To forecast demand for HCV medication under several scenarios (over a 10-year period)

Transcription

1

2 Objectives To forecast demand for HCV medication under several scenarios (over a 10-year period) To assess the responsivenness of the demand for HCV medication to changes in prices of the medicines How are changes in the prices of drugs (either through the introduction of generics or subsidization by a third party) will affect the demand? 2

3 Outline Methodological Approach o Mathematical Model Two Base Cases for Treatment Cascade o Current Best Regimen vs. Future Best Regimen Two Financing Scenarios o Free-Market vs. Third Party Coverage Working Assumptions Main Results: o Low-income, Lower-Middle & Upper-middle income countries Sensitivity Analysis Discussion & Policy Recommendations 3

4 Methodological Approach A Susceptible-Infected-Removed (SIR) Model (e.g. Elbasha 2013): New infections are assumed to be of the naïve nature (No reinfection) Population is assumed to mix homogeneously N Susceptible S Acutely infected I Immune R Chronically infected F m 4

5 Methodological Approach ν Susceptible S(t) x Acute HCV I(t) ε 0 F0(t) ε 1 F1(t) ε 2 F2(t) ε 3 F3(t) ε 4 F4(t) λ σ μ μ μ μ μ μ Recovered R(t) μ Mortality Unrelated to HCV μ 5

6 Methodological Approach The chronically infected go through the diagnosis cascade starting with HCV detection and ending once treatment is accepted. Outline of the Stages of Detection, Referral to Care and Treatment Chronically Infected α 1 HCV Detected α 2 Referred to Care α 3 Fibrosis Staging α 4 Medically Eligible Patients (based on scenarios) α' 5 Health System Absorption Capacity α" 5 Price α'" Elasticity Ability-to- 5 of Demand α 5 Pay (Among those with ATP) Treatment Received 6

7 Methodological Approach: Diagnosis Rate (r 0 = α 1 *α 2 *α 3 ) Diagnosis rate reflects the probability that a chronically infected individual would have had HCV detected been referred to care and gone through fibrosis staging. For the current best case, this rate (r 0 ) is assumed to be 5% in the year 2013 Stage 2013 to 2014 to 2015 to 2016 to 2017 to 2018 to 2019 to 2020 to 2021 to 2022 to Acute 0,1% 0,1% 0,1% 0,2% 0,2% 0,2% 0,2% 0,3% 0,3% 0,4% F0 0,5% 0,6% 0,7% 0,8% 0,9% 1,0% 1,2% 1,3% 1,5% 1,8% F1 1,0% 1,2% 1,3% 1,5% 1,7% 2,0% 2,3% 2,7% 3,1% 3,5% F2 2,5% 2,9% 3,3% 3,8% 4,4% 5,0% 5,8% 6,7% 7,6% 8,8% F3 10,0% 11,5% 13,2% 15,2% 17,5% 20,1% 23,1% 26,6% 30,6% 35,2% F4 15,0% 17,3% 19,8% 22,8% 26,2% 30,2% 34,7% 39,9% 45,9% 52,8% For the future best case, this rate (r 0 ) is assumed to be 15% in the year 2013 Stage 2013 to 2014 to 2015 to 2016 to 2017 to 2018 to 2019 to 2020 to 2021 to 2022 to Acute 5,0% 5,8% 6,6% 7,6% 8,7% 10,1% 11,6% 13,3% 15,3% 17,6% F0 10,0% 11,5% 13,2% 15,2% 17,5% 20,1% 23,1% 26,6% 30,6% 35,2% F1 15,0% 17,3% 19,8% 22,8% 26,2% 30,2% 34,7% 39,9% 45,9% 52,8% F2 20,0% 23,0% 26,5% 30,4% 35,0% 40,2% 46,3% 53,2% 61,2% 70,4% F3 25,0% 28,8% 33,1% 38,0% 43,7% 50,3% 57,8% 66,5% 76,5% 80,0% F4 35,0% 40,3% 46,3% 53,2% 61,2% 70,4% 80,0% 80,0% 80,0% 80,0% 7

8 Methodological Approach: Eligibility(α 4 ) The probability that an individual will be offered treatment is based on their eligibility for treatment. All individuals eligible for treatment are offered treatment after fibrosis staging. α 4 is the probability that the individual will fulfil the eligibility criteria (depending on WHO (2014) guidelines). Scope of eligibility for treatment: Initial analysis: only fibrosis stage F3 & F4 Sensitivity analysis: treatment of everyone who is chronically infected (F0 to F4). 8

9 Methodological Approach: Treatment Rate (α 5 ) Individuals are at liberty to either accept or refuse treatment based on conditions: Side effects of the drug (acceptability of treatment) Costs involved (Ability-to-Pay (ATP)) Sensitivity analysis to take into account differential acceptability of treatment & various thresholds for ATP In the Two Cases: of people willing and able to accept treatment depends on three factors: o Health System s Absorption Capacity (AC) α 5 o Individuals Ability-To-Pay (ATP) α 5 Treatment rate α 5 o Price elasticity of demand (ξ d q) α 5 9

10 1 0 Objectives Methodology Results Discussion Methodological Approach: Treatment Rate (α 5 ) For the current best case, the initial absorbtion capacity of the health system is assumed to be 5% in the year 2013 and increasing (annually) by 5% over time, i.e. AC t = AC t For the future best case, the initial absorbtion capacity of the health system is assumed to be 5% in the year 2013 and increasing (annually) by 15% over time, i.e. AC t = AC t (with a maximum 80% capacity at end of period) The price elasticity of demand measures the change in demand as a result of a 1 percent decrease in prices. For the current best case, it is assumed that elasticity is 0.9 for the highest income quintile and reduces to 0.5 for the lowest income quintile in each country. For the future best case, it is assumed that elasticity is 0.9 irrespective of the income quintile. Individuals are considered as able-to-pay for medication if the cost of the regimen represents at most 40% of their per-capita GDP.

11 1 1 Objectives Methodology Results Discussion Methodological Approach: Prices & Treatment For the initial analysis drugs are assumed to cost the same in all countries irrespective of their income levels. Current lowest prices are used as the reference prices Prices are assumed to fall by an annual rate of 3.4% as in the price of brand ARVs. The total cost of medication will depend on the genotype. Genotype Cost (12 Weeks) in USD 1, 4 Sofosbuvir 840 Ribavirin 53 Peg-IFN Sofosbuvir 840 Ribavirin 53 3 Sofosbuvir 840 Ribavirin 53 All Genotypes Sofosbuvir 840 Another DAA 420 Recommended Duration (Weeks) Current Best WHO (2014) 12 Future Best

12 1 2 Objectives Methodology Results Discussion Treatment Module: Two FINANCING Scenarios: (Free Market vs. Subsidised) Free Market Prices: Demand for drugs depends on ATP Subsidised: Third party willing to cover a fraction of the avg. costs (say, e.g., 50%) Demand for drugs based on ATP after subsidies have been applied. Major sensitivity analysis Changes in Eligibility Criteria Changes in third-party subsidies Changes in ATP thresholds Changes in price reduction rates

13 Countries Used in Analysis Low Income Countries ($664) Lower-Middle Income Countries ($2751) Upper Middle Income Countries ($9025) Included Benin, Burkina Faso, Gambia, Guinea, Guinea Bissau, Liberia, Mali, Mauritania, Niger, Sierra Leone, Togo, Central African Republic, Chad, Democratic Republic of Congo, Burundi, Comoros, Ethiopia, Kenya, Madagascar, Malawi, Mozambique, Rwanda, Tanzania, Uganda, Zimbabwe, Haiti, Afghanistan, Bangladesh, Nepal, Cambodia, Kyrgyz Republic, Tajikistan Cape Verde, Côte d Ivoire, Ghana, Nigeria, Senegal, Cameroon, Congo, São Tomé & Principe, Lesotho, Swaziland, Djibouti, Zambia, Egypt, Morocco, Iraq, Sudan, Yemen, Bolivia, Paraguay, Belize, El Salvador, Guatemala, Honduras, Nicaragua, Guyana, Pakistan, India, Bhutan, Sri Lanka, Indonesia, Lao People s Democratic Republic, Philippines, Vietnam, Fiji, Papua New Guinea, Armenia, Georgia, Moldova, Ukraine, Uzbekistan, Albania, Mongolia Angola, Gabon, Botswana, Namibia, South Africa, Seychelles, Algeria, Islamic Republic of Iran, Jordan, Tunisia, Argentina, Brazil, Chile, Colombia, Ecuador, Peru, Uruguay, Venezuela, Costa Rica, Mexico, Panama, Dominican Republic, Jamaica, Suriname, Maldives, Malaysia, Thailand, China, Azerbaijan, Belarus, Kazakhstan, Turkmenistan, Russia, Bosnia & Herzegovina, Serbia, Montenegro, The Former Yugoslav Republic of Macedonia, Turkey, Bulgaria, Latvia, Lithuania, Romania Excluded Eritrea, Democratic Republic of Korea, Myanmar, Somalia Kiribati, Kosovo, Marshall Islands, Federal States of Micronesia, Samoa, Solomon Islands, South Sudan, Syrian Arab Republic, Timor-Leste, Tonga, Vanuatu, West Bank & Gaza Cuba, Dominica, Grenada, Lebanon, Libya, Mauritius, Palau, St. Lucia, St. Vincent and the Grenadine, Tuvalu 1 3

14 1 4 Objectives Methodology Results Discussion Data Sources Data & Variables Economic-related Data Share of GDP held by each quintile (income per capita (share of GDP /# of i's in each q) GDP overtime Population related Data: size, mortality, growth rate Prices of medication Recommended therapy per genotype Transition probabilities Incidence rates Sensitivity Parameters: treatment option, rate & type of third party coverage, & rate of price decrease) Data Sources World Development Indicators (WDI) International Monetary Fund s World Economic Outlook Database (April 2014) World Development Indicators (WDI) WB Projection Tables; World Health Statistics (2013) WHO guidelines (2014), Open Society pharmaceutical companies & other media sources exogenous & given by the natural history of the disease exogenous and obtained from literature exogenously determined but allowed to vary based on assumptions made NA: extrapolated using geography, genotypes, GDP

15 Main Results Some Definitions Expenditure: This refers to the total (cumulative) expenditure on drugs at the end of the period. Regimen: This is the average cost of the drug regimen at the end of the period for enduser (patient). 15

16 Results: Base case of Low Income Countries (LICs) [price decrease = 3.4, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] of Individuals (Annual) in 000s of Individuals in 000s Current Best Future Best LIC Regimen ($) LIC Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 16

17 Lessons from LIC Base Case Scenarios LIC markets are and will remain limited in size - there are fewer and fewer LICs - HVC prevalence is relatively low In the most pessimistic scenario (current best regimen, free market) there will still be over patients treated by 2023, for 400 Million USD over 10 years In the most optimistic scenario (future best regimen, 50% subsidy) there would be almost 2.2 million people treated by 2013, for 2.2 Billion USD over 10 years. In none of 4 scenarios we see the number of new patients treated annually decline over time. Alternative actions (steeper price declines, higher levels of subsidy, improved diagnostic access and health system performance are needed to realize meaningful impact on the epidemic in LIC. 17

18 Results: Base case of Lower-Middle Income Countries (LMICs) [price decrease = 3.4, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] of Individuals (Annual) in 000s of Individuals in 000s Current Best Future Best Cummulative LMIC Regimen ($) Cummulative LMIC Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 18

19 Lessons from LMIC Base Case Scenarios LMIC markets will be massive In the most pessimistic scenario (current best regimen, free market) over 13 Million patients will have been treated by 2023, for a total of almost 15 Billion USD over 10 years. In the most optimistic scenario (future best regimen, 50% subsidy) over 30 Million patients will have been treated, for a total of almost 33 Billion USD over 10 years The number of new patients treated decline over time only in the future best regimen case scenarios with 50% subsidy (most optimistic) For public health impact either lower prices than projected, or subsidy for treatment will be needed, along with improved access to diagnostics and health system performance. 19

20 Results: Base case of Upper-Middle Income Countries (UMICs) [price decrease = 3.4, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] of Individuals (Annual) in 000s of Individuals in 000s Current Best Future Best UMIC Regimen ($) UMIC Regimen ($) By Free By Free Market By Market By By Subsidised By Subsidised By % By % 20

21 Lessons from UMIC Base Case Scenarios UMIC markets will be massive too and may reach saturation of demand by 2023 In the most pessimistic scenario (current best regimen, free market) over 23 Million patients will have been treated by 2023, for a total of over 26 Billion USD over 10 years. In the most optimistic scenario (future best regimen, 50% subsidy) over 29 Million patients will have been treated, for a total of over 33 Billion USD over 10 years The number of new patients treated declines over time in all scenarios. For impact on the epidemic, better performance in case finding and health system organization will be needed than in the current best regimen scenario. Subsidy would improve the impact marginally. 21

22 Results: Base case of All Countries [price decrease = 3.4, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] of Individuals (Annual) in 000s of Individuals in 000s Current Best Future Best Low-to-Middle Income Regimen ($) CumUlative Low-to-Middle Income Regimen ($) By Free By Free By Market By Market By Subsidised By Subsidised By % By % 22

23 Summary: Lessons from Base Case Scenarios LIC markets will be small - additional price decreases, better diagnostic/health system performance and high levels of subsidy are needed to have a meaningful public health impact LMIC and UMIC markets will be massive In LMIC, third party coverage would boost demand (X2.0 to X3.0) In LMIC, impact on the epidemic will require both greater price decreases than projected, better access to diagnostics and health system organization; In UMIC saturation of demand, high coverage, and impact on the epidemic may be reached if diagnostics and health system organization can be improved, and subsidy will impact demand in a limited way. Comparison of "current" and "future" best regimen case scenarios suggests that simplification in drug regimens and access to diagnostics might boost LIC, LMIC and UMIC markets in the next 5 years (X 2.5 to X4.0) and increase demand at the 2023 horizon. 23

24 Sensitivity Analysis: Expand eligibility for treatment from F3-F4 to F0-F4 Main Lesson In all cases, demand would be very sensitive to an extension of current treatment eligibility guidelines. If it happened now, there would be, in L+MIC combined, by the end of 2023: 57 million people treated, compared to 36.5 million in the current best regimen case/free market baseline scenario 198 million people treated, compared to 62 million in the future best regimen case/50% subsidy baseline scenario 24

25 Results: Base case of Lower-Middle Income Countries (LMICs) [price decrease = 3.4, subsidy = 50%, Treatment Group = F0 to F4, ATP =40%] Sensitivity F0 to F4 of Individuals (Annual) in 000s Base Case F3 to F4 of Individuals (Annual) in 000s Current Best Future Best LMIC Regimen ($) LMIC Regimen ($) By By Free Free Market By By Market By Subsidised By Subsidised By % By % 25

26 Results: Base case of All Countries [price decrease = 3.4, subsidy = 50%, Treatment Group = F0 to F4,, ATP =40%] Sensitivity F0 to F4 of Individuals (Annual) in 000s Base Case F3 to F4 of Individuals (Annual) in 000s Current Best Future Best Low-to-Middle Income Regimen ($) Low-to-Middle Income Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 26

27 Sensitivity Analysis: Subsidy Rate 0% - 100% (25% Interval) Main Lesson Free of charge drugs for end-users at the point of delivery (100% subsidy) would be required to have a significant impact on the number of annual cases in need of treatment in LIC. High-level of third-party coverage ( 75% subsidy) would be needed to boost demand (X2.0 or more) in LMIC. This is shown in the current best regimen case scenarios in the next 3 slides. 27

28 Results: Sensitivity Analysis of Low Income Countries (LICs) [price decrease = 3.4, subsidy = 0% - 100%, Treatment Group = F3 to F4, ATP =40%] Current Best of Individuals (Annual) in 000s of Individuals in 000s Current Best Current Best By 2018 LIC By 2023 LIC Subsidy Expenditure Subsidy Regimen ($) ($) Regimen ($) 0% % % % % % % % % %

29 Results: Sensitivity Analysis of Lower-Middle Income Countries (LMICs) [price decrease = 3.4, subsidy = 0% - 100%, Treatment Group = F3 & F4] Current Best of Individuals (Annual) in 000s of Individuals in 000s Current Best Current Best By 2018 LMIC By 2023 LMIC Subsidy Subsidy Regimen ($) Regimen ($) 0% % % % % % % % % %

30 Sensitivity Analysis: Rate of Price Reduction 25% ( ) then 5.2% Main Lesson In LIC, an initial price shock would help create a significant market The value of the LIC market will reach almost 1billion US$ in the most pessimistic scenario (current best case/free market) compared to 400 Million over 10 years in the base case, without "price shock In LMIC, it will increase access to treatment by a factor of 1.5 to 2.0, while maintaining total value of market over the next ten years In UMIC, it will increase demand by 30% to 50% in the next five years (without affecting total $ value of market). It will reach similar levels of coverage at the 2023 horizon (with circa 15% to 25% decrease in total $ market value) Prices would still be well above the marginal costs at the 2023 horizon for DAA-only treatment 30

31 Results: Base case of Low Income Countries (LICs) [price decrease = 25% ( ) then 5.2%, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] Sensitivity 25% decrease ( ) then 5.2% price decrease of Individuals (Annual) in 000s Base Case 3.4% price decrease of Individuals (Annual) in 000s Current Best Future Best LIC Regimen ($) LIC Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 31

32 Results: Base case of Lower-Middle Income Countries (LMICs) [price decrease = 25% ( ) then 5.2%, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] Sensitivity 25% decrease ( ) then 5.2% price decrease of Individuals (Annual) in 000s Base Case 3.4% price decrease of Individuals (Annual) in 000s Current Best Future Best LMIC Regimen ($) LMIC Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 32

33 Results: Base case of Upper-Middle Income Countries (UMICs) [price decrease = 25% ( ) then 5.2%, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] Sensitivity 25% decrease ( ) then 5.2% price decrease of Individuals (Annual) in 000s Base Case 3.4% price decrease of Individuals (Annual) in 000s Current Best Future Best UMIC Regimen ($) UMIC Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 33

34 Results: Base case of All Countries [price decrease = 25% ( ) then 5.2%, subsidy = 50%, Treatment Group = F3 & F4, ATP =40%] Sensitivity 25% decrease ( ) then 5.2% price decrease of Individuals (Annual) in 000s Base Case 3.4% price decrease of Individuals (Annual) in 000s Current Best Future Best Low-to-Middle Income Regimen ($) Low-to-Middle Income Regimen ($) By By Free Market Free Market By By By Subsidised By Subsidised By % By % 34

35 Sensitivity Analysis: Ability-to-Pay 10% - 50% (10% Interval) Main Lesson Results are quite robust to economic uncertainties on the demand-side : they are only moderately sensitive to assumptions about elasticity of demand to price under households budgetary constraints (ability-to-pay). However, monitoring of patients actual preferences and behaviors will remain key for scaling up access to treatment. 35

36 Results: Base case of Lower-Middle Income Countries (LMICs) [price decrease = 3.4, Treatment Group = F3 & F4, ATP = 10% - 50%] Free Market Current Best of Individuals (Annual) in 000s of Individuals in 000s Current Best Current Best By 2018 LMIC By 2023 LMIC ATP Expenditure ATP ($) Regimen ($) Regimen ($) 10% % % % % % % % % %

37 Results: Base case of Lower-Middle Income Countries (LMICs) [price decrease = 3.4, Treatment Group = F3 & F4, ATP = 10% - 50%] Free Market Future Best of Individuals (Annual) in 000s of Individuals in 000s Future Best Future Best By 2018 LMIC By 2023 LMIC ATP Expenditure ATP ($) Regimen ($) Regimen ($) 10% % % % % % % % % %

38 Some limitations of the forecasting exercise Model does not take into account potential re-infections and is limited to initial treatment initiation. Model does not fully take into account the impact of treatment on rates of transmission (incidence remains exogenous). There is an intrinsic difficulty in predicting ex-ante human behaviours and market dynamics Heterogeneity and uncertainty on epidemiological data used? Some assumptions on price dynamics derived from the experience of HIV/AIDS drugs? 38

39 Some major lessons of the forecasting exercise There is a significant opportunity for a massive scale up of access to HCV treatment in health system absorption capacity will be required. in treatment eligibility from F3-F4 to F0-F4 would greatly increase demand. a significant in initial drug prices is needed for LIC and LMIC, and should be possible because it would increase the total value of the LIC and LMIC markets over a 10 year horizon. significant subsidy will be needed in LIC markets and some subsidy in LMIC to decrease the number of new cases needing treatment. The willingness to subsidize will be greater if the drugs are more affordable and would increase the value of the market. There is a rationale for adapting policies for HCV treatment access according to countries level of development. 39