Antiretroviral therapy is a life-long intervention for HIV infected individuals, since no cure is currently available. The treatment presently available in South Africa, is the highly active antiretroviral therapy (HAART). The maintenance therapy is a combination of three drugs in three different regimes (see table 1).
Table 1: Recommended HAART regimens (1)
Stavudine (NRTI) + Lamivudine (NRTI) and Efavirenz (NNRTI)
Stavudine (NRTI) + Lamivudine (NRTI) and Nevirapine (NNRTI)
Zidovudine (NRTI) +Didanosine (NRTI) and Lopinavir (Protease inhibitor)
Since 2004, the South Africa National Department of Health (1) has set in guidelines the CD4 count<200CD4 cells/mm3 or WHO stage IV(2) as medical criteria to consider patient selection and indication for HAART in women and men >= 15 years. A first line and second line HAART treatments along with the routine monitoring, which includes CD4 and viral load count are therein recommended.
In regards planning, in 2007 the department of health of South Africa published the "HIV and AIDS and STI strategic plan for South Africa, 2007-2011", initiative of the South African AIDS council, with the outlined objective, among others, of increasing access to appropriate treatments to up to 80% of HIV infected individuals over five years (3).
Additionally, recent studies are suggesting that the treatment should be initiated at higher level CD4 (4-5) and the latest WHO rapid advice on antiretroviral therapy for HIV infection published online in November 2009 recommend to "start antiretroviral treatment in all patients with HIV who have CD4 count <=350 cells/mm3 irrespective of clinical symptoms" (6). Thus, it can be suggested that a future implementation of the new WHO HIV treatment guideline in South Africa and the ongoing execution of the strategic plan are both going to increase the needs of ARTs within the HIV infected population substantially. A proxy estimate measure of it , limited to the progression of the strategic plan, can be extrapolated from a recent WHO progress survey (7). The latter shows that, over four year period, the percentage of women and man >= 15 years who ever tested for HIV increased respectively from 32.9% to 59.7% (+81.5%) and from 27.6 to 43% (+55.8%). If this increase is compared to the HIV prevalence in the South Africa population for the same age group an approximate can be calculated.
Although an HIV national program has already been implemented in South Africa
Schulpher (8), suggests that economic evaluations should be iterative to allow the maximisation of cost-efficiency. This is even more relevant before a possible extension of the treatment to patients with CD4 <350 cells/mm3. Therefore, the aim of this economic evaluation is to assess, with a health service perspective, the cost effectiveness of HAART for HIV infected individuals starting the treatments with 200<=CD4<=350 cells/mm3 compared to no therapy.
Since it is particularly adapted for the modelling of chronic diseases, a Markov model with the evolution of the disease with and without HAART over 20 cycles of one year was used to model the progression of HIV infected individuals all entering the model when 18 years old and at 200<=CD4<350 cells/mm3 (9). The structure of the model was dictated by the CD4 count in three mutually exclusive states (10), with the chances of disease progression and the treatment effectiveness assumed constant within lower and upper limits at 95% CI and no possibility of reversion to a previous state.
The time horizon was limited to 20 cycles of one year in view of the result of a recent study which showed a median survival of 21.5 years from diagnosis (11), although not generalisable to South Africa, it provides a rough estimate for conservative approach. Considered the provided uncertainty of the treatment effect beyond 10 years, the effect of the treatment was assumed as null from 11 to 20 years. The time horizon was reduced to 10 years in the sensitivity analysis to assess methodological uncertainty and demonstrate the extent of change of the cost utility ICER with the change of time horizon (12) . The HAART treatment in the model was assumed to aggregate the first line and second line regime in the first 10 cycles, but for the sensitivity deterministic analysis and probabilistic sensitivity analysis (PSA), it was assumed that all patients had switched to the second line treatment for the following 10 years with no effect, although the cost was imputed to ensure a relative conservative analysis (13). HAART was therefore considered and analysed as one single category and the model resulted simplified. First and second line treatments were assumed to correspond to South Africa regimen guidelines (1).
The incremental cost-effectiveness ratio (ICER) was the calculation of the ratio of the difference in discounted costs and discounted QALYs (effects) between HAART and non HAART models, see formula adapted from (9) ; were CT = Cost Treatment, CnT= Cost non Treatment, ET= Effects Treatment (measured in QALY), EnT= Effects non Treatment.
The Markov model cannot represent all the variability found in the reality (12), therefore the limits of our model will be considered in the discussion section .
A sensitivity deterministic analysis was performed to deal with different forms of uncertainty (parameter, methodological, model structure). The base-line discount rate of 8% was applied, since it is based on long term government bonds (13) and therefore broadly used in South Africa . In the sensitivity analysis all parameters are kept constant a part of the selected ones. The process was repeated for every parameter, but entry age and cycles of one year were considered as constant. Two different deterministic analyses were performed. First, a one-way deterministic analysis interested all parameters, but health state costs and utilities values, which were subsequently the object of two respective multi-way deterministic analyses (Table 3). Furthermore, the discount rate was modified to 6% and 3% for both costs and effects to provide its effect on the ICER and allow for comparison.
The basic sensitivity analysis provides an impact of a change in one or more variables on the ICER and consequently the relative decision. Nevertheless, these changes are arbitrary and results can be difficult to interpret, in particular if more than one parameter is modified more than one at a time (14).
Patients characteristic where assumed to be uniform in regards, co-morbidity, socio-economic factors and ethnicity (14). All individuals infected with HIV were entering the model at the age of 18 years (+/- 6 months) and with CD4 count 200<=CD4<350 cells/mm3 within 6 months from the start of the treatment to limit variability.
Probability of disease progression was based on CD4 count only for the no treatment model and on CD4 count and treatment effect for the HAART model.
The transition probability data, provided for patients in the no treatment model, were assumed to be natural history data of disease progression in Africa within the four health states (12, 15), whereas for patients in the treatment model, data provided were assumed to result from observational studies in similar settings of patient infected with HIV with CD4 count 200<=CD4<350 cells/mm3 with characteristics comparable to the and being 18 years old and with regimens equivalent to the South Africa Guidelines observed over 10 years period. Treatment failures, lack of adherence or retention to the treatment were assumed included in the direct transitions from states to death.
Treatment costs took into account all the costs in relation with the health service perspective. The costs provided were average costs per cycle of one year and were separated in two categories, first the costs for treatments of Opportunistic Infections (OIs) in the three different disease stages, second the cost of HAART treatment alone.
The cost of treatments included the cost of treating OIs, primary and secondary prophylaxis, as well as the cost of laboratory tests, imaging, and overhead costs (personnel wages, use of hospital beds, clinics and programme cost) (16).
The cost of HAART included costs of the specific clinical and laboratory monitoring, adverse events, and additional blood tests according to the national ART guideline (1) as well as overhead costs.
Cost were expressed in 2008 prices (see tables 1and 3). The average US dollar conversion rate of 2008 was applied (US dollar = 8.248 South African Rands) (17)
Utility scores are assumed to be established on health state evaluations of South African HIV infected patients upon the three stages of disease progression in the Markov model, based on the completion of EUROQOL (EQ-5D) questionnaire (18). In the absence of local evaluations utilities values between 0 (death) and 1 (full health) were established using the time trade-off method of HIV patients in the UK.
The QALY gained with HAART is the result of the difference on the QALY calculation between the HAART model and the non treatment model, which is the sum of utilities values per disease stage multiplied by the life years gained in every cycle adjusted with a discount rate of 8%.
A second-order Monte Carlo simulation was performed to seize parameter uncertainty. The starting age was kept at 18 years and the length of the cycle remained one year since they were considered not probabilistic, whereas other parameters were made probabilistic according to the distribution and the certainty of the variable. We assumed the high and low range of the parameters of table 2 correspond to the 95% confidence interval (CI), where only Standard Deviation (SE) was available the CI was calculated. Uniform distribution was applied to transition probabilities to subsequent stages and to death, gamma distribution was applied to the cost of HAART and the health treatment costs in every disease stage, finally lognormal distribution was applied to drug effectiveness and utility values to allow for more uncertainty on utilities scores derived from a UK population. The model was run 1000 times with all the parameters varying simultaneously. The cost effectiveness acceptability curves (CEACs) obtained provided a summary of uncertainty at different discount rates (12).
We could not find an agreement in regards the ideal willingness to pay (WTP) threshold for QALY gain for South Africa. The WHO (19), although related to DALY gained, proposes a threshold at 1-3 times the country GDP in its guide to cost-effectiveness analysis, and other studies mentioned 2 times the GDP according to the LYs gained (20). Nevertheless, the ceiling ratio for a cost-effective intervention per QALY gained in this economic evaluation was considered to be realistic at twice the per capita Gross National Income (GNI) (21) and a human capital approach was held implicit (22). The threshold was calculated on the basis of the 2008 South Africa GNI per capita in US dollars (23) and rounded up to 11650 US dollars/QALY gained.
Results in table 2 are expressed in LY and QALY gained, both undiscounted and discounted. Considered that a number of studies e.g. (20) have results expressed in LY gained, this will allow for comparability although analysis will be based on QALY only. The ICER per QALY discounted at baseline rate of 8% for HAART versus no HAART was 12323 US$ and was reduced by 571 US$ (-4.4%) compared to the 10 cycles model. In both cases ICERs were above the threshold.
The sensitivity one-way analysis shows ICER per QALY above the threshold for all the variables when the higher range was applied. The same was found with the multi-way analysis individually performed for costs and utilities. Nevertheless, the model was particular sensitive to HART costs (+/- 35%), drug effect (-15 to 22%) and direct transition to death from the first state (-21to 26%) when the higher or lower variable range were individually applied. This is resumed in the bar graph figure 2.
Figure 2: Bar graph sensitivity analysis variation ICER QALY with low-high variable range
The second-order Montecarlo simulation of the baseline model provided a probability of cost-effectiveness at 43% for a WTP at 11650 US$ with both costs and effects discounted at 8%. In comparison, the probability resulted increased to 46% , 49%, and 54% with a discount rate also for cost and effects at 6%, 3% and 0% respectively. The CEACs for the four different discount rates show these variations in figure 3. In regards the cost-effectiveness plane showed that HAART was always more effective but more costly. Comparatively, over 10 cycles the probabilities of cost-effectiveness for the same WTP and discount rates were 35%, 41%,43% and again 43%.
Figure 3: Cost-effectiveness acceptability curve for different discount rates (costs and effects)
There are two main issues to discuss, the methodology and policy issues.
The Markov model shows that individuals infected with HIV and 200<=CD4<=350 have a higher life expectancy with HAART compared to no treatment, hence the chronicity of the condition engenders greater health expenditures, although it can be argued that more expensive hospital treatment are also deferred to later time . The CEAC at 43% for the baseline PSA might be considered unsatisfactory to support decision. On the other hand, HAART cost have shown to be very sensitive and future ARV drug cost reduction up to 30% (24) will have a positive impact on the CEAC . Nevertheless, the Markov model used was too simplified and results are likely to be fallacious since important aspects related to the disease progression were ignored (25). Hence, if uncertainty is too broad sensitivity analysis and PSA might not interpret uncertainty correctly. Thus, to improve the model and obtain more accurate results reducing uncertainty and increasing robustness to inform policy-makers the following measures, although not exhaustive, are suggested (9, 12).
- To introduce natural death risk according to age groups (the starting age was already entered), separated from HIV-AIDS specific mortality to implement time-dependency.
- To add transition to first and second line regimes
- To allow for reversion in a previous state
- To "build memory" with multiple "remission states"
This could be done without the need of further research since some data e.g. life tables (26) are available either from other studies or online. The model would also benefit from sub-grouping according to case-mix profiles (e.g. TBC, HCV) and from cycles reduced to 6 months. For the latter "instantaneous rates need to be derived and, from that, 6-month probabilities calculated" (12). Only then, the need to undertake further research to collect information on uncertain variables will be considered in relation to the expected value of perfect information (EVPI) (14).
Further, we would recommend that future version of the Markov model to evaluate HAART should be based on the latest available researches and possibly with data from observational studies from the same population under study (e.g. utilities values) to allow for results' comparability and generalizability (27).
South Africa has an economy classified as upper - middle income (28) and we considered the threshold of two times the GNI to be cost-effective, however, this approach does not take into account the burden of disease. Considered the HIV estimated prevalence at 15.5-18.4 %, HIV treatment has an important impact on the South Africans' heath care annual expenditures, thus it is also a challenge in term of health policy and planning since it has also "affordability and resource implications" (29). Moreover, opportunity costs of HAART treatment are not measurable in this economic evaluation since in its design is limited to the establishment of the cost-efficiency of HAART for patients starting the maintenance therapy at 200<=CD4 350 and it does not take into account the societal benefits (Nord et al 1995, cited in 30). Looking at equity, Cleary (30), argues that a more cost-effective HIV program could reach lower socio-economic groups achieving greater coverage.
On the basis of this economic evaluation the intervention is 43% effective at 8% discount rate for a WTP of 11650 US$ per QALY gained. Nevertheless, it has been argued that the model does not reflect effects of HAART completely. To provide evidence to policy-makers the model should be improved, and the need for further information assessed in view of the new findings. The CEA based on reasonable modelling will support decision-makers with greater evidence and interpretation of uncertainty (Weinstein et al. 1996, cited in 31). In addition, priorities should be considered by policy-makers in regards equity, opportunity costs and resources implications (also within sectors). With also a particular view on the current economic crisis and its implications in regards development aid from developed countries
1. National Department of Health. National Antiretroviral Treatment Guidelines:Section 1: Antiretroviral treatment (ART) in adults. 2004 [cited 2010 15.03]; Available from: http://www.doh.gov.za/docs/facts-f.html.
2. National Department of Health. National Antiretroviral Treatment Guidelines Appendix 1:World Health Organisation Adults HIV and AIDS Staging System. 2004 [cited 2010 15.03]; Available from: http://www.doh.gov.za/docs/facts-f.html.
3. South African National AIDS Council. HIV and AIDS and STI Strategic plan for South Africa, 2007-2011. 2007 [cited 2010 06.03]; Available from: http://www.info.gov.za/otherdocs/2007/aidsplan2007/khomanani_HIV_plan.pdf.
4. When to start consortium. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. The Lancet. [doi: DOI: 10.1016/S0140-6736(09)60612-7]. 2009;373(9672):1352-63.
5. Kitahata MM, Gange SJ, Abraham AG, Merriman B, Saag MS, Justice AC, et al. Effect of Early versus Deferred Antiretroviral Therapy for HIV on Survival. New England Journal of Medicine. 2009;360(18):1815-26.
6. WHO. Rapid Advice: Antiretroviral therapy for HIV infection in adults and adolescents. 2009 [cited 2010 15.03]; Available from: http://www.who.int/hiv/pub/arv/rapid_advice_art.pdf.
7. WHO. Towards Universal Access: Scaling up priority HIV/AIDS interventions in the health sector 2009 [cited 2010 15.03]; Available from: http://www.who.int/hiv/pub/tuapr_2009_en.pdf.
8. Sculpher M, Drummond M, Buxton M. The iterative use of economic evaluation as part of the process of health technology assessment. J Health Serv Res Policy. [; Research Support, Non-U.S. Gov't]. 1997 1997;2(1):26-30.
9. Briggs A, Sculpher M. An introduction to Markov modelling for economic evaluation. Pharmacoeconomics. 1998;13(4):397-409.
10. Sanders GD, Bayoumi AM, Sundaram V, Bilir SP, Neukermans CP, Rydzak CE, et al. Cost-effectiveness of screening for HIV in the era of highly active antiretroviral therapy. New England Journal of Medicine. 2005;352(6):570-85.
11. Fang CT, Chang YY, Hsu HM, Twu SJ, Chen KT, Lin CC, et al. Life expectancy of patients with newly-diagnosed HIV infection in the era of highly active antiretroviral therapy. Qjm-an International Journal of Medicine. 2007;100(2):97-105.
12. Briggs A, Schulpher MJ, Claxton K. Decision Modelling for Health Economic Evaluation - Oxford University Press. Oxford2006.
13. Drummond MF, Jefferson TO. Guidelines for authors and peer reviewers of economic submissions to the BMJ. British Medical Journal. 1996;313(7052):275-83.
14. Briggs AH. Handling Uncertainty in Cost-Effectiveness Models. Pharmacoeconomics. 2000:479-500.
15. Morgan D, Mahe C, Mayanja B, Okongo JM, Lubega R, Whitworth JAG. HIV-1 infection in rural Africa: is there a difference in median time to AIDS and survival compared with that in industrialized countries? Aids. 2002;16(4):597-603.
16. Beck EJ, Santas XM, DeLay PR. Why and how to monitor the cost and evaluate the cost - effectiveness of HIV services in countries. AIDS - the Official Journal of the International AIDS Society. 2008;22(SUPP/1):S75-S86.
17. Federal Reserve Statistical Release. Foreign Exchange Rates (Annual). 2009 [cited 2010 21.03]; Available from: http://www.federalreserve.gov/releases/G5A/20090102/.
18. Cheung K, Oemar M, Oppe M, Rabin R. EQ-5D User Guide:Basic information on how to use EQ-5D. EuroQol Group; 2009 [cited 2010 16.03]; Available from: http://www.euroqol.org/fileadmin/user_upload/Documenten/PDF/User_Guide_v2_March_2009.pdf.
19. WHO. WHO Guide to Cost-effectiveness analysis. 2003 [cited 2010 06.03]; Available from: https://who.int/choice/publications/p_2003_generalised_cea.pdf.
20. Badri M, Maartens G, Mandalia S, Bekker L-G, Penrod J, Platt R, et al. Cost-Effectiveness of Highly Active Antiretroviral Therapy in South Africa. PLoS Med. 2005;3(1):e4.
21. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. Journal of Health Economics. [doi: DOI: 10.1016/S0167-6296(96)00506-1]. 1997;16(1):1-31.
22. Shillcutt SD, Walker DG, Goodman CA, Mills AJ. Cost Effectiveness in Low- and Middle-Income Countries A Review of the Debates Surrounding Decision Rules. Pharmacoeconomics. [Review]. 2009;27(11):903-17.
23. World Bank. Gross National Income per capita 2008, Atlas method and PPP. 2008 [cited 2010 22.03]; Available from: http://siteresources.worldbank.org/DATASTATISTICS/Resources/GNIPC.pdf.
24. Anso T. New ARV tender taking shape. 2010 [cited 2010 18.03]; Available from: http://www.health-e.org.za/news/article.php?uid=20032641.
25 . Sheldon TA. Problems of using modelling in the economic evaluation of health care. Health Economics. 1996;5(1):1-11.
26. WHO. Life Tables for WHO Member States: South Africa 1990. 2008 [cited 2010 18.03]; Available from: http://apps.who.int/whosis/database/life_tables/life_tables_process.cfm?path=whosis,life_tables&language=english.
27. Drummond M, Manca A, Sculpher M. Increasing the generalizability of economic evaluations: Recommendations for the design, analysis, and reporting of studies. International Journal of Technology Assessment in Health Care. 2005;21(2):165-71.
28. WHO. Classification of low-middle-income countries, 2008. 2008 [cited 2010 06.03]; Available from: http://www.who.int/entity/hiv/data/tuapr2009_annex_countries.xls.
29. Edejer TT-T. Improving the use of research evidence in guideline development: 11. Incorporating considerations of cost-effectiveness, affordability and resource implications. Health Res Policy Syst. 2006;4:23.
30. Cleary S. Commentary: Trade-offs in scaling up HIV treatment in South Africa. Health Policy Plan. 2010;25(2):99-101.
31. Corzillius M, Muhlberger N, Sroczynski G, Jaeger H, Wasem J, Siebert U. Cost effectiveness analysis of routine use of genotypic antiretroviral resistance testing after failure of antiretroviral treatment for HIV. Antiviral Therapy. 2004;9(1):27-36.
32. Mangham L, Hanson K. Scaling up in international health: what are the key issues? Health Policy Plan. 2010;25(2):85-96.
If this essay isn't quite what you're looking for, why not order your own custom Economics essay, dissertation or piece of coursework that answers your exact question? There are UK writers just like me on hand, waiting to help you. Each of us is qualified to a high level in our area of expertise, and we can write you a fully researched, fully referenced complete original answer to your essay question. Just complete our simple order form and you could have your customised Economics work in your email box, in as little as 3 hours.
This Economics essay was submitted to us by a student in order to help you with your studies.
This page has approximately words.
If you use part of this page in your own work, you need to provide a citation, as follows:
Essay.uk.com, Is treatment with HAART for HIV cost-effective in South Africa?. Available from: <http://www.essay.uk.com/free-essays/economics/treatment-with-highly-active-antiretroviral-treatment.php> [04-10-15].
If you are the original author of this content and no longer wish to have it published on our website then please click on the link below to request removal: