Cost and cost-effectiveness of voluntary medical male circumcision in street-connected youth: findings from an education-based pilot intervention in Eldoret, Kenya

Eldoret is the administrative capital of Uasin Gishu (UG) County with a population of nearly 300,000. Eldoret is home to the Moi Teaching and Referral Hospital (MTRH) and the Academic Model Providing Access to Healthcare (AMPATH) program, a large HIV care and treatment program [35, 36]. In 2010, 51.3% of UG County lived below the national poverty line and approximately 52% of the population were age 20 years or younger [37].

Male SCY aged 12–24 who had lived on the street for more than 3 months and spent more than 75% of their days and nights on the street or with other SCY in a shared shelter were approached. SCY who were spending both days and nights on the streets and had limited-to-no guardian contact or who had a caregiver to whom they returned at night were also eligible. Those who had not yet been circumcised were invited to participate in the program. Youth were recruited through a snowball-sampling approach by community leaders and peer navigators who conducted outreach and sensitization activities. Eligible SCY were invited to attend a series of local community assemblies (‘mabaraza’) [38, 39] where the intervention was explained to them in a central location before registering at the start of each intervention cycle. Additional details about the eligible population and recruitment strategy is described elsewhere [33].

The VMMC intervention was piloted in three separate cohort groups over a total of 30 days. Each cohort group participated in the intervention over a 10-day period and received the same program components. Forty SCY consented to participate in groups 1 and 2 and 36 SCY consented to participate in group 3. A target of 40 SCY per group was set based on the number of circumcisions that could reasonably be carried out in a 1-day period. Group 1 was conducted in December 2016 and Groups 2 and 3 were conducted in May 2017. The intervention took place on the grounds of a community organization and rehabilitation center for street-connected boys located on the outskirts of Eldoret.

On day 1, SCY consented or assented to the intervention. Social workers conducted baseline quantitative surveys and a clinical officer (CO) and nurse screened participants to ensure suitability for circumcision. Provider-Initiated Testing and Counselling (PITC) staff tested SCY for HIV. On day 2, the CO and nurses conducted the circumcision surgery in accordance with Kenya’s National Voluntary Medical Male Circumcision Strategy [27]. From days 3–5, the SCY recovered from the procedure under the supervision of a nurse and nurse assistant. From days 6–9 trained mentors facilitated educational modules which focused on life skills, sexual health, and HIV prevention. The modules were taught after circumcision because, following local cultural beliefs, males are thought to enter manhood after being circumcised and it was the hope that SCY would be increasingly open to the subjects that were taught during this time. Teaching the modules after the procedure further allowed wounds to heal under medical supervision and in a sanitary setting. At the end of the modules, SCY completed the end-of-intervention quantitative survey and program evaluation. A graduation ceremony was held on day 10 and included community leaders, the County Children’s Officer, and other support staff and relatives. SCY received a t-shirt and a monogrammed bracelet as an award for completing the program. If a participant indicated wanting to leave the streets, a social worker met with him to begin the process of returning the youth to his family home or Charitable Children’s Institution (CCI) [33, 34].

We used a micro-costing approach, from the perspective of the program implementer, to document and account for resources used to implement the VMMC program. Costs were recorded in real time on an ongoing-basis over the course of the intervention. Costs were recorded in Kenyan Shillings (KES) and later converted to United States Dollars (USD) using the average exchange rate during the 30-day intervention period [40]. All costs are presented in USD ($).

All costing information was based on program implementation records; no human subjects’ data were used in this analysis. Costs above the service-delivery level, such as program management, supply chain and staff training costs were excluded from this analysis. These higher-level costs were largely fixed costs for the overall VMMC program, so they are less relevant for a pilot study and micro-costing approach. Experience from other programs suggested that supply chain costs alone could add significantly to total program costs [41, 42].

The basic assumptions and program information for this intervention are presented in Table 1. All information and calculations used in this analysis are presented in Additional file 1. All unit costs for resources used (inputs) were based on costs to the program (e.g. program invoices), standard staff and volunteer salary scales, and local retail prices (e.g. minor items procured locally in small amounts). Costs were estimated per individual and per each SCY cohort.

Cost of the intervention was itemized by direct procedure costs, procedure support costs, costs of the education component, and research-related costs. Personnel costs included costs to hire the 24 staff recruited to oversee and implement the intervention. AMPATH staff members were paid a daily volunteer rate according to their existing pay grade and volunteer staff were paid based on responsibilities and the local pay scale.

Materials for meals and firewood for cooking were bought locally and in bulk to keep costs to a minimum. All supplies, consumable and reusable, were obtained through the AMPATH procurement department and purchased from the supplier with the lowest quote.

To be able to better inform potential scale-up, we estimated the cost per DALY averted for the VMMC intervention using sensitivity analyses. Price estimates, coverage, and years to scale-up were varied and parameters are reported in Table 1.

The total direct cost of the VMMC surgical procedure was $994 for the 116 SCY, $8.57 per individual. The average cost of medical supplies was $648 and contributed the greatest proportion to the direct procedure costs (65.2%). Total personnel costs were $291, contributing 29.2% to the direct procedure cost. One clinical officer, two nurses and two nursing assistants were involved in the clinical procedure. Clinical staff were paid a daily salary of $9.69 and worked six of the 30 intervention days. Equipment costs amounted to 5.6% ($56) of the direct procedure costs.

Procedure support costs contributed more than 75% ($83.53) to the total intervention cost per individual. Support costs included the cost of supplies to prevent infection after the procedure (e.g. underwear, petroleum jelly, shorts), site rental costs, and costs of non-clinical personnel which accounted for 33% of all procedure support costs. The cost of food for the 10-day intervention period was $18.65 per individual, one-fifth of all support costs.

Education-related costs included personnel costs for mentors to lead modules on life skills, sexual health, and HIV prevention ($10 per individual), costs of education-related materials, and graduation costs of less than $3 per individual. Education costs contributed 13% to the total intervention cost.

Research-related costs were negligible ($1.80 per individual) and included personnel costs and printing costs for consent/assent forms.

The total cost of the intervention for 116 SCY over the 30-day period was $12,526, $108 per individual. Personnel costs totaled $4789 for the 30-day period and contributed the largest proportion to the cost of the intervention (38.2%). Non-clinical personnel were the highest wage earners. The cost to employ four education mentors was $1163 and accounted for 9.28% of the total costs. Cooks, general helpers and security guards received the second highest earnings, $827 each for the 30-day period.

After personnel costs, the greatest expenses related to the retreat components. Personal effects totaled $3415 and food amounted to $2163, contributing 27.3% and 17.3% respectively to the total cost. The cost to purchase shorts for the SCY amounted to 11.6% of the total intervention cost, more than the cost of surgical equipment and medical supplies combined.

The cost to conduct 4 days of education modules per intervention group was driven largely by the cost to employ mentors with minimal costs for materials. Research and evaluation costs contributed less than 1.0% to the total program cost.

Using the WHO’s definition of cost-effectiveness (that is, the cost per disability-adjusted life-year [DALY] saved is less than three times the per capita gross domestic product) [43], the VMMC intervention was highly cost-effective when targeted to male SCY in Kenya (Table 2). Varying the VMMC intervention cost from $8.57 to $108, the cost per DALY averted ranged from $8.5 to $267 per individual and averted more than 8500 new HIV infections in male SCY within 5 years. Scaling-up the VMMC intervention over 10 years averted 16,000 new HIV infections at a cost per DALY averted of up to $141 per individual and increased the intervention’s cost-effectiveness.