

Ethical approval

Botswana’s Health Research and Development Committee (HRDC00724) and McGill University (A12-B72-18A) granted ethical approval considering this study minimal risk. We registered oral informed consent from panellists prior to surveys, interviews, and workshops (Appendices 2 and 3).

Acknowledgements

RW conceptualised the study; collected, analysed, and interpreted the data; wrote the first draft of the paper and revised it for submission. MK collected data; reviewed the analysis and interpretation; and critically reviewed the paper. AC, IV, MJ, and NA critically reviewed the paper. All authors approved the final manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

RW was supported by a CIHR Vanier Canada Graduate Scholarship; The authors thank the Quebec Population Health Research Network for its contribution to the financing of this publication.

Notes

1 HIV-sensitive social protection for unemployed and out-of-school young women in Botswana: an exploratory study of barriers and solutions.