Costing of a Multiple Family Group Strengthening Intervention (SMART Africa) to Improve Child and Adolescent Behavioral Health in Uganda

ABSTRACT.

Reliable cost estimates are key to assessing the feasibility, affordability, and cost-effectiveness of interventions. We estimated the economic costs of a multiple family group (MFG) intervention—child and adolescent mental health evidence-based practices (CAMH-EBP) implemented under the SMART Africa study, seeking to improve family functioning and reduce child and adolescent behavior problems—delivered through task-shifting by community health workers (CHWs) or parent peers (PPs) in school settings in Uganda. This prospective microcosting analysis was conducted from a provider perspective as part of a three-armed randomized controlled trial of the MFG intervention involving 2,391 participants aged 8-13 years and their caregivers in 26 primary schools. Activity-specific costs were estimated and summed, and divided by actual participant numbers in each study arm to conservatively calculate total per-child costs by arm. Total per-child costs of the MFG-PP and MFG-CHW arms were estimated at US$346 and US$328, respectively. The higher per-child cost of the MFG-PP arm was driven by lower than anticipated attendance by participants recruited to this arm. Personnel costs were the key cost driver, accounting for approximately 70% of total costs because of intensive supervision and support provided to MFG facilitators and intervention quality assurance efforts. This is the first study estimating the economic costs of an evidence-based MFG intervention provided through task-shifting strategies in a low-resource setting. Compared with the costs of other family-based interventions ranging between US$500 and US$900 in similar settings, the MFG intervention had a lower per-participant cost; however, few comparisons are available in the literature. More costing studies on CAMH-EBPs in low-resource settings are needed.

INTRODUCTION

Mental health is integral to overall health and well-being. ¹ Yet in low- and middle-income countries (LMICs), a staggering 33% of young children are at risk for developing mental health disorders, with potentially far-reaching health and socioeconomic consequences. ²

Evidence-based practices (EBPs) for prevention and treatment of child and adolescent mental health (CAMH) disorders exist. ³ Resource constraints in LMICs, including the dearth of mental health professionals, have prompted calls for innovative approaches to expand provision of mental health services in community settings. ⁴ As a result, CAMH-EBPs have been adapted for delivery by lay health workers, consistent with task-shifting strategies. ⁵ Despite the extensive evidence on the transportability of these EBPs to low-resource settings, ⁶ including sub-Saharan Africa (SSA), ^{7, 8} large-scale implementation and evaluation studies to guide the expansion of services and improve the range of feasible and affordable interventions in such settings are sparse. ⁹ Further, there is little evidence on the economic costs and cost-effectiveness of CAMH-EBPs in low-resource contexts. ^{5, 10}

In Uganda, children and adolescents make up more than 50% of the population. ¹¹ The majority live in conditions that place them at risk for developing mental health disorders, such as high rates of chronic poverty, ¹² domestic violence, ¹³ and physical violence against children. ¹⁴ Overall, estimates for Uganda suggest that every fourth child suffers from an externalizing or internalizing mental health problem. ¹⁵

Against this background, the Multiple Family Group (MFG) intervention—a CAMH-EBP designed to improve family functioning and reduce child and adolescent behavior problems—was adapted to be delivered to children and their caregivers in primary school settings in Uganda, under the Strengthening Mental Health and Research Training in Africa (SMART Africa) study. ⁹ The effects and feasibility of the MFG intervention delivered through task-shifting by community health workers (CHWs) or parent peers (PPs) were evaluated in a randomized controlled trial (RCT). An evaluation of intervention effects at 16-week follow-up showed that children in both the MFG-CHW and MFG-PP groups reported significantly lower levels of disruptive behavioral disorders than children in the control group. ²¹

A rigorous examination of economic costs of mental health interventions alongside efficacy and effectiveness trials is pivotal to assessing their feasibility, affordability, and cost-effectiveness. Resources needed to deliver mental health interventions depend on intervention setting (institutional, community-based), type of care providers (medical, nonmedical) and delivery format (individual, group). ¹⁶ Cost evaluation facilitates more realistic planning and budgeting of the introduction, integration, and scale-up of interventions, especially in resource-constrained healthcare settings with competing health priorities. ¹⁷ Despite their importance, little is known about the economic costs of interventions designed to improve CAMH outcomes through task-shifting strategies in community-based settings. ¹⁰ This study aims to estimate the economic costs of the MFG intervention delivered by CHWs or PPs in a school-based setting in Uganda and help fill the current knowledge gap on the costs of CAMH-EBPs in LMICs.

MATERIALS AND METHODS

Study setting and design.

The MFG intervention targeted youth attending schools and was evaluated in an RCT involving 2,391 children aged 8 to 13 years and their caregivers in 26 primary schools in southwestern Uganda. ¹⁸ The trial was registered on ClinicalTrials.gov on March 16, 2017, NCT03081195 (https://clinicaltrials.gov/ct2/show/NCT03081195). The region is characterized by high poverty rates, ¹⁹ in part due to civil wars and HIV/AIDS, with estimates indicating that two out of five children are orphaned by AIDS each year. ²⁰

The study was conducted with the Masaka Diocese, which operates more than 420 schools in Greater Masaka. ⁹ Schools were randomly assigned to three study arms: 1) MFG delivered by trained PPs (MFG-PP arm, N = 8 schools), 2) MFG delivered by trained CHWs (MFG-CHW arm, N = 8 schools), and 3) bolstered standard of care (BSOC; control arm, N = 10 schools) for mental health promotion through provision of mental health wellness materials developed by the Ugandan Ministry of Health. ⁹ Teachers and children received educational materials, and children school lunch. This is the first CAMH-EBP evaluated for effects and implementation feasibility through task-shifting approaches in the context of a large-scale RCT in Uganda. ¹⁸

MFG intervention.

Adapted from an evidence-based family-strengthening intervention from the United States and guided by the 4Rs (Rules, Responsibility, Relationships, Respectful Communication) and 2Ss (Stress and Social Support) skills and family processes, ²² the MFG intervention was delivered by either PPs or CHWs to children and their caregivers. A total of 48 PPs (six per school) and 48 CHWs (six per school) served as MFG facilitators and received training and supervision by program staff.

Schools organized initial meetings to inform and mobilize families about the MFG intervention and the study. After informed consent processes, caregivers completed a screening measure for child disruptive behavioral disorders. Children who screened positive were invited to enroll in the study, together with their families, and completed a full baseline assessment before the start of MFG sessions. Children who screened negative were also invited, together with their families, to complete an abbreviated baseline assessment, and attend MFG sessions so as to decrease stigma in the school community and enhance parenting skills and functioning of all families. MFG sessions were delivered weekly to groups of 10 to 12 families in a school setting for 16 weeks, with each session taking 60 to 90 minutes. At least two generations of a family (a child–caregiver dyad) were present in each session. Content and practice activities fostered learning and interaction within and among families. All MFG sessions were supervised by program staff, and fidelity assessments were conducted for a quarter of the sessions.

Cost data collection and analysis.

Using an activity-based micro-costing method, ²³ we conducted cost analysis of each of the three study arms from the program provider perspective. Micro-costing entailed a three-step approach where we identified, measured, and valued resource use in each study arm. To accurately capture resource use, we first identified all program-related activities in each study arm through periodic interviews with key personnel and a review of study protocols and records. Program-related activities included school identification and family mobilization; participant screening and recruitment; provision of school lunch and educational materials to all participating children; standard of care mental health promotion; training and supervision of MFG facilitators; MFG delivery; and stakeholder engagement and dissemination. Next, we measured and valued all resources used for each program activity or cost category (Table 1). Shared resources across activities and study arms (personnel, donated resources, overheads, and capital costs) were measured and valued as separate cost categories and apportioned to the study arms (Table 1). Relative use of staff time for program (80%; versus research, 20%) activities and for each study arm (20% standard care; 40% MFG-PP; 40% MFG-CHW) was determined based on interviews conducted by the key program personnel. The results were then summed to obtain component-specific and total costs, which were then divided by the number of children in each study arm to calculate the total per-child costs over the study period, using the intent-to-treat (ITT) and treatment-on-the-treated (TOT) samples as denominators. The TOT sample consists of persons who actually participate in a study and hence allows us to conservatively calculate per-child costs by study arm.

Table 1.

Identification, quantification, and valuation of main cost categories

Cost categories	Cost items	Cost calculation methods
Personnel	On the basis of administrative and financial expenditure records of the study, time devoted to program activities (as opposed to research) in each study arm by program staff	Extract the number of hours each staff member devoted to program activities each year, multiply the total hours by average hourly salary rate of staff (estimated based on average annual gross wage rate), and calculate the total cost across all staff. Apportion the total cost incurred each year based on level of effort dedicated by staff to program activities (80%) (versus research, 20%) and to each study arm (20% standard care; 40% MFG-PP; 40% MFG-CHW), and divide by number of participants in each study arm.
Identification of schools and school visits	Facilitation incentives for teachers (airtime for phones used to mobilize families), provision of educational materials to teachers and students (teacher handbooks, student notebooks, pens, textbooks, and geometry sets), participation incentives for families, snacks and refreshments	Divide the total cost incurred each year by number of participants in each study arm.
Screening and recruitment of study participants	Facilitation incentives for teachers (airtime for phones used to mobilize families), participation incentives for families, snacks and refreshments, interviewer compensation for screening	Divide the total cost incurred each year by number of participants in each study arm.
Provision of school lunch	School lunch (porridge) to all participating children	Divide the total cost incurred each year by number of participants in each study arm.
Standard mental health promotion	Mental health wellness materials, participation incentives for families, facilitation incentives for teachers (airtime for phones used to mobilize families)	Divide the total cost incurred each year by number of participants in each study arm.
Training of MFG facilitators	Participation incentives for PPs and CHWs, meals for program staff, snacks and refreshments	Divide the total cost incurred each year by number of participants in each treatment arm.
Delivery of MFG sessions	Facilitation incentives for PPs and CHWs for MFG sessions, facilitation incentives for teachers (airtime for phones used to mobilize families), meals for program staff, snacks and refreshments for families	Divide the total cost incurred each year by number of participants in each treatment arm.
Donated resources	Time spent by teachers and head teachers on mobilization of families for program-related activities, classroom space used for program-related activities at participating schools	Based on an hourly time cost of teachers and daily cost for classroom space, calculate the total cost incurred each year. Apportion the total cost to each study arm, and divide by number of participants in each study arm.
Program overheads	Utilities (water, electricity, communication), transportation (fuel, taxi fare, car hire), security services and insurance, maintenance (equipment, vehicles and facilities), office supplies and printing, and other miscellaneous costs	Apportion the total cost incurred each year based on level of effort dedicated by staff to program activities (80%) (versus research, 20%) and to each study arm (20% standard care; 40% MFG-PP; 40% MFG-CHW), and divide by number of participants in each study arm.
Stakeholder engagement and dissemination	Transport refund for stakeholders and meeting-related costs	Divide the total cost incurred each year by number of participants in each study arm.
Capital costs	Capital items with an expected useful life of more than one year, such as equipment, vehicles, and furniture	Calculate equivalent annual costs by annualizing all capital costs over the useful life of capital items (3 years for equipment, 5 years for furniture, and 10 years for vehicles), and apportion the total cost incurred each year based on level of effort dedicated by staff to program activities (80%) (versus research, 20%) and to each study arm (20% standard care; 40% MFG-PP; 40% MFG-CHW), and divide by number of participants in each study arm.

CHW = community health workers; MFG = multiple family group; PP = parent peer.

To enhance the reliability and validity of the cost data, we measured resource use prospectively over the trial period, rather than reconstructing them retrospectively. Time costs of program staff were calculated based on average annual gross wage rates and apportioned according to estimated time devoted to program- and research-related activities. Other recurring program costs included transport refund/facilitation for families, transport refund/facilitation for PPs and CHWs for MFG session implementation, facilitation for teachers to mobilize families (airtime for cell phones), provision of school lunch (porridge) for students, provision of educational materials for teachers and students (teacher handbooks, student notebooks, pens, geometry sets and textbooks), and program overheads, including utilities (water, electricity, communication), transportation of program staff to/from schools for implementation (fuel, taxi fare, and car hire), security services and insurance, maintenance of equipment, vehicles and facilities, office supplies and printing, and other miscellaneous expenses. Expenditure and other pertinent data for these cost items were extracted from the administrative and financial records of the study.

We also considered and valued any volunteered and donated resources used for program activities to arrive at the economic costs of the intervention. Specifically, classroom space donated at participating schools and time donated by teachers and head teachers for program activities were estimated, valued, and included in the cost analysis. Recurring costs were differentiated from capital (nonrecurrent) costs, which typically include the costs of capital items with an expected useful life of more than 1 year (e.g., equipment, vehicles, and furniture). For all capital items, annual depreciation costs were calculated assuming an appropriate useful life for each item, and were apportioned to the study arms based on level of effort dedicated by program staff to the activities in each arm. All costs related to the evaluation of the MFG intervention by the research team, such as baseline and follow-up assessments of study participants, were excluded from the analysis, as our objective was to estimate the resource requirements of reproducing the intervention in a nonresearch setting. All costs were adjusted for inflation using the Ugandan Consumer Price Index in Masaka, ²⁴ discounted at an annual rate of 3% to the start year of the trial, and presented in 2018 US dollars. ²⁵

RESULTS

Table 2 presents per-child costs by cost category and study arm (Control arm, N = 1,000; MFG-PP, N = 1,000; MFG-CHW, N = 1,000), using the ITT sample. From the healthcare provider perspective, the per-child costs were similar for the MFG-PP and MFG-CHW arms at US$254 and US$255, respectively. This is because the delivery format of MFG sessions and the facilitation incentives for PPs and CHWs were the same across the two treatment arms. We observed relatively higher numbers of family members attending the MFG sessions delivered by CHWs compared with PPs, which led to a slight increase in cost of participation incentives provided to families in the MFG-CHW arm. The incremental cost of the treatment over the control arm was US$133 and US$134 for the MFG-PP and MFG-CHW arms, respectively.

Table 2.

Per-child costs of the multiple family group (MFG) family-strengthening intervention in Uganda using the intent-to-treat (ITT) sample, by study arm

Costs	Control arm	MFG-PP arm	MFG-CWH arm
Personnel (salaries)	332,950	665,899	665,899
Identification of schools and school visits	13,712	13,773	13,712
Educational materials	13,039	13,039	13,039
Participation incentives for families	171	232	171
Facilitation incentives for teachers	434	434	434
Snacks and refreshments	68	68	68
Screening and recruitment of study participants	22,512	21,654	20,929
Participation incentives for families	10,313	9,455	8,730
Interviewer compensation	6,147	6,147	6,147
Snacks and refreshments	5,412	5,412	5,412
Facilitation incentives for teachers	640	640	640
Provision of school lunch	22,699	29,098	26,383
Bolstered standard of care	1,569	1,569	1,569
Participation incentives for families	1,102	1,102	1,102
Mental health wellness materials	364	364	364
Facilitation incentives for teachers	103	103	103
Training of MFG facilitators	–	8,567	9,947
Participation incentives for PPs and CHWs	–	6,149	8,451
Facilitation incentives for teachers	–	1,664	741
Snacks and refreshments	–	755	755
Delivery of MFG sessions	–	88,363	97,426
Participation incentives for families	–	57,807	65,020
Facilitation incentives for PPs and CHWs	–	28,621	30,430
Facilitation incentives for teachers	–	1,924	1,965
Snacks and refreshments		11	11
Donated resources	12,473	23,685	23,685
Teachers’ time	316	632	632
Classroom space	12,157	23,053	23,053
Program overheads	40,937	81,875	81,875
Utilities	4,557	9,115	9,115
Transportation	22,036	44,072	44,072
Communication	2,366	4,732	4,732
Maintenance	4,713	9,426	9,426
Materials and office supplies	7,265	14,530	14,530
Stakeholder engagement and dissemination	2,112	4,223	4,223
Meeting space	1,322	2,643	2,643
Participation incentives	790	1,580	1,580
Capital costs	3,205	6,410	6,410
Total costs	452,169	945,116	952,058
Total costs (in 2018 USD)	121	254	255

CWH = community health workers; MFG = multiple family group; PP = peer parents.

The largest drivers of program costs were personnel and overheads across all study arms, accounting for approximately 70% and 9% of total costs, respectively. Personnel costs in the two treatment arms were largely driven by staff time devoted to the supervision of MFG facilitators during the intervention period. The implementation costs of the MFG intervention, including facilitator training, MFG session delivery, donated time by teachers for family mobilization, and donated space for MFG sessions by participating schools, were important drivers of the costs in the two treatment arms, comprising approximately 13% to 14% of total costs.

Table 3 presents per-child costs using the TOT sample, which is based on actual participant numbers in the study arms (control arm, N = 880; MFG-PP arm, N = 733; MFG-CHW, N =  778). This estimation method provides more conservative cost estimates for future cost-effectiveness studies because the TOT sample is smaller than the ITT sample. The total per-child costs for the MFG-PP and MFG-CHW arms were hence higher and estimated at US$346 and US$328, respectively, whereas the incremental costs of the two treatment arms over the control arm were at US$208 and US$190, respectively.

Table 3.

Per-child costs of the MFG family-strengthening intervention in Uganda using the using the treatment-on-the-treated (TOT) sample, by study arm (All costs are in 2018 Ugandan Shillings unless otherwise indicated)

Costs	Control arm	MFG-PP arm	MFG-CWH arm
Personnel (salaries)	378,352	908,457	855,911
Identification of schools and school visits	15,582	18,790	17,625
Educational materials	14,817	17,789	16,760
Participation incentives for families	194	316	220
Facilitation incentives for teachers	550	660	622
Snacks and refreshments	21	25	23
Screening and recruitment of study participants	25,582	29,542	26,902
Participation incentives for families	11,719	12,899	11,221
Interviewer compensation	6,985	8,386	7,901
Snacks and refreshments	6,150	7,383	6,956
Facilitation incentives for teachers	728	874	822
Provision of school lunch	25,795	39,696	33,911
Bolstered standard of care	1,783	2,141	2,017
Participation incentives for families	1,252	1,503	1,416
Mental health wellness materials	414	497	468
Facilitation incentives for teachers	117	141	131
Training of MFG facilitators	–	11,688	12,785
Participation incentives for PPs and CHWs	–	8,388	10,863
Facilitation incentives for teachers	–	2,270	952
Snacks and refreshments	–	1,030	970
Delivery of MFG sessions	–	120,550	125,226
Family incentives	–	78,864	83,573
Facilitation incentives for PPs and CHWs	–	39,047	39,113
Facilitation incentives for teachers	–	2,624	2,526
Snacks and refreshments	–	15	14
Donated resources	14,174	32,313	30,444
Teachers’ time	359	863	813
Classroom space	13,815	31,450	29,631
Program overheads	46,520	111,698	105,237
Utilities	5,179	12,435	11,716
Transportation	25,041	60,125	56,648
Communication	2,689	6,456	6,082
Maintenance	5,355	12,859	12,115
Materials and office supplies	8,256	19,823	18,676
Stakeholder engagement and dissemination	2,400	5,761	5,428
Meeting space	1,502	3,606	3,397
Participation incentives	898	2,155	2,031
Capital costs	3,642	8,744	8,239
Total costs	513,829	1,289,380	1,223,725
Total costs (in 2018 USD)	138	346	328

CWH = community health workers; MFG = multiple family group; PP = peer parents. All costs are in 2018 Ugandan Sshillings unless otherwise indicated.

DISCUSSION

An increasing body of literature demonstrates that family-based mental health interventions seeking to build parenting skills and strengthen family relationships have the potential to improve CAMH outcomes by fostering a supportive family environment that can buffer the effects of high-stress and unstable contexts. ^{10, 22} In an era of rising demand for evidence-based healthcare spending, a demonstration of intervention effectiveness alone is insufficient for policy-making. For countries to prioritize these proven interventions, decision-makers need evidence on the implementation feasibility and affordability for their introduction and scale-up in light of other competing health interventions. ²⁶ Evidence-based decision-making requires an understanding of context-specific resource needs and gaps (e.g., human resources) for successful implementation and a formulation of feasible and acceptable strategies (e.g., task-shifting) to address any identified gaps. Reliable cost estimates are key to public health programming and improving the usefulness of economic evaluations to inform resource allocation decisions. ^{10, 17} Yet the push for building an economic evidence base on mental health interventions is relatively recent in LMICs, and there is a clear evidence gap on the costs and cost-effectiveness of mental health interventions. ¹⁰

To the best of our knowledge, this is the first study to fully explore the economic costs of an evidence-based MFG family-strengthening intervention in a low-resource environment. Although costing studies of health interventions are growing in number, most cost analyses are performed retrospectively and rely on gross costing, and health interventions in low-resource settings are rarely evaluated. ²⁷ A major strength of this costing study is that data were captured prospectively alongside an RCT by trained staff. Furthermore, costing data collection tools and methods used are consistent with the recommendations by the U.S. Panel on Cost-Effectiveness in Health and Medicine. ²⁸ Although labor- and time-intensive, previous research has shown that prospective microcosting analysis improves the reliability and validity of cost estimates ²⁹ and is particularly useful for evaluating costs of new interventions for which no previous cost estimates exist. ³⁰

In SSA, CAMH service and policy gaps are particularly wide. ³¹ Uganda is one of the few SSA countries where CAMH policy guidelines have been developed. Government mental health expenditure in Africa is $0.1 per capita, with 43% of patients paying for mental health services out-of-pocket. ³¹ Given the serious shortage of mental health professionals in LMICs, including Uganda, the WHO’s Mental Health Action Plan 2013–2020 recommends community-based delivery of mental health interventions through task-sharing models with nonspecialist or lay health workers. ³² Training, supervision, and support of these paraprofessionals and quality assurance are critical for the effectiveness of interventions that rely on task-shifting ³³ and are likely to render these interventions resource-intensive.

We expanded the current literature on CAMH-EBPs by estimating the economic costs of an evidence-based MFG intervention delivered through task-shifting in a low-resource environment. Costs of training and supervision of paraprofessionals and costs of donated resources were considered in the analysis. The inclusion of such costs ensured a realistic representation of resource requirements of the intervention in a real-world setting. Our results showed that per-child costs of delivering the MFG intervention by PPs and CHWs was US$346 and US$328, respectively. As anticipated, personnel costs comprised the largest share of the total costs because of intensive supervision and support provided to MFG facilitators and quality assurance efforts related to intervention implementation. To date, only a few cost evaluations of family-based interventions have been published. In Uganda, a family-based economic empowerment intervention seeking to improve mental health outcomes of children and adolescents affected by HIV/AIDS cost between US$418 and US$426 per child over 2 years. ³⁴ Another study reported the cost of running small-scale pilot programs designed to improve parenting practices in Liberia and Thailand; per beneficiary cost of these programs ranged between US$650 and US$900. ³⁵ A recent costing study in South Africa estimated the costs of a parenting program for the prevention of violence against adolescents at US$504 per family. ³⁶ Compared with these interventions, the MFG intervention had a lower cost per-participant; however, there are few available comparisons in the literature. The higher per-child cost of the MFG-PP arm was driven by lower than anticipated attendance. Further, it is unclear whether these intervention costs are acceptable to policymakers in LMICs. If we consider scale-up to a larger target group, the actual costs of replicating the MFG intervention would likely be lower than the costs we reported here. For example, some fixed costs associated with MFG delivery, such as staff training, would be spread over a larger denominator, resulting in lower per-participant costs in those cost categories as target population size increases. It should be noted that, in reality, most costs are fixed for a certain level of activity and increase incrementally when a threshold is exceeded, for instance, costs associated with needing to recruit and train more CHWs or PPs when their maximum capacity is reached, or when the program is expanded. Yet, even with high costs, these interventions may prove to be economically viable due to potentially high cost-savings from reduced risk of future health problems, delinquency, and school dropout, as well as improved school performance, ³⁷ in addition to potentially increased earnings in adulthood. ³⁸ More health economic studies are needed on this topic.

The COVID-19 pandemic resulted in substantial disruptions to implementation in 2020. Adaptations were made to intervention delivery and follow-up assessments to ensure the safety of program staff and participants. Staff spent time on these adaptations, for example, validating a phone interview in lieu of an in-person assessment. In addition, US$3,500 were spent in 2020 on personal protective equipment and disinfectants. These expenses, which amounts to slightly more than US$1 per child, were not included in this cost analysis. However, every program should have contingency funds for these types of unexpected events, which may require unplanned expenses.

There are a few limitations to this study. First, the MFG intervention was implemented in schools through task-shifting in specific districts of Uganda. The cost analysis made use of direct study expenditures, and variation in cost categories may be difficult to gauge across different implementation and geographic settings. Our findings may not be generalizable to other settings or modalities. Second, the microcosting of the MFG intervention was performed in the context of a large-scale evaluation study. Although we took great effort to exclude all purely research costs, this might have resulted in over- or underestimation of costs in some cost categories (e.g., participant recruitment). Some degree of monitoring, evaluation, and quality assurance are necessary for effective implementation of interventions, and should be considered as programmatic costs; however, costs incurred for such activities in research studies may be higher than real-life implementation settings. It is important to prospectively collate and categorize costs as programmatic or research costs to facilitate subsequent analyses. Third, staff time measurement mainly relied on information provided by key program staff. Although this approach minimized the burden associated with collecting time-use data compared with alternative methods (e.g., daily logs), it is potentially subject to bias. To limit this bias, we elicited such information from multiple program staff and used an activity-based costing approach in which we mapped resource use to specific program activities.

Despite these limitations, this study presented much-needed data on the levels and types of resources necessary for community-based implementation of a family-based mental health EBP through task-shifting. Drawing on detailed resource use and cost data, this prospective micro-costing analysis provided reliable cost estimates of delivering the MFG intervention by PPs or CHWs in a low-resource setting and set the stage for its economic evaluation to inform whether one task-shifting approach is a more efficient use of resources than the other and to provide important insights into its scale-up and sustainability in the long-term.