Addressing health equity concerns in costeffectiveness

Transcription

1 Addressing health equity concerns in costeffectiveness analysis? Distributional Cost-Effectiveness Analysis Miqdad

2 Overview 1) Introduction 2) Distributional CEA 3) Case study (cancer screening UK) 4) Case study (rotavirus vaccine Ethiopia) 5) Conclusion Miqdad Asaria 2

3 1. Introduction

4 The WHO UHC Cube Miqdad Asaria 4

5 Equity is Normative Inequality to economists just means variation or differences Equity refers to a fair or socially just allocation Defining what we mean by fair requires us to make social value judgements Equity does not always imply equality Miqdad Asaria 5

6 Equality vs Equity Source: The Partnership for Southern Equity (PSE) Miqdad Asaria 6

7 Relative inequality Equality Measured How? Difference between 40 years and 50 years equivalent to difference between 80 years and 100 years Absolute inequality Difference between 40 years and 50 years equivalent to difference between 80 years and 90 years Miqdad Asaria 7

8 Horizontal & Vertical Equity Horizontal equity means the equal treatment of equals in relevant respects Vertical equity means the unequal treatment for those who are unequal in relevant respects Miqdad Asaria 8

9 2. Distributional CEA

10 The Economic Problem Resources are scarce Decision makers need to prioritise Cost-effectiveness analysis is about doing as much good as possible with fixed budget In this case maximise overall health benefits Miqdad Asaria 10

11 Cost-Effectiveness Analysis Less effective more costly Q Reject D Cost Reject More effective more costly? Accept Health Opportunity Cost Reject Less effective less costly? Accept Accept D Effectiveness More effective less costly R Miqdad Asaria 11

12 Cost-Effectiveness Analysis Cost of funding one health policy is the health we lose by not funding an alternative health policy CEA only focusses on maximising total health has nothing to say on the distribution of health Miqdad Asaria 12

13 Social Welfare Analysis Equity efficiency trade off Less equitable more efficient? Accept D Health Impact More equitable more efficient R Reject Accept D Equity Impact Reject Accept Less equitable less efficient Q Reject More equitable less efficient? Miqdad Asaria 13

14 A Primer in Distributive Justice Health of person 1 (disadvantaged; e.g. poor childhood circumstances) Rawlsian utilitarian social indifference curves Line as close to Equality equality as possible Cost-effectiveness: the point with the largest sum total health is efficient MaxiMin point Egalitarian point Health (not Pareto efficient) maximising point Starting point Possibility frontier Health of person 2 Miqdad Asaria 14

15 Equally distributed equivalent Lifetime Health Distribution 62 Most deprived Q2 Q3 Q4 Least deprived Average = 69 QALYs Rawlsian EDE Health Utilitarian EDE Health Most deprived Q2 Q3 Q4 Least deprived Plausible range of EDEs Most deprived Q2 Q3 Q4 Least deprived September 2018 Miqdad Asaria 15

16 Comparing health distributions Health Distribution A Inequality aversion Average = 70 QALYs Most deprived 62 Most deprived Q2 Q3 Q4 Least deprived Health Distribution B Q2 Q3 Q4 Least deprived Average = 71 QALYs social welfare function EDE A EDE B Choose policy with max EDE September 2018 Miqdad Asaria 16

17 Social Welfare Functions SWFs allow us to quantitatively evaluate this equity efficiency trade off They require parameterisation with an inequality aversion parameter to specify the curvature of the indifference curves to give something between the utilitarian (parameter=0) and Rawlsian (parameter= ) extremes Atkinson SWF (relative) Kolm SWF (absolute) Miqdad Asaria 17

18 Focus group exercises to elicit inequality aversion Miqdad Asaria 18

19 Inequality Aversion in England 60% 50% 49% 84% of people are willing to sacrifice some health for more equal distribution 40% 30% Traditional CEA 31% 20% 14% 10% 0% Pro-Rich 2% Health Maximiser Weighted Prioritarian 4% MaxiMin Egalitarian Miqdad Asaria 19

20 The Inequality Aversion Parameter SWF Median* (95% CI) Implied weight** (95% CI) Atkinson (ε) Kolm (α) ( ) ( ) ( ) ( ) * Median preference and confidence intervals identified through bootstrapping; population weights used * * Implied weight of marginal health gain to poorest fifth of the population compared to the marginal health gain to the richest fifth of the population Miqdad Asaria 20

21 3. Case Study NHS Bowel Cancer Screening Program

22 NHS Bowel Cancer Screening Programme Bowel cancer is the second most common cause of cancer death in the UK more than 16,000 in 2010 Free national screening programme rolled out in 2006 to all year olds in England Those who attend screening have a 25% reduction in their risk of dying from bowel cancer Less than 60% of those eligible for screening participate Miqdad Asaria 22

23 Inequality in Health Quality Adjusted Life Expectancy Most Deprived IMD 4 IMD 3 IMD 2 Least Deprived Miqdad Asaria 23

24 Inequality in Screening Uptake 75% 70% 69% 65% 65% gfobt Uptake (%) 60% 55% 55% 61% 50% 45% 45% 40% Most Deprived IMD 4 IMD 3 IMD 2 Least Deprived Miqdad Asaria 24

25 Redesign Options Two redesign options considered both having same total cost Option A: additional reminder sent by doctor targeted at the most deprived areas increase in uptake 12% Option B: a standard reminder sent to everybody increase in uptake 6% Miqdad Asaria 25

26 Impact of Redesign on Health Incremental Per Person QALYs targeted universal Most Deprived IMD 4 IMD 3 IMD 2 Least Deprived Miqdad Asaria 26

27 DCEA Results 1,500 1,000 Universal Better Universal EDE - Targeted EDE (Population QALYs) ,000-1,500-2,000 Targeted Better -2, Constant Relative Inequality Aversion (Atkinson ε) Miqdad Asaria 27

28 4. Case Study Rotavirus vaccination program Ethiopia

29 Rotavirus vaccine in Ethiopia Rotavirus responsible for a third of global diarrhoea- related deaths Prior to vaccine Ethiopia had fifth highest number of rotavirus related deaths (28,218 in 2013) Vaccine reduces severe rota-virus disease by > 60% in the first year of life Vaccine coverage is 56% (DHS 2016) Standard vaccine roll-out vs pro-poor vaccine roll-out targeted to rural populations There are higher delivery costs in rural areas these areas are typically inhabited by poorer citizens Miqdad Asaria 29

30 Health distribution Ethiopia 70 Health life expectancy at birth Q1 (poorest) Q2 Q3 Q4 Q5 (richest) Miqdad Asaria 30

31 Distribution of health opportunity cost Proportion of opportunity cost (%) 22% 21% 21% 20% 20% 19% 19% 18% 18% 17% 21% 21% 21% 19% 19% Q1 (poorest) Q2 Q3 Q4 Q5 (richest) Miqdad Asaria 31

32 Distribution of vaccine uptake Rotavirus vaccine coverage (%) Q1 (poorest) Q2 Q3 Q4 Q5 (richest) standard pro-poor Miqdad Asaria 32

33 Cost-effectiveness results Miqdad Asaria 33

34 DCEA Results Miqdad Asaria 34

35 Extended Cost-Effectiveness Analysis Looks at impacts on financial risk protection as well as health Essentially a benefits incidence analysis Does not include opportunity costs in either dimension Does not evaluate trade-offs Can be seen as a first step in DCEA Miqdad Asaria 35

36 5. Conclusion

37 DCEA DCEA requires distributions of: population health impacts of policies (benefit incidence) health opportunity cost DCEA does not look at non-health impacts of policies Miqdad Asaria 37

38 Conclusion Economics can help provide tools to think about and quantify health inequality Economics can help identify efficient policies to address inequalities and make trade-offs if and when necessary Social value judgements need to be made in order to make trade-offs, analysts are not the people who should be making these Miqdad Asaria 38

39 6. References

40 Key readings Asaria, M. Method of the month: Distributional cost effectiveness analysis. AHE Blog. Cookson, R., Mirelman, A.J., Griffin, S., Asaria, M., Dawkins, B., Norheim, O.F., Verguet, S. and Culyer, A.J., Using cost-effectiveness analysis to address health equity concerns. Value in Health, 20(2), pp Asaria, M., Griffin, S., Cookson, R., Whyte, S. and Tappenden, P., Distributional cost-effectiveness analysis of health care programmes a methodological case study of the UK bowel cancer screening programme. Health economics, 24(6), pp Love-Koh, J., Asaria, M., Cookson, R. and Griffin, S., The social distribution of health: estimating qualityadjusted life expectancy in England. Value in Health, 18(5), pp Robson, M., Asaria, M., Cookson, R., Tsuchiya, A. and Ali, S., Eliciting the level of health inequality aversion in England. Health economics, 26(10), pp Dawkins, B.R., Mirelman, A.J., Asaria, M., Johansson, K.A. and Cookson, R.A., Distributional cost-effectiveness analysis in low-and middle-income countries: illustrative example of rotavirus vaccination in Ethiopia. Health policy and planning, 33(3), pp Miqdad Asaria 40