SEARCH

SEARCH BY CITATION

Keywords:

  • cost-effectiveness;
  • visceral leishmaniasis;
  • India;
  • treatment;
  • monotherapy;
  • combination therapy
  • ]coût/efficacité;
  • leishmaniose viscérale;
  • Inde;
  • traitement;
  • monothérapie;
  • thérapie combinée
  • costo-efectividad;
  • leishmaniasis visceral;
  • India;
  • tratamiento;
  • monoterapia;
  • terapia de combinación

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

Objectives  To assess the cost-effectiveness of current monotherapies and prospective combinations for treating visceral leishmaniasis (VL) in Bihar, India in terms of years of life lost (YLL) averted as well as deaths averted.

Methods  We employed two methods to estimate the number of avertable deaths in our analysis: one using estimated mortality, the other using direct incidence estimates for VL. Costs of care are based on an average private hospital in Bihar, and data on drug costs were obtained both locally and from the World Health Organization.

Results  The cost of monotherapies per averted YLL ranged from US$2 for paromomycin in an outpatient setting to US$20–22 for AmBisome® at 20 mg/kg. The corresponding costs per death averted ranged from US$53–54 to US$523–527. Combinations ranged US$5–8 per YLL averted and US$124–213 per death averted.

Conclusion  The available treatments for VL are cost-effective, and our mortality and incidence-based methods produce consistent estimates. The combinations considered here were more cost-effective than most monotherapies. Having multiple treatment options and combining drugs are also likely to reduce drug pressure and prolong the drugs’ life-span of effective use.

Projections du rapport coût/efficacité des monothérapies et thérapies de combinaison de la leishmaniose viscérale à Bihar, en Inde

Objectifs:  Evaluer le rapport coût/efficacité des monothérapies actuelles et des traitements prospectifs de combinaisons de la LV à Bihar, en Inde, en termes d’années de vie perdues (AVP) évitées et de décès évités.

Méthodes:  Nous avons utilisé dans notre analyse, deux méthodes pour estimer le nombre de décès évitables: une basée sur l’estimation de la mortalité et l’autre sur des estimations directes de l’incidence de la LV. Les coûts des soins ont été estimés par rapport à un hôpital privé moyen de Bihar et les données sur les coûts des médicaments ont été obtenues au niveau local et de celui de l’Organisation mondiale de la santé.

Résultats:  Le coût des monothérapies par AVP évitée vont de 2$ US pour la paromomycine en ambulatoire à 20-22$ US pour l’AmBisome®à 20 mg/kg. Les coûts correspondants par décès évité varient de 53-54$ US à 523-527$ US. Les combinaisons allaient de 5$ US à 8$ US par AVP évitée et de 124$ US à 213$ US par décès évité.

Conclusions:  Les traitements disponibles pour la LV sont rentables et nos méthodes basées sur la mortalité et l’incidence produisent des estimations cohérentes. Les combinaisons considérées ici étaient plus rentables que la plupart des monothérapies. Le fait d’avoir multiples options de traitement et de combinaisons de médicaments est aussi susceptible de réduire la pression sur des médicaments et de prolonger la durée de vie de l’utilisation efficace de ceux-ci.

Proyecciones de costo efectividad de terapias monoterapias y de combinación para la leishmaniasis visceral en Bihar, India

Objetivos:  Evaluar la costo efectividad de las monoterapias actuales y posibles terapias de combinación para tratar la leishmaniasis visceral (LV) en Bihar, India en términos de años de vida perdidos (AVP) evitados, así como muertes evitadas.

Métodos:  Hemos utilizado dos métodos para estimar el número de muertes evitadas en nuestro análisis: uno utilizando la mortalidad estimada, y otro utilizando estimativos de la incidencia directa de LV. Los costes de cuidados están basados en la media de hospitales privados en Bihar, y los datos sobre costes de medicación se obtuvieron tanto localmente como de la OMS.

Resultados:  El coste de las monoterapias por AVP evitados estaba entre US$2 para la paromomicina en un centro de consulta externa a US$20-22 para AmBisome® a 20 mg/kg. Los costes correspondientes para muerte evitada estaban entre US$53-54 y US$523-527. Las combinaciones estaban entre US$5-8 por AVP evitado y US$124-213 por muerte evitada.

Conclusiones:  Los tratamientos disponibles para LV son costo efectivos, y nuestros métodos basados en la mortalidad e incidencia producen estimativos consistentes. Las combinaciones que se consideraron en este estudio fueron más costo efectivas que la mayoría de monoterapias. El tener múltiples opciones de tratamiento y el combinar medicamentos también podría reducir la presión sobre los medicamentos, prolongando su vida útil y uso efectivo.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

Three countries in the Indian subcontinent (India, Bangladesh and Nepal) hosting half of the world’s burden of visceral leishmaniasis (VL) have agreed to join efforts towards eliminating the disease as a public health problem (WHO 2008). The availability of cost-effective treatments is one of the key elements of the strategy, particularly in Northern Bihar, India, where the traditional first-line treatment, sodium antimoniate, has been lost to parasite resistance (Olliaro et al. 2005). Fortunately, effective alternatives exist: in addition to amphotericin B deoxycholate, oral miltefosine and intramuscular paromomycin are now registered in India and the price of liposomal amphotericin B (AmBisome®) has been reduced from 200 to 20US$ per vial (agreement between Gilead and WHO of 14th March 2007). This creates options for case management with single and combined agents.

However, these treatments have limitations and liabilities. For amphotericin B deoxycholate, drug costs combined with the requirement for intravenous devices and long hospitalization make treatment expensive and impractical; the drug also has immediate and delayed toxicities. AmBisome® remains expensive even at the reduced cost if a full treatment is given, while a single dose of AmBisome® on daycare is less effective. The ability to take miltefosine orally is both an advantage and disadvantage; while oral delivery is more convenient in that it allows the patient to self-treat at home, adherence to 4-week treatment in this outpatient setting may be inconsistent. In addition, restrictions for women of child-bearing age because of teratogenic effects may limit its use, and the cost (for the product on the private market) remains high. Paromomycin is cheap but requires intramuscular injection.

All monotherapies impose selective pressure upon the parasites, possibly leading to resistant mutants being recycled rapidly in areas such as Bihar, where transmission is anthroponotic. Combining drugs with different modes of action could reduce treatment duration, costs and risk of developing resistance without inducing added toxicity.

Visceral leishmaniasis imposes both societal and individual costs to already impoverished societies. These costs can be directly related to treatment, whether medical (drug costs, devices, laboratory tests, etc.) or non-medical (transportation costs, etc.), or they can be indirect (costs incurred due to time lost from productive activities). A study in Bihar estimated that when using Amphotericin B deoxycholate, 55% of societal costs are medical, 29% non-medical and 36% indirect, while for individuals the highest component is indirect costs (59%) followed by medical (23%) and non-medical costs (15%) (Meheus et al. 2006). The latter study documents the substantial economic burden of VL – total costs from the household perspective to manage one episode of VL (median duration of hospitalization: 18 days; median duration of the illness: 70 days) were estimated to be Indian rupees 9420 while the median household income for the sample was only Rs. 2200/month (approximately US$ 209 and 49 respectively in 2006).

The purpose of this paper was to provide cost and effectiveness projections for VL treatment using original data from Bihar. We also estimate the cost-effectiveness of currently available monotherapies and potential combination therapies that are being studied but not yet in routine use. The relative cost-effectiveness of interventions is an integral part of selecting the most appropriate treatment options.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

Treatments

We considered the following monotherapies in our cost projections:

  • • 
    Liposomal amphotericin B (AmBisome®, AB) i.v. infusion at either 2 mg/kg × 5 doses every other day (e.o.d.) (total dose 10 mg/kg), 3 mg/kg × 5 doses e.o.d. (total dose 15 mg/kg), 4 mg/kg × 5 doses e.o.d. (total dose 20 mg/kg), or 5 mg/kg single dose (WHO 2005) (though currently used, there is no formal clinical trial published on the 20 mg/kg dose in India);
  • • 
    Miltefosine (Impavido®, MF) orally 50 mg/day if body weight <25 kg or 100 mg if body weight >25 kg per day × 28 days and
  • • 
    Paromomycin (PM) intramuscular injections at 15 mg/kg/day × 21 days.

We did not consider amphotericin B deoxycholate because, although very effective, it is not an option for wide deployment in Bihar due to limited availability, inconvenience, costs of long hospitalization and toxicity.

All these treatments are registered and available in India. We allowed for delivery of care to inpatients for all treatments, outpatients for paromomycin and miltefosine, and daycare for single-agent AmBisome®. Efficacy rates from clinical trials were derived from Olliaro et al. (2005); compliance in outpatients for miltefosine is derived from a Phase IV trial (Bhattacharya et al. 2007), and is assumed for paromomycin (not formally tested yet).

Costs of drugs were obtained both locally and from WHO (courtesy of Dr J. Alvar, WHO/HTM/NTD). The new preferential price for AB is US$ 20/vial. Miltefosine’s price in the private sector in India is $75 and $150 for the two weight categories but a preferential price has been agreed for the public sector via WHO, depending on the size of the order: $61.5 and $84.5 or, if >75 000 packs ordered, $53.8 and $71.7 respectively. The price of paromomycin is US$10 for an adult and US$5 for a paediatric treatment (Olliaro & Sundar 2009).

Costs of delivering care are calculated for an average private care hospital in Muzaffarpur, Northern Bihar [derived from data collected by S. Sundar locally from a 60-bed charitable facility, the Kala-azar Medical Research Centre (KAMRC), providing free treatment in Muzaffarpour, Bihar]. These cost estimates incorporate admission fees, room and board, medical materials, personnel/physician fees and lab tests (Table 1). Outpatient treatments do not require admission fees or payments for room and board. Daycare treatment requires payment for a hospital bed but does not include food costs or admission fees. Our analysis takes a health system perspective and includes only direct medical costs (drug costs and the costs of care outlined above). Our estimated costs are incremental to those of managing the health programme and staff training as these are likely to be the same across all treatment options.

Table 1.   Unit costs for healthcare delivery for an average private care hospital in Muzaffarpur, Northern Bihar – data collected in 2007
Type of careUnit cost (US$)
Bed (per day)2.0
Food, patient (per day)2.0
Food, caretaker (per day)2.0
Syringes0.2
Tubing1.0
Personnel fees, injections0.2
Personnel fees, infusions1.0
Physician fees (per consultation)2.0
Lab tests4.0
ECG2.0

Potential options for combination treatment are: (i) AB single i.v. infusion of 5 mg/kg followed by MF orally at 50 mg × 2 doses every day for 7, 10, or 15 days; (ii) AB single i.v. infusion of 5 mg/kg followed by PM for 15 or 10 days and (iii) PM + MF for 15 or 10 days. Combinations of these drugs have been tested experimentally in vitro and in vivo (Seifert and Croft 2006); the AB + MF regimens have already been reported in patients with VL (Sundar et al. 2008), others are currently being tested clinically (clinicaltrial.gov identifiers: NCT00371995 and NCT00696969).

We used local anthropometric data (age, weight and sex) of patients with VL (Olliaro & Sundar 2009) to: (i) calculate the number of drug units (vials, tablets) and costs per treatment (for AB we counted all open vials whether used whole or in part, i.e. no vial sharing) and (ii) extrapolate frequencies according to the age/weight distribution of the population and costs of treatment. In short, data were obtained from 1496 consecutive patients presenting with VL between March 2002 and January 2006 at the Kala-azar Medical Research Centre in Muzaffarpour, Bihar. Patients were predominantly young, small and male. Median (inter quartile range) age was 14 (8–30) years and weight was 35 (20–50) kg. The male:female ratio was 6:4. Overall, 41% weighed <25 kg and 59%≥25 kg (the weight categories for MF treatment).

The main characteristics of the treatments used for our calculations are summarized in Tables 2–4. The primary reason for variation in efficacy across treatments is to account for differences in expected compliance. Oral miltefosine is currently delivered in an unsupervised outpatient setting, for which we expect lower adherence to treatment, accounting for a variance in the cure rate (85% in an outpatient setting vs. 94% in an inpatient setting). All monotherapies are highly efficacious (85% or greater), even with adjustments for compliance.

Table 2.   Projections of cure rates and costs of care (direct costs) for treating 100 patients with single-agent treatments in Bihar – 2007
 InpatientOutpatientDaycare*
AB 20AB 15AB 10AB 5PMMFPMMF (market price)MF (WHO price)AB 5
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine; e.o.d., every other day; q.d, once a day; b.i.d, twice a day; IV, intravenous infusion; IM, intramuscular injection; PO, per os.

  2. *Daycare requires that patient stay at the hospital for one day for the infusion but is less costly than treatment as an inpatient.

  3. †Costs of care include unit costs (Table 1) scaled as appropriate for a given treatment. Therefore, cost of care varies across treatments based on the route of administration and the duration of hospitalization.

  4. ‡Total costs are the sum of the cost of care and drug costs, excluding cost of failures and retreatment.

Total dose (mg/kg) 201510515100151001005
Schedulee.o.d.e.o.d.e.o.d.Onceq.d50 mg b.i.dq.d50 mg b.i.d50 mg b.i.dOnce
Hospital stay (days)101010121280001
Route of adminIVIVIVIVIMPOIMPOPOIV
Cure rate (%)94949491949490858590
Cost of care†$14 400$11 400$8400$2000$16 440$20 200$3210$2800$2800$700
Drug cost$28 600$22 950$16 360$6892$745$11 925$745$11 925$7513$6892
Total cost‡$43 000$34 350$24 760$8892$17 185$32 125$3955$14 725$10 313$7592
Table 3.   Projections of costs of care with combination therapies in Bihar
 AB 5 +  MF 7AB 5 +  MF 10AB 5 +  MF 15AB 5 +  PM 15MF 10 +  PM 10
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine. Total costs are the sum of the cost of care and drug costs.

AmBisome (mg/kg)5555
Miltefosine100 mg/kg for 7 days100 mg/kg for 10 days100 mg/kg for 15 days100 mg/kg for 10 days
Paromomycin (days)1510
Cost of care$2300$2700$2700$3100$4400
Drug cost (market price for MF)$9873$10 827$12 855$7642$4435
Drug cost (WHO purchase of MF)$8562$9096$10 231$2704
Total cost (market price for MF; excluding costs of failure)$12 173$13 527$15 555$10 742$8835
Total cost (WHO purchase of MF; excluding costs of failure)$10 862$11 796$12 931$7104
Table 4.   Projections of efficacy with combination therapies in Bihar
 AB 5AB 5 +  MF 7AB 5 +  MF 10AB 5 +  MF 14
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine.

  2. Definitive cure determined at 9 months.

AmBisome (mg/kg)5555
Miltefosine100 mg/kg for 7 days100 mg/kg for 10 days100 mg/kg for 15 days
No. patients enrolled in trial group45454645
Completed treatment45454645
Apparent cure (day 16)45454645
Relapses3112
Deaths1000
Lost-to-follow-up0000
Definitive cure (%)41 (91)44 (98)45 (98)43 (96)
95% CI78–9787–10087–10084–99

Cost estimates for potential combination therapies are noted in Table 3. For combination therapies involving AB and MF, efficacy estimates are reported in Table 4. These estimates are based on trial data from Bihar that indicate high treatment efficacy (>95%) (Sundar et al. 2008). Data on the efficacy of the remaining combination therapies listed above are lacking and hence we have not considered these options in our cost-effectiveness analysis.

Cost-effectiveness calculations

For each monotherapy considered, we estimated the cost-effectiveness of expanding treatment to the uncovered population. Cost-effectiveness analysis provides a way to compare interventions in terms of cost per unit of health gained. The ratio of treatment costs relative to the health gains afforded by treatment can therefore be interpreted as the ‘price’ of purchasing a unit of health using a given intervention. In this analysis, we enumerate health gains in terms of averted deaths or averted years of life lost (YLL), which incorporate discounting (assuming a 3% discount rate) to account for the fact that health losses associated with death occur over the span of predicted life expectancy. Our assumption of a constant 3% discount rate accords with standard guides for cost-effectiveness analysis (Gold et al. 1996). However, to assess the sensitivity of our results to this assumption, we also calculated cost-effectiveness ratios using a 6% discount rate. These estimates are reported in Tables 7 and 8.

Table 7.   Cost-effectiveness of single-agent treatments, assuming a 6% discount rate
 InpatientOutpatientDaycare
AB 20AB 15AB 10AB 5PMMFPMMF (market price)MF (WHO price)AB 5
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine; YLL, years of life lost.

Mortality-based method
 $/death averted$523$418$301$112$209$391$50$198$139$95
 $/YLL averted$36$29$21$8$14$27$3$14$10$7
Incidence-based method
 $/death averted$527$421$303$113$210$394$51$199$140$96
 $/YLL averted$34$27$19$7$13$25$3$13$9$6
Table 8.   Cost-effectiveness of AB + MF combination therapies, assuming a 6% discount rate
 AB 5 +  MF 7AB 5 +  MF 10AB 5 +  MF 14
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine; YLL, years of life lost.The ranges above represent estimates using various efficacy rates: (i) the definitive cure rate reported by Sundar et al. (2008) (also shown in Table 4 above), (ii) the endpoints of 95% confidence intervals for cure rate results reported by Sundar et al. (2008), and (iii) the efficacy for AB monotherapy. Note that the ranges also span the cost range for MF (market price vs. WHO price).

Mortality-based method
 $/death averted$124–160$135–178$139–198
 $/YLL averted$9–10$9–12$10–14
Incidence-based method
 $/death averted$125–161$136–179$150–213
 $/YLL averted$8–9$9–11$10–14

Additional inputs of population-level and disease-specific data were needed to calculate cost-effectiveness ratios. We estimated total incidence for Bihar at 100 000 and stratified by age using the original anthropometric sample data from Olliaro and Sundar (2009). We assumed that current coverage (treatment) rates were 8–10% and that cure rate with current (antimonial) treatment in Bihar is 35% (Sundar et al. 2000). We used a case fatality rate (CFR) of 0.9 for untreated VL (Bora 1999). We also utilized preliminary estimates of VL mortality from the Centre for Global Health Research based on India-specific survey data. To estimate Bihar-specific deaths from this national-level mortality data, we applied the same proportions as VL incidence in Bihar relative to incidence in all of India.

Our cost-effectiveness analysis followed two methods to estimate avertable deaths, one based on estimated VL mortality and another based on estimated incidence of VL. For the mortality method, we calculated avertable deaths as the cure rate (efficacy) of the treatment multiplied by the number of expected deaths, which is based on our calculations of the population-wide, underlying mortality rate. The underlying mortality rate is equal to VL deaths as a proportion of the sum of the untreated population plus the segment of the treated population for whom treatment is ineffective. We used the above-mentioned CFR of 0.9 to estimate the incidence associated with the mortality data and then used these age-specific incidence numbers to scale up per patient costs and estimated the total treatment costs of expanding treatment and re-treating failures.

The alternative, incidence-based method used direct estimates of incidence, CFR, and treatment efficacy to calculate avertable deaths. In this case, we multiplied the efficacy of a given monotherapy by the number of new cases we would expect to result in death based on the CFR and current treatment failure rate. We again scaled up per patient costs by the incidence estimates to project total costs of treatment. Cost-effectiveness results from these two methods are reported in Table 5 for monotherapies and Table 6 for combination therapies. (As noted above, we also estimated cost-effectiveness ratios assuming a 6% discount rate; these results are reported in Tables 7 and 8).

Table 5.   Cost-effectiveness of single-agent treatments
 InpatientOutpatientDaycare
AB 20AB 15AB 10AB 5PMMFPMMF (market price)MF (WHO price)AB 5
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin; MF, miltefosine; YLL, years of life lost.

Mortality-based method
 $/death averted$523$418$301$112$209$391$50$198$139$95
 $/YLL averted$22$18$13$5$9$17$2$8$6$4
Incidence-based method
 $/death averted$527$421$303$113$210$394$51$199$140$96
 $/YLL averted$20$16$12$4$8$15$2$8$5$4
Table 6.   Cost-effectiveness of AB + MF combination therapies
 AB 5 +  MF 7AB 5 +  MF 10AB 5 +  MF 14
  1. AB, liposomal amphotericin B AmBisome®; PM, paromomycin, MF, miltefosine; YLL, years of life lost.The ranges above represent estimates using various efficacy rates: (i) the definitive cure rate reported by Sundar et al. (2008) (also shown in Table 4 above), (ii) the endpoints of 95% confidence intervals for cure rate results reported by Sundar et al. (2008) and (iii) the efficacy for AB monotherapy. Note that the ranges also span the cost range for MF (market price vs. WHO price).

Mortality-based method
 $/death averted$124–160$135–178$139–198
 $/YLL averted$5–6$6–8$6–8
Incidence-based method
 $/death averted$125–161$136–179$150–213
 $/YLL averted$5–6$5–7$6–8

It should be noted that this is a marginal cost-effectiveness analysis, meaning we do not take into account the dynamic feedback effects of treatment on incidence due to limitations in modelling these impacts. In this sense, our cost-effectiveness estimates understate the health benefits of these treatments.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

Direct costs (drug price and healthcare delivery) of treating 100 patients with various monotherapies are presented in Table 2 and Figure 1. The most expensive treatment used was AmBisome® at 20 mg/kg total dose. Hospital-based treatments were more expensive than treatment of outpatients or on daycare.

image

Figure 1.  Direct costs of monotherapies in Bihar. Bar colours indicate the component of cost: cost of care, open bar; drug cost, solid bar. The labels on the Y axis include: Name of the drugs: AB, AmBisome® at 20, 15, 10 mg/kg total dose; MF, miltefosine (m, market price; p, preferential price); PM, paromomycin; type of delivery of care: i, inpatient; o, outpatient; d, daycare.

Download figure to PowerPoint

As shown in Table 5, treatment with paromomycin in an outpatient setting was the most cost-effective single-agent intervention in our evaluation at approximately US$53 per death averted. AmBisome® via single-dose infusion of 5 mg/kg was the most cost-effective inpatient intervention at US$112 per death averted and AmBisome® in a daycare setting was notably cost-effective as well at < US$100 per death averted. The cost-effectiveness of treatment with AmBisome® quickly degrades as total dosage increases, which also causes drug costs to increase. It should be noted that while current international guidelines recommend a total dose of 20 mg/kg in other settings, 10 or 15 mg/kg are allowed in India (WHO 2007). The cost-effectiveness of these interventions in terms of the cost per averted years of life lost ranges from US$2 per YLL averted for PM in an outpatient setting to US$20–22 per YLL averted for AB at 20 mg/kg.

Cost-effectiveness results for the combination therapies involving AmBisome® and miltefosine are shown in Table 6. Ranges are presented to account for the fact that the efficacies reported by Sundar et al. (2008) are based on an inpatient trial while in practice the miltefosine component of the therapy would be taken in an outpatient setting (likely lowering compliance, and therefore efficacy as well).

In addition to estimates based on the reported efficacies for these combination therapies, we produced estimates using the efficacy for a single-dose of AmBisome® at 5 mg/kg only (i.e. assuming ‘zero compliance’ for the miltefosine component of a combination therapy) as well as efficacies corresponding to the 95% confidence intervals reported by Sundar et al. (2008). Table 6 presents the cost-effectiveness results for these scenarios. At US$124–213 per death averted, the combination therapies were more cost-effective than most monotherapies in our evaluation primarily because of their greater efficacy.

As shown in Tables 7 and 8, under the assumption of a 6% discount rate, our cost-effectiveness estimates for monotherapies (Table 7) and combination therapies (Table 8) are quite similar to our primary results, which assume a 3% discount rate. As expected, the results based on a 6% discount rate show a slightly higher cost per YLL averted.

Visceral leishmaniasis treatments are highly cost-effective largely because of their relatively low costs. Cost-effectiveness ratios are also influenced by age patterns of VL incidence: nearly 90% of cases occur in people younger than 50 years and over 30% of cases occur in patients 20 years old or younger. Successful treatment at these ages still affords patients many expected years of life.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

The purpose of this study was not to identify a single ‘winning’ intervention for VL but to present a comparative framework and initial estimates to support decision-making on appropriate VL treatment in Bihar. Promoting the use of only one intervention will impose selective pressure potentially leading to the emergence of resistant strains. The fact that our analysis identifies various cost-effective options that can be delivered to outpatients as well as inpatients may be beneficial from a public health perspective. We also identify several combination therapies that are competitive with monotherapies in terms of cost-effectiveness. When the value of delayed emergence of resistance, which would accompany the use of combination therapies, is also considered, these options appear quite promising. Furthermore, given that the total cost of all combination therapies, not just those for which we have efficacy estimates, is well within the range for available monotherapies, we might expect that these treatment options will also have favourable cost-effectiveness ratios if the cure rate is proven to be fairly high. Even in this analysis, the similarities in efficacy across monotherapies suggest that variations in costs are driving the cost-effectiveness results. While changes in drug costs will be inconspicuous for paromomycin, the margins are still sizeable for AmBisome® and miltefosine – for the latter price is uncertain due to a recent change in manufacturer.

However, practical considerations and indirect costs should also be brought into the equation. Inpatient treatment over 10–28 days is likely to impose a significant burden, potentially beyond the costs enumerated in this analysis, as it would prevent the patient and members of the family assisting them to earn money while at the hospital away from home (Meheus et al. 2006). Other costs should also be considered; for instance, the costs of averting pregnancy if treating with miltefosine, a potential teratogen (Sundar & Olliaro 2007).

Our results concur with the existing cost-effectiveness literature. Cattand et al. (2006) estimate the average cost of VL treatment in India per disability-adjusted life year averted to be US$18.40 (for treatment with either miltefosine or amphotericin B). Vanlerberghe et al. (2007) find that a miltefosine-based strategy costs US$328, which is similar to our estimate of US$391–394 (note also that the treatment we consider involves a higher total dosage). The variance in our cost estimates (and the inclusion of a daycare treatment option) explains the difference in our finding that treatment with AmBisome® ranges in cost-effectiveness from US$95 to US$527 per death averted while Vanlerberghe et al. find AmBisome® to be the least cost-effective of the interventions they considered at US$1622 per death averted.

The reported cost-effectiveness ratios for all interventions are independent of the estimation method (mortality-based or incidence-based) employed, suggesting significant consistency between the mortality data derived from verbal autopsies and incidence estimates derived from the existing literature.

It is not possible to predict the minimum required threshold of efficacy below which the intervention will no longer be considered effective and resistance is generated. Furthermore, these drugs have different modes of action and are intrinsically different in their vulnerability to parasite resistance. Insufficient doses proved to have caused resistance to sodium stibogluconate in this area where resistant strains are circulated rapidly through inter-human transmission (Olliaro et al. 2005; Alvar et al. 2006). Therefore treatments which fail to clear parasites in one of every 10 patients (AmBisome® single dose, possibly Miltefosine while deployed without supervision) may not be cost-effective in the medium to long-term. This further supports combination therapies.

Cost-effectiveness ratios provide an important comparative tool for making health policy decisions when several interventions are available. This analysis should help illuminate policy options regarding both healthcare delivery and health financing. Given the remarkable cost-effectiveness of all these interventions when compared in terms of $US/YLL averted, researchers, policy-makers and healthcare professionals should explore the issue of how to cycle these drugs so that the supply over time and across locations is diversified. Additionally, this analysis illustrates the importance of cost-equalization policies (note the improvement in the cost-effectiveness of miltefosine under the WHO price) and provides guidance as to the magnitude of subsidies or other supply support policies that would be required to encourage the use of certain treatments.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

We thank Jorge Alvar for advise and information provided and Dan Chisholm for critically reviewing the manuscript. We thank the Centre for Global Health Research for providing mortality estimates.

Disclaimer

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References

The opinions expressed in this paper are those of the authors and may not reflect those of their employing organizations. PO is a staff member of the WHO; the authors alone are responsible for the views expressed in this publication and they do not necessarily represent the decisions, policy or views of the WHO.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclaimer
  9. References