Cost effectiveness of facility and home based HIV voluntary counseling and testing strategies in rural Uganda

Background: In Uganda, the main stay for provision of human immunodeficiency virus (HIV) voluntary counseling and testing (VCT) has been at health facilities. Home based VCT on the other hand, was initiated in the country to improve service coverage. Objective: To evaluate the cost effectiveness of facility- and home-based HIV VCT strategies in rural southwestern Uganda. Methods: Data on costs and effectiveness of facility- and home-based HIV VCT intervention strategies was collected in two sub-Counties in rural southwestern Uganda. Costing was performed using the ingredients approach. Effectiveness was measured as the number of HIV sero-positive clients identified. Incremental Cost-Effectiveness Ratios (ICERs) were calculated from the provider perspective. Results: The cost per client tested were US$6.4 for facility based VCT and US$5.0 for home based VCT. The corresponding costs per positive case identified were US$86.5 and US$54.7 respectively. The incremental cost to providers per additional positive case identified by facility based VCT was US$3.5. Conclusion: Home based VCT was the least costly strategy per client tested and was also cost effective in identifying HIV sero-positive clients in rural areas. This strategy should therefore be promoted to improve service coverage and thereby facilitate early and extensive detection of clients eligible for treatment.


Introduction
HIV VCT that relies mostly on individuals presenting themselves and giving voluntary informed consent 1,2 is a key component of national AIDS programs in sub-Saharan Africa 3 , especially in prevention and treatment 4,5 .Individuals or couples undergo pretest counseling, risk assessment, a same-day rapid HIV test, post-test HIV prevention counseling, and referral for medical and support services by trained counselors 6 .Through post-test counseling, high-risk HIV-uninfected individuals are able to re-think about their lives and change their risk sexual behavior while those who are positive can be linked to care and treatment.The timely identification of HIV seropositive clients through VCT and their subsequent enrollment on treatment enables them to be healthier and live longer, reduces the incidence of AIDS-related opportunistic infections and has broader population benefits by reducing onward transmission of the virus 7 .HIV infected clients are also able to access prevention of mother-to-child transmission services, cotrimoxazole and isoniazid prophylaxis 8 .
Different strategies have been used to offer VCT services, including at the work place, through mobile clinics and home-based models 9 .However, the main stay for provision of VCT in Uganda has been at health facilities (facility-based) supplemented by stand-alone clinics (such as the AIDS Information Centre) and mobile outreach programs.Facilitybased VCT is located preferably at a County level health centre (level IV) and above, where capacity and associated HIV/AIDS services are available.However, Sub-county level health centres (level III) with adequate capacity (personnel, supplies and laboratory facilities) also provide VCT.The health facility may be either public or non-government aided such as faith-based health units.VCT services at such sites are integrated into existing health services on a daily basis or if a facility is inadequately staffed, specialized VCT clinic days are established 10 .However, for various reasons particular groups of clients such as the young and men may not be willing to visit health facilities for VCT.Further, with inadequate staffing at health facilities, occurrence of emergencies may interrupt the provision of VCT services 2 .The home based model was therefore initiated as an alternative to improve equity in access to VCT.In this model, counseling and testing is offered within the home.It involves the use of counselors or community health workers who offer and provide counseling and testing to family members 11 .
In Uganda, the estimated number of testing and counseling health facilities per 100,000 of the adult population was 13 in 2010, up from 8.5 in 2009 12 .While the number of people 15 years and older who had received HIV VCT was 18.1% in 2010 slightly up from 16.5% in 2009 12 .The Government of Uganda has set a target of increasing the proportion of people who know their HIV status from currently 38% to 70% by the year 2015 13 .However following the global recession, financial resources for the response appear to have peaked and flattened 7,13 .UNAIDS reports that the future of AIDS resourcing depends on smart investing: spending now to curtail the need to 'spend moreforever' 7 .It therefore becomes critical that decision makers are guided to allocate the available resources to the most cost effective strategies 14 .Studies conducted on cost effectiveness of HIV counseling and testing interventions have predominantly examined particular service provision models [15][16][17] .Further, these studies have used cost per HIV infection averted or cost per quality-adjusted life-year or disability adjusted life year saved as outcomes [15][16][17][18][19][20][21][22] .This article evaluates the cost effectiveness of facility-and home-based VCT models with the primary outcome measure of "extra cost per HIV sero-positive case identified."The use of an intermediate outcome measure makes the study findings more readily interpretable 23 for health service managers at primary care level in making decisions on suitable alternatives for VCT provision in rural communities in Uganda based on operational efficiency 20,24 .

Study setting
A pre-post VCT intervention study was conducted in Rugando sub-County (Mbarara district) and Kabingo sub-County (Isingiro district) of Uganda from November 2007 to March 2008.Mbarara and Isingiro districts have estimated populations of 418,300 and 385,500 respectively 25 .Prior to providing the VCT intervention in both sub-Counties, the services were promoted through the local community leaders.During the study, facilitybased VCT was offered to clients at a health centre level IV in Rugando while home based VCT was offered at the household level in Kabingo sub-county.A detailed description of the study setting and the intervention is given elsewhere 26 .The facility-based and home-based strategies are briefly described below.

Facility based VCT
At the Rugando health centre, clients willing to receive VCT were registered by a nursing assistant in the outpatient department and then referred to the counselor.Pre-and post test counseling was provided by a trained counselor (a midwife).HIV testing was performed by a laboratory assistant in the health centre laboratory.The nationally recommended three test algorithm was used 27 .The first test (HIV Determine I/II) is very sensitive, the second (HIV Unigold I/II) is a confirmatory test run on only positive samples and a third test (HIV Statpak) breaks the tie when necessary.

Home based VCT
Kabingo parish was randomly selected in Kabingo sub-county, and home-based VCT offered in its nine villages.The households to be visited were randomly selected from household lists generated by the village leaders.An adult aged between 18 and 59 years was offered VCT in each of the households.Pre-and post-counseling and testing was provided by integrated trained counselors (ITCs).The ITCs had previously been trained by an Integrated Community Based Initiatives project that provided district-wide door-door counseling and testing in the neighboring district of Bushenyi from January 2005 to December 2006 28 .Both the counselor at the health facility and the ITCs were re-oriented using the national counselor training manual prior to commencement of the study 29 .As with the facility-based strategy, the three HIV test algorithm was used.The two ITCs who provided home based VCT collected test kits and sundries from Rugando health centre at the beginning of every week throughout the one month study period.The ITCs also provided weekly reports to the health facility on the number of clients counseled and tested and supplies used.

Data collection and analysis
A questionnaire was administered to the study participants.Information on socio-demographic profile and HIV risk reduction behavior was collected.Ethical approval for the study was sought and obtained from the Institutional Review Board at Mbarara University of Science and Technology.
Costs A cost analysis using the "ingredients" approach 23 and based on a programmatic perspective considering economic costs was performed.The resources used (type, quantity and unit price) for each of the strategies were documented prospectively during the study implementation.However financial accounts and reports were also analyzed where necessary.Costs of the research exercise such as data collection by the study team were excluded.The cost for each resource was calculated by multiplying quantity by unit cost and total cost was derived by adding up all the individual costs 23 .All costs were expressed in 2008 US Dollars at an exchange rate of 1700.94 per US Dollar (US$) 30 .

Capital costs
Capital costs that included buildings, furniture, and training (home based VCT) were annualized using a discount rate of 3% with an assumed lifespan of 30 (buildings) and 10 (furniture) years 23 .Capital costs for facility based VCT included; buildings (outpatient department and laboratory) as well as the furniture used by clients in these two areas.Training of the counselor (midwife) was not included as a capital cost because counseling is part of the preservice training curriculum for midwives.Capital costs for the home based approach included; building costs (for record keeping) and the initial cost of training the ITCs that was met by the Integrated Community Based Initiatives project.

Personnel costs
Personnel costs were calculated based on monthly salaries or allowances and percent hours spent on VCT activities.The mean time spent on counseling for facility based VCT was 12 minutes (95%CI: 11-13) and 16 minutes (95%CI: 15-17) for home based VCT.The personnel for the facility based approach included; counselor (midwife), laboratory assistant, nursing assistant (who registered VCT clients in the out-patient department), records assistant (who managed health records) and two community mobilizers at the parish level.Personnel costs for home based VCT included; ITCs, the record assistant (responsible for health records returned to the health facility) and nine community mobilizers at the village level.The personnel costs for the two ITCs were calculated based on the outreach allowance for health facility staff who conduct weekly home visits as part of the facility primary health care outreach activities.

Other recurrent costs
The cost of supplies such as test kit and sundries were based on the Joint Medical Stores price catalogue.On-going training costs were incurred for the re-orientation of the counselor and ITCs.While, transport costs were computed as the reimbursement provided to the ITCs during, provision of home based VCT.

Effectiveness
The number of HIV sero-positive clients identified following the VCT intervention was used as the measure of effectiveness.In the era of antiretroviral treatment (ART) and cotrimoxazole preventive therapy, identification of HIV-infected persons through VCT remains pertinent 31 .

Cost effectiveness
Costs effectiveness analysis using a decision model was performed in TreeAge Pro 2009 32 .A cost effectiveness ratio was calculated for each of the VCT approaches, by dividing its cost by the numerical value of the effect (number of HIV sero-positive clients identified).To evaluate the additional cost per HIV sero-positive client identified by the least cost-effective strategy an incremental cost effectiveness ratio (ICER) is reported.

Sensitivity analysis
One way sensitivity analysis was conducted taking into consideration that HIV prevalence was likely to change over time.It has been reported that as prevalence rises, the costs of testing increase, although economies of scale may be observed as well 24 .
Additional sensitivity analysis was conducted by using an alternative discount rate of 5% 33 .

Background characteristics of clients
Nine hundred ninety four (500 for facility-and 494 home-based VCT) clients, were enrolled and tested.The characteristics of the clients by type of VCT strategy are shown in table 1.Both strategies attracted more; women than men, clients who had ever attended school, clients who were ever married, and those who had sex in the 12 months prior to the study.The median monthly income of clients in both groups was similar and the majority tested HIV sero-negative.

Costs
The total cost for provision of facility-based VCT was US$ 3114, while it was US$ 2462 for homebased VCT over the study period.The cost by category and strategy is shown in table 2.

Cost effectiveness and incremental cost effectiveness ratio (ICER)
The number of HIV sero-positive clients identified was 36 (7%) and 45 (9%) for the facility-and homebased strategies respectively.The calculated unit cost per HIV sero-positive client identified for both strategies is shown in table 3.
The unit cost per HIV positive client identified was US$ 86.5 and US$ 54.7 for the facility-and homebased strategies respectively.Based on these costs and effectiveness measures the home-based dominates the facility-based strategy with an ICER of US$ 3.5.

Sensitivity analysis
An increase or decrease in HIV prevalence and change in the discount rate did not influence the cost effectiveness of home-based strategy in identifying HIV sero-positive clients.The home-based strategy remained less costly and more effective.

Discussion
This study reports that the home-based VCT strategy is more effective at identifying HIV sero-positive individuals and at a lower cost to health service providers, than the facility-based strategy.The HIV prevalence for facility-(7%) and home-based VCT (9%) to a large extent reflects the prevalence among the 30-34 year olds reported at the time in Uganda 34 .The cost per client tested for the home-based strategy was comparable to that (US$ 5.88 -8.29) reported in other studies [35][36][37] .Although the cost per client tested in the facility-based strategy reported here was higher than that of home-based VCT, it was lower than that reported elsewhere 31,36,38,39 .The level of care and range of services provided by a facility affects the costs and therefore contributes to the variation in costs observed.
The lower cost per client tested using the home based approach suggests that the home based VCT would be the more suitable approach for improving service coverage when compared to the facility based approach.Another study undertaken in Uganda has also shown that compared to HIV counseling and testing models such as stand alone and hospital based, the home based approach was the least expensive per client tested 36 .
The additional cost per HIV sero-positive case identified by the facility-based strategy was US$ 3.5 which is considered reasonable given that this cost is lower than the estimated expenditure on health per capita for 2007/08 in Uganda of US$ 8.4 13 .However, the current study findings show that the home-based strategy is the most cost effective in identifying those HIV sero-positive individuals in rural areas.It has been suggested that, HIV testing must become simple and ubiquitous as home-based pregnancy test kits 7 .This will ensure that the costs of maintaining dedicated HIV testing and counseling centres are lowered and individuals are empowered to access HIV treatment and care services in a timely and confidential manner 7 .VCT provided in the home environment can be convenient and conducive for counseling and testing and eases the workload on the existing health infrastructure 27 .Further, provision of VCT through a home-based strategy is at no cost to the clients and physical accessibility is ensured by reaching people in their homes 37 and therefore greatly improves service coverage.

Study limitations
The facility-based group was self selected which may have probably been determined by their risk taking behavior.The study findings however show that the HIV prevalence was not significantly different among study participants in the two groups.

Conclusion
Home based VCT was the least costly strategy per client tested and is a cost effective strategy for enabling the early identification and subsequent referral of HIV sero-positive individuals that is critical for Uganda's HIV prevention approach.This strategy will ensure an improvement in HIV counseling and testing service coverage and effectiveness among rural communities.

Table 2 :
Cost by category and VCT strategy VCT= voluntary counseling and testing, FB= facility-based, HB= home-based The unit cost per client tested was US$6.4 and US$ 5.0 for the facility-based and home-based VCT strategies respectively.

Table 1 :
A comparison of the background characteristics by VCT strategy

Table 3 :
Incremental cost effectiveness ratios