Health Economics – SOCE3B11 – Autumn 04/05 Lecture 21: The Economics of AMR Richard Smith Reader in Health Economics School of Medicine, Health Policy.

Health Economics – SOCE3B11 – Autumn 04/05

Lecture 21: The Economics of AMR

Richard SmithReader in Health Economics

School of Medicine, Health Policy & Practice


Overview of lecture

What is antimicrobial resistance (AMR)?

Economic conceptualisation of AMR Cost of resistance - country, hospital,

disease Micro-economic evaluation of

strategies to contain AMR Macro-economic evaluation of impact

of AMR and strategies to contain AMR


Antimicrobial Resistance (AMR)

AMR occurs where a micro-organism previously sensitive to an antimicrobial therapy develops resistance to its effect, rendering it ineffective

It is associated with antimicrobial usage (over & under use) and the interaction of micro-organisms, people and the environment

It is potentially irreversible once developed: some resistances are linked (therefore reduction in

all associated antimicrobials is necessary) the resistance mechanism/gene encoding may

provide an unrelated selective advantage to the organism

the 'genetic cost' to the organism of maintaining AMR in the absence of selection pressure may be minimal


Importance of AMR

“Despite the multifactorial nature of antibiotic resistance the central issue remains quite simple: the more you use it, the faster you lose it” (The Lancet, 15/4/95)

“We may look back at the antibiotic era as just a passing phase in the history of medicine, an era when a great natural resource was squandered, and the bugs proved smarter than the scientists” (Cannon, 1995)

“We are further away from mastering infectious diseases than we were 25 years ago” The Times, 4/4/95


Importance of AMR


Economic Conceptualisation of AMR

Externality = Effect on those other than the immediate consumer (cross-sectional & temporal ext.)

Resistance = Negative externality (i.e. cost) associated with consumption

of antimicrobials now

Implication = Sub-optimal (over) consumption of antimicrobials


A Difficult Balance

The best interests of the individual

Society’s need for sustainable

antibacterial use


Form of Negative Externality

ERt = f(At, Xi

t)

ERt = extent of externality (AMR) in time t

At = quantity of AMs consumed in time t

Xit = vector of exogenous factors


Form of Positive Externality

EPt = f(At, ER

t, Xit,)

EPt = externality associated with reduced

transmission of disease during time t

At = quantity of AMs used in time t

ERt = extent of externality (AMR) in time t

Xit = vector of exogenous factors


Optimisation of AM Use

NBAt = f(Bt, Ct, St, Dt, EP

t, ERt, At, Xi

t)

NBAt = net benefit from AMs used in time t

Bt = direct benefit to patient of AMCt = drug (+ administration) costSt = cost associated with side-effectsDt = represents difficulties in diagnosis(EP

t, ERt, At, Xi

t as before*)*EP

t = externality associated with reduced transmission during time t; At = quantity of AMs used in time t; ER

t = extent of externality (AMR) in time t; Xit = vector of exogenous factors


Implications Of AMR As Externality

NOT eradication, but containment of AMR

Importance of optimisation over time - use (and benefit from) AMs now and in future

Importance of assessing costs and benefits of AM use and strategies to contain AMR


Cost of AMR

Additional investigations Additional treatments Longer hospital stay Longer time off work Reduced quality of life Greater likelihood of death Impact on wider society (health and

economic)


Cost of AMR

By country (e.g. USA) $4-7bn pa to medical care sector (American

Soc. for Microbiology, 1995; John & Fishman, 1997)

By institution (e.g. hospital) ~£500,000 to contain 5 week outbreak of MRSA

in general hospital (Cox et al, 1995) By disease (e.g. Tuberculosis)

Double cost of standard treatment ($13,000-$30,000) (Wilton et al, 2001)


Micro-economic Evaluation Of Strategies To Contain AMR

Systematic review of strategies (GFHR/ WHO)

Specific economic policies (WHO, CMH, UNDP, CIDA/Health Canada, US NAS)

Development of WHO ‘Global Strategy’


Strategies To Contain AMR

FOCUS OF STRATEGYLEVEL OFSTRATEGY Reduce transmission Prevent emergence

Micro E.g. handwashing inhospitals

E.g. ‘cycling’ drugswithin hospitals

Macro E.g. restrictinginternational travel

E.g. restriction policies(eg taxation, permits)



Objective Strategy Intervention Selection Pressure* Antimicrobial use in

humans and agriculture

Education of professionals Education of patients Rapid diagnosis of bacterials Control of sensitivity data released

to prescribers Antibiotic policies Restriction of availability Financial incentives/disincentives Antimicrobial cycling Regulation on the use of

antibiotics in agriculture Opportunity forResistance Emerging*

Optimal use of existingagents

Ensuring optimal agent, dose anddose frequency for each infection

Removal of potential septic foci Emphasising/ensuring compliance Use of antibiotic combinations

Range of AgentsAvailable

Consider use ofalternative treatmentoptions

Antiseptics Cranberry juice for UTI Probiotics

Requirement forAntimicrobials

Immune competence Vaccination Nutrition Minimise time patient is

immunocompromised



Objective Strategy Intervention Transmission Early recognition of

resistant organisms• More rapid techniques• Surveillance• Screening patients/staff

Infectivity Use of antimicrobials OPPORTUNITIESFOR TRANSMISSION

Isolation Handwashing General Hygiene Patient/Staff ratios Bed spacing

Susceptibility toinfection

Immunity Nutrition

3. DEVELOP NEW ANTIMICROBIALS Range of AgentsAvailable

Discover/develop newagents

1. Modification of existing agents/discovery of new antimicrobials

2. Discovery of new drug targetsthrough microbial gene analysis

4. Genetic manipulation5. Computer modelling


Evidence: Literature Review

127 studies of strategies to contain AMR. Most are:

of poor methodological quality (high risk of bias)

from developed nations (principally the USA) not measuring the cost impact of AMR micro (institution) not macro (community) concerned with transmission not emergence


Importance Of Transmission Versus Emergence

Equilibrium resistance

X X+n

% of organism resistant to an anti-microbial

Time

Lag phase

0


Importance of Time

Because of uncertainty, evaluation of strategies to reduce transmission easier to undertake than evaluation of strategies to control emergence

Because of discounting of future benefits, strategies to reduce transmission likely to appear to be more cost-effective than strategies to control emergence


The Problem

Micro policies – generally to contain transmission – are more likely to be rigorously evaluated ...

BUT ... macro policies – generally to contain emergence – are more likely to be socially optimal (and) in the long-term.


Macro-economic Impact Of AMR

As Lecture 18, national ‘health’ and ‘wealth’ are closely linked

BUT: micro-economic evaluation typically concentrate on impact only to health care sector (+ patient/family)

Such ‘micro’ analyses may result in loss of social efficiency, especially so when evaluating ‘macro’ interventions

An alternative is macro analysis – CGE is most ‘popular’ (e.g. energy, environment, taxation, economic development, international trade, agriculture, trade reform etc).


Overview of CGE Approach

CGE is general equilibrium model solved to find prices at which quantity supplied equals quantity demanded across all markets (sectors)

Describes economy using representative agents: consumers, producers, and government

Consumers allocate time to employment/leisure and income to consumption/saving to max utility

Producers combine labour/capital inputs to max profit

Government collects tax revenue to finance expenditure & redistribute as benefits


The Modelling Process

Microconsistent benchmark data set for single year (SAM)

Exogenous elasticities

Calibration(model spec.)

Replication check

Calculation of benchmark

equil.

Policy evaluation - pairwise comparison between counterfactual & benchmark equil.


The Social Accounting Matrix (SAM)

Organises data from diverse sources into consistent /comprehensive database of revenue & expenditure for every commodity/agent. E.g....

Goods Factors Firms Consumers Column Total

Goods 80 200 280 Factors 200 200 Firms 280 280 Consumers 200 200

Row Total 280 200 280 200 Note: Rows=receipts from column accounts, columns=expenditure on row accounts. E.g. firms ‘spend’ (expenditure) on goods & factors, and goods receive ‘income’ (receipts) from firms and consumers who buy them


The basic AMR-CGE model

The basic model represents a simple static, closed-economy model, with 10 sectors of the UK economy using SAM from 1995 (last IO data)

Modifies the 8-sector version of 123 sector IO table to disaggregate health sector into 3 separate sectors based on OECD accounts and NHS expenditure allocation for 1995


Data – SAM for UK (1995 £million)

AGR PHA CHE R&

D EDU SOW AD

M HOS FHS OTH CAPZ LABZ CGZ HH INVZ TOTAL

AGR 2110 52 27 4 65 1 2 7 3 14243 15 3942 231 20702 PHA 6 168 24 0 0 2 0 0 0 5857 5 421 23 6506 CHE 562 0 0 0 54 31 1 3 1 15784 43 596 191 17266 RD 35 48 57 66 193 4 29 104 47 3109 354 11 50 4107 EDU 16 28 50 110 507 33 15 54 24 21480 72 6319 7 28715 SOW 0 0 0 0 1 288 7 26 12 7410 2 3430 9 11185 ADM 16 2 4 2 5 7 22 82 37 3001 47 452 2 3679 HOS 60 9 15 9 18 24 82 297 134 10662 39 1648 44 13041 FHS 27 4 7 4 8 11 37 134 60 4985 14 743 6 6040 OTH 6426 2969 7872 1065 2650 1184 621 1905 1020 497819 66 360788 205568 1089953 GOV 125137 125137 HH 240236 383723 85659 709618 SAV 206131 206131 CAPZ 8488 1144 3408 895 2064 1096 720 2621 1181 213759 4860 240236 LABZ 2956 2082 5802 1952 23150 8504 2143 7808 3521 291844 33961 383723 TOTAL 20702 6506 17266 4107 28715 11185 3679 13041 6040 1089953 240236 383723 125137 709618 206131 AGR – agriculture, PHA – pharmaceuticals, CHE – chemicals, R&D - research and development, EDU – education, SOW - social works activities, ADM - health administration, HOS -hospital services, FHS - family health services, OTH - All other industries aggregate, CAPZ – baseline capital endowment, LABZ– baseline labour endowment, HH- final consumer consumption demand, INVZ-baseline investment demand, GOV- government consumption demand, SAV- aggregate domestic savings






Replication check


equil.



Model Specification

Dimensions of the model – number/type of sectors/industries, whether analysis is static/dynamic, whether model is closed or open economy, etc

Types of functional forms to describe agents behaviour – normally constant elasticity of substitution (CES) function, usually Cobb-Douglas (CD)






Replication check


equil.



Exogenous Elasticities

For a CD function, a single price and quantity observation is sufficient to determine the parameters of the function uniquely

But, for more general CES functions extra values of substitution elasticity parameters are required to infer the curvature of indifference curves and isoquants. These are taken as exogenous inputs to the model


Elasticity Parameters In Model

Sectors Own price Elasticities of Consumer Demand

functions (LES estimates) Agriculture 0.334 Pharmaceuticals 0.685 Chemicals 0.685 Research and Development 0.578 Education 0.488 Social Services 0.488 Health Administration 0.488 Hospitals 0.488 Family Health Services 0.488 Other Sectors 0.488






Replication check


equil.



Calculating Benchmark Equilibrium

Analysis requires calculating a ‘benchmark’ equilibrium year (from data for a single year, or average from a number of years)

As benchmark uses values from the SAM, the model can reproduce the SAM as an equilibrium solution (consistency check)

Once calibration is complete, a fully specified numerical model is available for policy analysis


Solving The Model

CGE models rely on well established solution algorithms available as software packages

The Generalised Algebraic Modelling System (GAMS) is popular

GAMS is a set of syntax which declare sets, data, parameters, variables and equations, and assigns model relationships






Replication check


equil.

Policy evaluation - pairwise comparison between benchmark equil. & counterfactual


Policy Evaluation

Policy analysis in CGE models is ‘comparative static’ – what would happen if some policy is introduced as exogenous ‘shock’ to the model?

The ‘post-change’ counterfactual equilibrium is compared with the benchmark equilibrium

Policy is evaluated in terms of macro-economic indicators – national income, employment, inflation, welfare criteria [compensating (CV) and equivalent (EV) variations)]


Macroeconomic Impact Of AMR

AMR is a (negative) exogenous shock on the labour supply and productivity of inputs, and a (positive) shock (cost) to healthcare delivery

No UK data of impact on productivity or labour supply so use data from other areas/countries

Assumptions: Prevalence of AMR ~40% in UK AMR reduces labour supply by 0.1% to 0.8% AMR reduces productivity by 0.5% to 10% AMR increases healthcare cost by 0.5% to 10%


Macroeconomic Impact of AMR in UK

Change in: Different Scenarios Productivity (%) -0.5 -0.5 -1.0 -1.0 -0.5 -1.5 -1.5 -2.0 -2.0 Healthcare delivery cost (%) +0.5 +0.5 +1.0 +1.0 +0.5 +10 +5.0 +5.0 +10 Labour supply (%) -0.1 -0.8 -0.2 -0.8 -0.5 -0.2 -0.1 -0.8 -0.8 Impacts on macroeconomic indicators in percentage Household income -0.070 -0.098 -0.140 -0.174 -0.085 -0.234 -0.257 -0.327 -0.327 Government transfers +0.740 +0.721 +1.690 +1.407 +0.753 +2.092 +2.119 +2.802 +2.817 Tax Revenues -0.081 -0.110 -0.160 -0.196 -0.086 -0.216 -0.269 -0.352 -0.324 Unemployment +4.210 +4.919 +9.630 +9.159 +4.644 +12.98 +13.66 +17.71 +17.80 Household utility -0.590 -0.612 -1.200 -1.211 -0.600 -1.816 -1.821 -2.413 -2.425 Real GDP -0.400 -0.381 -0.810 -0.776 -0.387 -1.199 -1.174 -1.574 -1.582 Welfare (EV/GDP) -0.270 -0.282 -0.560 -0.559 -0.276 -0.834 -0.841 -1.121 -1.121 Inflation (CPI index) +0.004 +0.004 +0.010 +0.008 +0.004 +0.013 +0.013 +0.017 +0.017 Total Savings -0.500 -0.531 -1.000 -1.017 -0.525 -1.449 -1.512 -1.990 -1.980 Healthcare and social -0.566 -0.585 -1.131 -1.156 -0.575 -2.621 -2.105 -2.613 -3.139 Services (average) Social services -0.537 -0.558 -1.074 -1.100 -0.547 -1.803 -1.717 -2.242 -2.344 Health administration -0.575 -0.594 -1.150 -1.174 -0.584 -2.894 -2.234 -2.736 -3.404 Hospitals -0.576 -0.594 -1.150 -1.175 -0.584 -2.894 -2.235 -2.737 -3.405 Family health services -0.575 -0.594 -1.150 -1.174 -0.584 -2.892 -2.234 -2.735 -3.403


Summary Results

GDP loss = ~£3-11 billion (~ 6-20% of total NHS expenditures) Welfare losses imply society willing to pay ~ £8 billion to avoid

AMR

Parameter of interest Impact of MRSA (% change) Household Income -0.070 to -0.327 Government Transfers +0.721 to +2.817 Tax Revenues -0.081 to -0.352 Unemployment +4.210 to +17.800 Household Utility -0.590 to -2.425 GDP (real) -0.381 to -1.582 Welfare(EV/GDP) -0.270 to -1.121 Inflation(CPI) +0.004 to +0.017 Total National Savings -0.500 to -1.990 Health and Social Services Social Services -0.537 to -2.344 Health Administration -0.575 to -3.404 Hospitals -0.576 to 3.405 Family Health Services -0.575 to -3.403

Note: All results are relative to 1995 economy in the absence of MRSA.


Macro-Economic Strategies To Contain AMR

Charges/taxes (equal to marginal external cost of AMR) – likely most efficient, but difficult to administer/inequitable

Regulation – easy to introduce, less efficient than more direct ‘economic’ incentives

Tradable permits (licences) - set quantity and let price adjust in market through physician ‘trading’


Evaluation Of ‘Macro’ Strategies

User-charge/tax increase cost of AM use - 40% tax leads to reduction of 10% in use of AMs (based on price elasticities observed)

Regulation yields 10% decrease in prescription of antimicrobials, but inflexible in who is targeted

Permits are combination policy – yield 10% decrease in AM consumption overall, but flexibility in who consumes them


Key Results

Impact of Impact of Impact of

Parameter of interest Regulation Taxation Permits Household Income +0.014 +0.005 +0.015 Government Transfers -0.150 -0.048 -0.151 Tax Revenues +0.017 +0.005 +0.017 Unemployment -0.862 -0.278 -0.870 Household Utility +0.119 +0.039 +0.120 GDP (real) +0.078 +0.025 +0.079 Welfare(EV/GDP) +0.055 +0.018 +0.055 Inflation(CPI) +0.000 +0.000 +0.000 Total Savings +0.101 +0.033 +0.102 Health and Social Services Social Services +0.108 +0.035 +0.109 Health Administration +0.115 +0.037 +0.116 Hospitals +0.115 +0.037 +0.116 Family Health Services +0.115 +0.037 +0.116 CAM and MRSA CAM (%) -10.00 -0.968 -10.00 MRSA Level (%) -10.00 -0.968 -10.00

Note: All results are relative to 1995 model that includes the adverse impacts of MRSA on the economy in the absence of any intervention.


Key Conclusions of Macro Approach

AMR substantially affects wider economy, not just healthcare

Concentrating on healthcare sector alone may therefore underestimate the societal impact of AMR/strategies

Of ‘macro’ strategies, taxation appears to be the least efficient & tradable permits the most efficient


Extensions To Basic Model

Capture international transactions, given the potential for global transmission of resistance

Include temporal dynamics to account for structural adjustment to strategies

Disaggregate consumer sector to understand better the distribution of the costs of AMR


Conclusions – Applying Economics To The Analysis of AMR

Conceptualisation of problem: Optimisation and balance Importance of temporal factors (trade-off now vs

future) Technical analysis:

Micro-economic evaluation of strategies Macro-economic assessment

Strategies: Financial incentive structures (e.g. permits) Tackling ‘public good’ issues globally (lecture 20)


Further references

Externality & micro-economic evaluation: Coast J, Smith RD, Miller MR. Superbugs: should antimicrobial resistance be

included as a cost in economic evaluation? Health Economics, 1996; 5: 217-226.

Coast J, Smith RD, Karcher AM, Wilton P, Millar MR. Superbugs II: How should economic evaluation be conducted for interventions which aim to reduce antimicrobial resistance? Health Economics, 2002; 11(7): 637-647.

Wilton P, Smith RD, Coast J, Millar MR. Strategies to contain the emergence of antimicrobial resistance: a systematic review of effectiveness and cost-effectiveness. Journal of Health Services Research and Policy, 2002; 7(2): 111-117.

Macro policies & macro-economic analysis: Coast J, Smith RD, Millar MR. An economic perspective on policy for

antimicrobial resistance. Social Science and Medicine, 1998; 46: 29-38. Smith RD, Coast J. Controlling antimicrobial resistance: a proposed

transferable permit market. Health Policy, 1998; 43: 219-32. Smith RD, Coast J. Antimicrobial resistance: a global response. Bulletin of the

World Health Organisation, 2002; 80: 126-133. Smith RD, Coast J. Resisting resistance: thinking strategically about

antimicrobial resistance. Georgetown Journal of International Affairs, 2003; IV(1): 135-141.

Yago M, Smith RD, Coast J, Millar MR. Assessing the macroeconomic impact of a healthcare problem: the application of computable general equilibrium analysis to antimicrobial resistance. Journal of Health Economics (in press).

Health Economics – SOCE3B11 – Autumn 04/05 Lecture 21: The Economics of AMR Richard Smith Reader in Health Economics School of Medicine, Health Policy.

Documents

time t b t

time t e r t

externality n

importance of amr n

n resistance

time t x

work n

n implication