Performance, costs and cost-effectiveness analysis of the Tay Ho HIV integrative prevention and care & treatment outpatient clinic, Vietnam. : is the model worth scale up?

Since the early 1990s, Vietnam has been experiencing an HIV/AIDS epidemic with a general prevalence rate of 0.42 % in 2008 and a projected prevalence rate in 2012 of 0.47%. Although the general HIV prevalence rate is considered low, the virus heavily affects some at-risk population groups in Vietnam, including commercial sex workers, injectable drug users and the men who have sex with men. According to studies, prevalence among these groups is considerable, up to 65%. Risky sexual behaviours are common practice in all three groups, and the chain of infection is intertwined. Besides, the situation is rendered even more complex in respect of HIV transmission prevention and PLHA treatment because of a governmental zero tolerance policy in respect of drug-users and a high level of stigma and discrimination towards PLHA in the Vietnamese society.
In that context, the active fight against HIV in Vietnam began in earnest at the end of 2003, with the arrival of one major international donor scheme, namely PEPFAR, followed in 2006 by the start of disbursement of the Global Fund, and the active work of the World Bank and DFID and other bilateral agencies, amounting to US$ 114 million in the single year 2008. At the same time, the Vietnamese government was spending US$9.7 million on the fight against HIV amounting to less than 1% of the Ministry of Health’s budget. As a consequence, Vietnam became highly dependant on international aid to finance its fight against the epidemic. Meanwhile, it is estimated that about 30% of the needs are covered in terms of prevention activities and medical assistance to PLHA.
There is thus a critical need for identifying the most cost-effective models of intervention in the Vietnamese context to help scale-up programmes in the country and meet the needs in respect of both prevention and treatment.
It is in this context of limited resources and high social barriers for at-risk population that the French non-governmental organisation, Médecins du Monde, developed an integrated prevention and care model, implemented at the end of 2005 in Tay Ho, a district of Hanoi. The MDM has undertaken both financial and technical support and the main components of its model include a prevention component consisting of a mobile outreach team and the VCT unit, and a care and treatment department including adherence training, support through home-based care and HAART.
The assumptions that led Médecins du Monde to implement such of model of action were that integrating prevention, detection, and care & treatment services within the same structure would help better targeting and attracting at-risk populations, hence increase programme performance, and finally build a cost-effective response through cost-savings and internal programme synergies.
Goal and objectives
The goal of this research is to test the hypothesis on which this model of intervention relies: that the integration of outreach, detection and care & treatment components within the same outpatient clinic, in the Vietnamese context, results in a high caseload of at-risk clients and patients along with structural economy of scale, translating in high cost-effectiveness levels for the model’s key components.
As such, the goal of the research translates into the following objectives:
->Assessing model’s outputs by analysing prevention, testing and care & treatment components performance in term of provision, coverage, utilisation and impact
->Analysing central components of the model (VCT and HAART) cost- effectiveness, vs. the null-set scenario, and standards reflected in international literature
->Analysing potential sources of synergies within the program and their impact on the cost-effectiveness level of its key components
Method
This research is based on a bottom-up three-layer analysis:
->one related to each programme’s component performance and output;
->another related to each programme component's costs and unit costs; and
->a third related to the cost-effectiveness analysis of the programme's two central components, the VCT and the care and treatment services
Results are presented as follows:
->A first part presenting performance results
->A second dedicated to the financial and economic analysis, laying out:
->Model’s components financial unit-costs
->Key components – VCT and care & treatment – cost- effectiveness analysis, with two sub-parts, a first one on the analysis of synergies within the model based on financial unit- costs analysis, and a sensitivity analysis based as well on financial unit-costs
For each layer, the method of data collection and analysis is designed to address some field-related constraints including that:
->the research is partly based on retrospective data;
->the field is not designed to conduct academic research; and
->the M&E system at field level is limited and cannot be adapted for the purpose of the research.
The theoretical foundation of the thesis is founded on:
->Habicht’s guidance on the development of programme performance indicators in terms of provision, coverage, utilisation and impact;
->the World Health Organisation-CHOICE guideline on cost-effectiveness analysis; and
->an adaptation of the step-down accounting methodology to allocate indirect
costs in a systematic fashion and ensure transferability of the results
Findings
The underlying hypothesises supporting the implementation of that model of intervention combining prevention and care & treatment components proved true.
The model promoted strong synergies, which contributed to the increase in numbers of patients attending at the OPC level. Critical harm reduction activities could be carried out directly in the city’s hotspots while at the same time the mobile outreach team was identifying potential PLHA in need of a treatment. The concentration of these two functions within one team reduced the cost per client referred for VCT, and helped to raise awareness of existing medical services offered by the OPC targeting directly the most at-risk populations.
The integration saved as well costs by boosting the demand for the clinic’s services and the use of the significant resources invested in the setting up of such a model in term of fixed capital and trainings. The model worked as a system with positive feedback loops preventing new infections and actively treating identified People Living with HIV/AIDS through levelling off social barriers. This system worked not only from outreach to treatment, but certainly as well from treatment to outreach by increasing VCT attraction, at-risk persons being aware of the presence of immediately free medical services within the same structure.
As calculated in our research, the average ICR of the VCT unit vs. the null set scenario was 12 I$/DALY(3,0) averted, well below World Health Organisation- CHOICE SEAR indicator of 40 I$, and 252 I$/DALY(3,0) averted for the care & treatment unit, well in line with international standards. A model in which outreach and detection services were not integrated with care and treatment service would have increased unit costs (by a factor of four (4)), resulting in the medical component running costs per DALY averted far below international standards. The same would have been the case had the VCT unit not been integrated with the mobile outreach service, at least in the first two (2) years of the programme's operation. Integrating the mobile outreach team with the services offered by the VCT unit, cut costs to the latter by a factor of three (3). Nonetheless factors related to adherence to treatment and the delay in identifying patients for testing and treatment hampered the global cost- effectiveness of the programme.
Conclusion
The model is cost-effective, yet limited.
First, the demonstrated synergy highly depends on the context in which the programme operates. Were the prevalence in the target population to decrease below 15%, the synergy between the mobile outreach team and the VCT unit would begin to disappear. Moreover, were VCT services to be mainstreamed in Vietnamese society, the extra-cost incurred by the work of the mobile outreach team would hinder this synergy. Second, because of contextual limitations the model showed only an average cost-effectiveness by international standards, especially within its medical component. The model was unable to retain pre-ARV patients in sufficient numbers, or to convince them to abide by the OPC protocols in the absence of critical
complementary services, such as Methadone Maintenance Therapy, and/or early access to ART. The introduction of a Methadone Maintenance Therapy in an environment in which about 60% of PLHA are opiate-users would change dramatically the outcome of HAART, not to mention reducing HIV transmission. Third, in our views, the main limiting issue of this model might be the intense technical support it needed to be implemented and supervised. Indeed, the presence of an external NGO, such as MDM, though necessary in the international co-operation scheme, added critical costs to that programme. Over three years, the share of NGO expenses was considerable, amounting to 58.1% of the total. This cost share reflected the complexity of setting up the programme in the Vietnamese environment and the necessity to channel international funding, control spending, report to donors, and ensure the overall technical supervision of the model. Besides, costs also rose because the general NGOs co-operation system in Vietnam creates significant market distortions as a result of a limited local pool of skilled labour creating niche job markets. It is hence likely that the international system in place inflated costs at the NGO level by creating not only job-market distortions, but also several time- consuming tasks, such as reports, proposal writing, seeking fund prolongation agreements, and juggling different accounting and report norms.
As such, the question remains on how transfer both financial and technical burden to local authorities in a context of limited resources.
The Vietnamese government spends US$1,100,000,000 on health care according to official figures from the National Office of Statistic, representing an expense per citizen of US$13.75, including general administrative costs. The sole medicine cost if the current number of PLHA (240,000) in Vietnam were to have access to first-line HAART rises to a minimum US$24,000,000 per year (or 3% of the total health budget), excluding medicines and management costs. Apply the model’s average cost to follow-up a patient for one year of HAART, including medical management and biological follow-up in an optimal situation (average caseload of 750 patients), and that cost would exceed US$200,000,000 a year, (or almost 20% of the annual health budget). This excludes integrating general supervision and management costs, which, depending on the efficiency of the system put in place by the Vietnamese authorities, could add an extra 30% to the total.
It seems that in the long term, the matter of the context and technical assistance are central. Though cost-effective and well adapted to the current constraints of the Vietnamese environment, the Tay Ho OPC approach is only a short-term solution until prevention and detection activities are mainstreamed and social obstacles lifted off. It could well be the best model to address HIV/AIDS in the Vietnamese context, or in any other places where concentrated epidemics are evident to quickly break an epidemic. Yet, the issue of the social and financial sustainability of such models remains and should be specifically explored. As such, it appears that research in the future should start focusing not only on the best mix of activities, but on the best model of technical assistance delivery, transfer and sustainability.