Introduction

Despite notable progress in HIV prevention and treatment worldwide, HIV remains a significant public health challenge in sub-Saharan Africa (SSA) [1]. As of 2023, SSA is home to approximately 65% of people living with HIV (PLHIV), and accounts for about 50% of new HIV infections worldwide [2]. Adolescent girls and young women in SSA are particularly vulnerable to HIV due to a complex interplay of biological, behavioral, and socio-structural risk factors [3, 4]. Evidence shows that household financial instability, school dropout, and mental health challenges increase the risk of HIV infection among these groups as these factors found to be strongly associated with sexual risk-taking behaviors such as age-disparate or transactional sex and early marriage [5–10]. Against this backdrop, a multi-level response to address the complex and inter-related drivers of HIV risk among adolescent girls in SSA is essential to achieving the 95-95-95 targets set by the Joint United Nations Programme on HIV/AIDS goal by 2030 [1].

In the Central Region of Uganda, which has a disproportionately high HIV prevalence of 10.6% compared to the national average of 6.2%, [11] a randomized control trial “Suubi4Her” was implemented between 2017 and 2022 [12]. Aimed at addressing the multi-dimensional drivers of HIV risk among adolescent girls, the Suubi4Her intervention integrated an asset-based financial inclusion and economic empowerment model with a family strengthening intervention to reduce HIV risk behaviors. An evaluation of the short-term impact of the combination intervention showed that adolescent girls in the treatment arms experienced a significant reduction in the prevalence of depressive symptoms after 12 months of implementation [13]. The positive effects on mental health outcomes persisted at 24 months, however no significant effects on sexual risk-taking behaviors and attitudes were observed, potentially due to the young age of the target population [14].

Given the growing emphasis on informed healthcare spending, evidence on the effectiveness of interventions alone is insufficient to justify investments. In addition to evaluating outcomes, a comprehensive assessment of the economic costs associated with the delivery of interventions is critical for intervention planning, resource allocation decisions, and priority-setting in the context of ongoing resource limitations, growing budget constraints, and shifting health needs, particularly in low-resource settings [15]. However, there is limited information available on the costs of combination interventions [15, 16]. This study seeks to contribute to the currently limited economic evidence in this area by estimating the economic costs of the Suubi4Her intervention, an evidence-based combination intervention aimed at improving mental health and reducing HIV risk behaviors, with the goal of reducing new HIV infections among adolescent girls in Uganda.

Methods

Cost data collection and analysis

To enhance the reliability and validity of the cost data, resource use in each study arm was captured prospectively over the trial period. Costs were then analyzed using an activity-based micro-costing approach from the program provider perspective. Micro-costing enumerates all resources used in intervention delivery and hence is considered the most reliable costing method [17]. The program provider perspective maintains the focus on costs pertinent to the implementation of an intervention and facilitates budgeting, financial planning, and technical efficiency analysis in the face of limited healthcare resources and competing interventions [18, 19].

To accurately capture resource use, we first identified all program-related activities in each study arm through a review of study protocols and administrative records and periodic interviews with key program staff. The shared activities across all three study arms included identification of schools and school visits, screening and recruitment of study participants, delivery of usual care, and stakeholder engagement and dissemination. We included the costs of school identification and visits and recruitment and screening processes as these costs may be incurred in real-life contexts depending on intervention’s eligibility requirements, delivery method, and implementation setting. In each treatment arm, the activities involved in the delivery of the interventions were also identified. In treatment arm 1, these activities included the delivery of four FLT sessions that cover basic principles of financial management including income generation, use of financial institutions, saving and asset-building as well as opening of and contributing to YDAs. In treatment arm 2, YDA activities in treatment arm 1 were combined MFG-related activities, which included the training and supervision of MFG facilitators and the delivery of MFG sessions. The costs of the activities conducted purely for the purposes of research such as baseline and follow-up assessments of the study participants through interviews and surveys were excluded from the cost analysis because our objective was to estimate the resource requirements of the Suubi4her intervention in a non-research setting.

Next, we measured and valued all resources used for each program activity by study arm. Shared resources across program activities and study arms (i.e., personnel, program overheads, donated resources, and capital costs) were measured and valued as separate cost categories. Personnel included a program coordinator and other program staff who monitored the implementation of the interventions for quality assurance and worked as research assistants for research activities. Time costs of program staff were calculated based on average annual gross wage rates and time devoted to the Suubi4Her intervention over the trial period based on interviews with key program staff (S1 Text). Other recurring costs included monetary incentives provided to families, CHWs, PPs, and teachers (S1 Text), provision of educational materials, and program overheads, including utilities (electricity, water, communication), transportation to/from schools for implementation (fuel, taxi fare, car hire), security services and insurance, maintenance of equipment, vehicles, and facilities, office supplies and printing, and other miscellaneous expenses. Expenditures and other pertinent data for the cost analysis were extracted from the financial and administrative records of the program. Where relevant, the costs related to a program activity were further broken down into and listed as cost items. For example, the delivery of the MFG sessions included participation incentives for families (time and transport), facilitation incentives for MFG facilitators (time and transport), community mobilization incentives for teachers (time and phone airtime) as well as the costs of training of the MFG facilitators and the donated classroom space.

We also considered and valued resources leveraged for the program activities to arrive at the economic costs of the interventions. Leveraged resources included the donated classroom space used for the screening and recruitment of study participants and delivery of the FLT and MFG sessions, and the donated time by school teachers and bank officials. Lastly, recurring costs were differentiated from capital costs, which typically include the costs of capital items with an expected useful life of more than one year (e.g., equipment, vehicles, and furniture). All capital costs were annualized using the replacement cost of each capital item, an appropriate useful life for each item, and an annual discount rate following the approach suggested by Walker and Kumaranayake [20]. This annualization method takes into account the opportunity cost of capital items, allowing us to estimate the economic cost of the capital in any year rather than its financial cost [20].

We then calculated the total cost per adolescent per study arm. To do this, all costs were first adjusted for inflation using the Ugandan Consumer Price Index (CPI) [21], discounted at an annual rate of 3% [22] to the start year of the intervention, and presented in 2018 US dollars [23]. The adjusted activity-based costs were then summed to calculate the total costs per study arm. Shared resources (program personnel costs, donated resources, overheads, and other capital costs) were apportioned to program (80%) activities versus research (20%) activities based on interviews with key program staff, and we only included the program costs in the analysis. By dividing the costs of the intervention activities in each study arm by the total costs of all intervention activities across the three arms we calculated ratios (UC, 17% vs. YDA, 35% vs. YDA+MFG, 48%) to apportion the costs of the shared resources to the three study arms. Total per-adolescent costs by study arm were calculated by dividing the total costs per study arm by the number of adolescent girls in each arm using the intent-to-treat (ITT) sample and the treatment-on-the-treated (TOT) sample. While the TOT sample includes persons who participated in an intervention, allowing a more conservative calculation of per-person costs, the cost analysis using the ITT sample yield per-person costs under the assumption that the intervention was taken up by all those who were offered. A summary of identified program activities, cost items, and measurement and calculation methods, is presented in Table 1. Our study follows the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) and the CHEERS checklist is available in S1 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Identification, quantification, and valuation of main cost categories.

https://doi.org/10.1371/journal.pmen.0000201.t001

Results

Table 2 presents per-adolescent costs by cost category and study arm (UC, N = 420; UC+YDA, N = 420; and UC+YDA+MFG, N = 420) using the ITT sample. The per-adolescent cost for the study arms was estimated at $264 for the UC arm, $534 for the UC+YDA arm, and $737 for the UC+YDA+MFG arm. The incremental cost over the control arm was $270 for the UC+YDA arm and $472 for the UC+YDA+MFG arm. Personnel costs had the biggest share of the program costs across all study arms, followed by program overheads. In the treatment arms, the economic empowerment intervention activities (YDAs, FLT sessions, and IGA workshops) were key cost drivers, following personnel and program overheads. The implementation costs of the family strengthening intervention (MFG sessions) and the economic empowerment intervention were comparable, each accounting for about 3% of the total cost of the UC+YDA+MFG arm.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Per-child costs of the economic empowerment and family strengthening intervention by study arm, using the intent-to-treat (ITT) sample (All costs are in 2018 Ugandan Shillings unless otherwise indicated).

https://doi.org/10.1371/journal.pmen.0000201.t002

Table 3 presents the cost per-adolescent by cost category and study arm using the TOT sample. These cost estimates were derived based on the actual number of adolescents participated in the study per arm, making them higher and more conservative than those estimates by the ITT sample across all study arms. The UC arm had 408 adolescents, the UC+YDA arm 471 adolescents, and the UC+YDA+MFG arm 381 adolescents. Cost per-adolescent estimates based on the TOT sample were $272, $476 and $812 for the UC, UC+YDA, and UC+YDA+MFG arms, respectively. The incremental cost over the UC arm was $204 for the UC+YDA arm, and $540 for the UC+YDA+MFG arm.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Per-child costs of the economic empowerment and family strengthening intervention by study arm, using the treatment-on-the treated (TOT) sample (All costs are in 2018 Ugandan Shillings unless otherwise indicated).

https://doi.org/10.1371/journal.pmen.0000201.t003

Discussion

In SSA, while access to HIV services has significantly improved since the early 2000s [1, 2], mental health services for children and adolescents remain severely limited, with substantial gaps in both policy and evidence [16, 24]. This is the first costing study of a combination intervention that combined an evidence-based family strengthening intervention together with a proven FEE intervention in the SSA context. Specifically, we expanded the current literature by estimating the economic costs of a combination intervention for adolescent girls in Uganda, a low resource setting. A key strength of this costing study is the prospective collection of resource use and cost data by trained staff alongside the Suubi4Her trial. We employed an activity-based micro-costing approach to improve the reliability and validity of cost estimates, a method particularly recommended for interventions without previous cost data [25]. The micro-costing approach, in contrast to the macro-costing approach, highlight the context-specific variations in the data, which is crucial for evaluating the feasibility and affordability of interventions within different implementation settings. Lastly, we closely followed the methodological recommendations of the Second Panel on Cost Effectiveness in Health and Medicine [18], carefully justifying the provider perspective based on the study’s aim and adjusting the cost data appropriately, including currency conversion and inflation, discounting, and annuitization. Our approach aims to lay the ground for future studies evaluating the costs of combination interventions in other low-resource environments and the cost-effectiveness of this intervention which was shown to be effective in positively improving mental health outcomes of school-going adolescent girls in Uganda [26].

Our analysis showed that the cost per adolescent for delivering the FEE intervention through YDAs on top of UC was US$476, while the cost for the combined FEE and MFG intervention was US$812. Personnel costs represented the largest portion of the total costs, as the intervention activities demanded extensive supervision and quality assurance measures. The body of economic evidence on FEE interventions is currently limited. Previous studies in Uganda have costs for FEE interventions aimed at improving mental health outcomes for children and adolescents affected by HIV/AIDS, ranging between US$418 and US$426 per child [27]. Similarly, few studies have evaluated the costs of the MFG intervention. For example, a recent study in Uganda reported per-child costs of delivering the MFG intervention by PPs and CHWs at US$346 and US$328, respectively [15]. Pilot studies focused on improving parenting practices in Liberia and Thailand estimated per-beneficiary costs between US$650 and US$900 [28]. A South African study estimated the cost of a parenting program designed to prevent violence against adolescents at US$504 per family [29]. Compared to these studies, the FEE intervention’s cost of US$476 was similar to existing findings, while the combination intervention’s cost of US$812 per adolescent fell within the range reported in these studies. However, the limited number of costing studies underscores the need for further research in diverse implementation contexts and country settings.

While it remains uncertain if the levels of investments required for these interventions would be deemed acceptable to policy-makers in low- and middle-income countries (LMICs), there are compelling reasons to justify increased investment in adolescent health and wellbeing. One notable reason is the increasing proportion of children and adolescents within the overall LMIC population driven by improved survival rates and declining fertility rates [30] and the pressures the social and economic adversities place on transition to adulthood, particularly in resource-constrained settings [31]. Moreover, evidence increasingly supports the argument that investing in interventions that address multiple facets of adolescent health and wellbeing may yield significant long-term benefits. These benefits include improved health in adulthood and economic benefits for countries, with the highest return on investments in low-income and lower-middle countries [32, 33]. In light of this, there is a clear need for greater investment in interventions targeting child and adolescent health outcomes, as well as a pressing need for more research on the economic costs and cost-effectiveness of such interventions.

This study has several limitations. First, the time spent on intervention-related activities by program staff, teachers, and bank officials was estimated based on interviews with key personnel, which may have introduced bias. An alternative method is the use of daily logs to capture time-use data more accurately. However, this approach is time-intensive and can place a significant burden on program staff involved in the program activities. Second, this costing study was undertaken alongside a randomized controlled study. Costs associated with research activities, such as baseline and follow-up assessments conducted through interviews and surveys, were excluded from the analysis to focus on estimating the resource requirements for the interventions in a non-research setting. Nonetheless, some costs may still be ambiguous. For instance, monitoring, evaluation and quality assurance are crucial for the successful implementation of interventions in real-world settings. However, these costs are typically higher in research studies due to the more intensive nature of service delivery. Furthermore, the shared nature of certain resources such as program personnel, donated resources, program overheads, and capital items, complicates the task of allocating costs between program or research activities. Due to the lack of a precise allocation method, we applied conservative apportioning ratios based on the information provided by key program staff. Lastly, the interventions were delivered through task-shifting in school settings across specific districts of Uganda. The resource utilization and associated costs may vary based on the delivery approach and geographical context, which limits the generalizability of our findings. This limitation is, however, balanced by the richness of the data micro-costing generates, providing a granular and comprehensive understanding of the inputs required for each intervention activity, which in turn allows for accurate and context-specific estimates. Further, the method highlights key cost drivers, enabling policymakers and program implementers to understand which factors contribute most to overall costs and where efficiency gains may be possible.

Conclusions

Despite these limitations, this study leveraged prospectively data on resource use and costs alongside implementation, providing robust cost estimates for delivering a combination intervention to school-going adolescent girls in Uganda, a low-resource setting. As a result, the study provides policymakers and program implementers with essential information to assess the feasibility, affordability, and long-term sustainability of the intervention within their own context. Furthermore, the results lay the groundwork for conducting a cost-effectiveness analysis of the combination intervention in comparison to other proven interventions aimed at improving mental health outcomes and reducing HIV risk behaviors among adolescents in resource-limited settings, ultimately guiding resource allocation decisions in HIV prevention and mental health.

Supporting information

Acknowledgments

We thank Abel Mwebembezi of Reach the Youth, Uganda, Rev. Fr. Joseph Kato Bakulu of the Masaka Catholic Diocese, and the entire field team of the International Center for Child Health and Development (ICHAD) for their contributions to the study design and implementation. In addition, we are grateful to the financial institutions that agreed to work with the adolescent girls in opening savings accounts, and the extension workers who have committed time to train the adolescent girls in conducting income-generating activities. Our thanks also go to the Ugandan Government Ministry of Education and the 47 secondary schools that have agreed to participate in the Suubi4Her study.