This presentation was made by Peter Smith, United Kingdom, at the 6th Meeting of the Joint OECD DELSA-GOV Network on Fiscal Sustainability of Health Systems, held at the OECD Conference Centre, Paris, on 18-19 September 2017
4. Do health services create value
• Yes!
• Estimates from England suggest that – at the
margin – USD 15,000 saves an extra life year.
• In low income countries the figure may be closer
to USD 150.
• Many interventions are much more cost-effective
even than this.
• But there are huge variations in treatments and
spending suggesting scope for vastly improved
productivity.
5. Why is low productivity a concern?
• It may deny health gain or create dissatisfaction for patients who
have received treatment, because they do not receive the best
possible care available within the health system’s resource limits;
• By consuming excess resources, low productivity may deny health
gain to other patients who could have benefited from treatment if
the resources had been better used;
• Poor productivity in the health sector may sacrifice consumption
opportunities elsewhere in the economy, such as education or
nutrition;
• Wasting resources may reduce society’s willingness to contribute to
the funding of health services, thereby harming social solidarity,
health system performance and social welfare;
• Low productivity threatened the sustainability of the health system.
6. Productivity: some preliminary issues
• Productivity: how well people and organizations convert inputs into
outputs.
• What is the unit of analysis?
– Clinician, Hospital, Health System etc?
• What are inputs?
– Just physical inputs, or social circumstances, other services etc?
• What are the outputs?
– Just physical outputs, or health gain? Other social outputs?
• Need to distinguish between productivity improvement and expenditure
control
– Improved productivity can arise from higher levels of attainment at the same cost,
as well as lower expenditure for the same attainment.
– Is the improved productivity in the right sector of the health system? Or is it doing
the ‘wrong things’ better?
• Often a need to adjust for uncontrollable constraints on better attainment
– E.g. diet; smoking; geography
7. Productivity: a health warning
• It is often the case that there are highly productive entities within
highly inefficient health systems.
• For example, hospitals may be highly productive (many patients,
low lengths of stay) but many of the treatments they offer may be
unnecessary if they were treated in a timely fashion in primary care.
• Allocative inefficiency
– Misallocation of inputs or outputs
– E.g. poor use of trained clinical skills; production of treatments with
low benefits relative to costs
• Technical inefficiency
– ‘Leakage’ of resources during the production process; waste
– E.g. excess prices for inputs; unnecessary duplication of tests
• Whilst technical inefficiency is closely related to productivity,
allocative inefficiency may be a more serious challenge in many
health systems.
9. Waste in the Australian health system
• Runciman et al. (2012) reviewed over 1000 Australian adults and their health care encounters and
found that 43 per cent received inappropriate care, based on evidence-based and consensus-based
guidelines.
• A 2007 study by the Commonwealth Fund found that 15 per cent of Australians reported
undergoing unnecessary repeat imaging. This has been associated with ‘treatment cascades’ —
that is, subsequent procedures that may be of low value to patients, or even unnecessary (Russell and
Doggett 2015).
• The former National Institute of Clinical Studies (2003, 2005) identified gaps between evidence
and practice in areas such as advising on smoking cessation, screening for lung cancer, and
vaccinating against influenza.
• In 2013-14, about 30 per cent of people presenting to general practitioners in Australia for
acute upper respiratory tract infection — the ‘common cold’ — were prescribed antibiotics,
even though antibiotics are ineffective for treating viral infections (SCRGSP 2015).
• Paracetamol is commonly recommended and prescribed for back pain in Australia. However, a
recent randomised trial of paracetamol for the treatment of acute lower back pain found no benefit
versus a placebo (Carpenter et al. 2014).
• Approximately 6.5 per cent of separations in public hospitals in 2012-13 were associated with
‘adverse events’ — where patients are harmed during hospitalisation — including injuries from
falls, adverse drug effects and surgical errors (SCRGSP 2015).
10. Areas of waste in health care (Adapted from
Berwick & Hackbarth, 2012)
Area Why is it wasteful? (and how to measure it)
Failures of care delivery Poor execution or lack of widespread adoption of best practices, such as effective
preventive care practices or patient safety best practices. Can result in patient injuries,
worse clinical outcomes, and higher costs.
Example indicators: patient injuries, avoidable mortality, delay in/lack of implementation
of best practices, unused drugs
Failures of care coordination Occur when patients experience care that is fragmented and disjointed--for example, when
the care of patients transitioning from one care setting to another is poorly managed.
Example indicators: unnecessary hospital readmissions; preventable emergency
admissions
Overtreatment Includes outdated modes of care, care driven by providers' preferences rather than those
of informed patients, care that ignores scientific findings, or that is motivated by something
other than provision of optimal care for a patient.
Example indicators: excessive use of antibiotics, use of brand-name drugs where a
generic substitute is available, unnecessary procedures, duplicate medical tests
Administrative complexity Occurs because private health insurance companies, the government, or accreditation
agencies create inefficient or flawed rules and unduly bureaucratic procedures.
Example indicators: high expenditures on administration
Price failures The price of a service exceeds that found in a properly functioning market, arising for
example from poor procurement.
Example indicators: Variations in unit prices paid for goods and services.
Fraud and abuse In addition to fake medical bills and scams, this category includes the cost of additional
inspections and regulations to catch wrongdoing.
Example indicators: Extent of qualified audits
11. OECD (2017), Tackling Wasteful
Spending on Health.
• Adverse events in 1/10 hospitalization, add between 13
and 17% to hospital costs and up to 70% could be
avoided
• Geographic variations in rates of cardiac procedures (x3)
and knee replacements (x5) are largely unwarranted
• Up to 50% of antimicrobial prescriptions are unnecessary
• 12% to 56% of emergency department visits
inappropriate
• Share of generic medicines varies between 10% and 80%
• Administrative expenditure varies more than seven-fold
• Loss to fraud and error averages 6% of payments
16. Barriers to productivity improvement may occur
at any stage of the transformation process
Partial views
•Are inputs purchased at minimum cost (sometimes
referred to as ‘economy’)?
•Is the correct mix of inputs being used?
•Is there unnecessary creation of physical outputs (tests,
visits, etc)?
•Is an appropriate level of patient quality being secured?
•Are there adverse health service consequences of the
treatment (eg emergency readmission)
Total view
•Is the episode of care produced cost-effectively?
17. Carter, P. (2016) Operational productivity and performance in
English NHS acute hospitals: Unwarranted variations
20. Unit costs as the backbone of most
productivity analysis
• Costs of producing some physical unit of health
services output
• Seriously underused
• Highly dependent on uniformity of accounting choices
• Will patient level information costing systems (PLICs)
help?
• Nothing about:
– Quality of output
– Impact on the broader system (eg future readmissions to
hospital)
• Can act as a broader input into composite measures of
comparison (diagnosis related groups)
21. HealthBasket Objectives
• To identify and develop a methodology for cost comparison
• To assess whether prices are a good estimate of costs of
individual services
• To explore the reasons underlying variations in the costs of
individual services
Busse, R., Schreyögg, J. and Smith, P. (eds) (2008), “Special issue: Variability in health care treatment
costs amongst nine EU countries”, Health Economics, 17(S).
22. Selection of Case-Vignettes
Need for care Age Type of Care ECHI*
Appendectomy 14-25 M In-patient Surgery Emergency -
Normal delivery 25-34 F In-patient Obstetrics Elective +
Hip-replacement 65-75 F In-patient Surgery Elective +
Cataract 70-75 M Day case Surgery Elective +
Stroke 60-70 F In-patient Medical Emergency +
AMI (PTCA) 50-60 M In-patient Medical Emergency +
Cough ~ 2 M Out-patient Paediatrics/GP Emergency -
Colonoscopy 55-70 M Out-patient Diagnostic Elective +
Tooth filling ~ 12 Out-patient Dental Emergency +
Physiotherapy (knee) 25-35 M Out-patient Rehabilitative - -
23. The HealthBasket Project:
Vignette 5: Stroke
• So far healthy female (i.e. no co-morbidity), 60-70 years old, with
sudden severe hemiparesis (right side) and dependency, with severe
aphasia:
• Admission to hospital (accident & emergency, medical or
neurological department depending on country/ hospital) by
ambulance car.
• Start of case vignette: hospital door. All the interventions including
diagnostic and treatment are delivered in the same hospital. The
patient is diagnosed and treated according to normal hospital
standards (which may or may not include a stroke unit, early
rehabilitation etc.); progress is average for age. Transient (TIA),
short and reversible (RIND) and prolonged and reversible (PRIND)
ischaemic neurological deficits are excluded.
• End of vignette: discharge to rehabilitative institution or home.
Busse, R., Schreyögg, J. and Smith, P. (eds) (2008), “Special issue:
Variability in health care treatment costs amongst nine EU countries”, Health Economics, 17(S).
24. Breakdown of stroke costs (€)
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
Denmark England France Germany Hungary Italy Netherlands Poland Spain
Diagnostic procedures
Stroke unit (staff time)
Ward (staff time)
Drugs
Overheads
25. Stroke: Labour contact hours
Physician
hours per
patient
Nurse
hours per
patient
Denmark 6.1 25.5
England 14.4 23.6
France 10.4 11.0
Italy 5.8 10.8
Netherlands 5.0 16.3
Poland 19.5 64.1
Spain 11.1 34.2
26. Diagnosis Related Groups (DRGs)
• Developed by Bob Fetter, Yale University
• Original intention was to act as a mechanism for
comparing costs or outcomes of hospitals
• Used to create an index of actual over expected
costs, adjusting for case mix
• Main challenge was to collapse thousands of
diagnosis codes and procedures into a manageable
number of clinically meaningful DRGs
• Intention was to minimize heterogeneity within
groups
Fetter, R. (1991). "Diagnosis related groups: understanding hospital performance." Interfaces 21(1): 6-26.
27. National average reference costs, selected non-elective inpatient
healthcare resource groups, 2005/06, England
Code HRG Label Count Average
Unit Cost
Lower
Quartile
Upper
Quartile
£ £ £
E02 Heart Transplant
75 32,113 7,895 47,437
E07 Pacemaker Implant for , Heart Failure or Shock
792 4,336 1,572 4,995
E08 Pacemaker Implant except for , Heart Failure or
Shock 9,575 3,605 1,540 4,068
E08DF Pacemaker Implant except for , Heart Failure or
Shock - Defibrillator Implant & Explant only 977 16,725 13,606 20,737
E11 Acute Myocardial Infarction with complications
23,219 1,695 1,097 2,401
E12 Acute Myocardial Infarction without complications
63,475 1,169 775 1,718
E15 Percutaneous Coronary Intervention
24,378 3,401 1,109 3,641
E15DF Percutaneous Coronary Intervention - Defibrillator
Implant & Explant only 81 15,906 14,747 20,214
E18 Heart Failure or Shock age >69 or with
complications 52,618 1,694 1,208 2,560
E19 Heart Failure or Shock age <70 and without
complications 10,009 1,390 854 1,963
E35 Chest Pain age >69 or with complications
59,057 603 504 1,123
E36 Chest Pain age <70 and without complications
95,136 458 409 848
E99 Complex Elderly with a Cardiac Primary Diagnosis
45,347 2,088 1,419 3,018
28. Extract from English NHS reference
costs 2013 to 2014
Adjusted Unadjusted
Org Code Organisation Name
Market Forces
Factor
Org-Wide
Index1 Elective / DC
Non-Elective
Inpatient
Org-Wide
Index1 Elective / DC
Non-Elective
Inpatient
RY8
Derbyshire Community Health Services
NHS Trust
0.9506 84 88 63 80 84 61
RP5
Doncaster and Bassetlaw Hospitals NHS
Foundation Trust
0.9541 91 82 95 87 79 91
RBD
Dorset County Hospital NHS Foundation
Trust
0.9647 77 92 70 74 89 68
RDY
Dorset Healthcare University NHS
Foundation Trust
0.9814 185 154 203 184 153 201
RC3
Ealing Hospital NHS Trust 1.1045 103 114 99 114 126 110
RWH
East and North Hertfordshire NHS Trust 1.0504 93 105 89 98 111 93
RJN
East Cheshire NHS Trust 0.9665 93 79 99 90 77 96
RVV
East Kent Hospitals University NHS
Foundation Trust
0.9630 101 94 107 98 91 103
RXR
East Lancashire Hospitals NHS Trust 0.9510 107 89 118 102 85 113
RXC
East Sussex Healthcare NHS Trust 0.9617 111 111 110 107 107 107
RVR
Epsom and St Helier University Hospitals
NHS Trust
1.0981 98 106 93 108 116 102
RDU
Frimley Health NHS Foundation Trust 1.0700 89 73 99 96 78 106
29. EuroDRG Project
• Investigates patient level data of
10 episodes of care (representing
different medical specialties and
diagnostic/ therapeutic
procedures) across 12 countries
(1) identifying ways to calculate comparable
payments in an adequate fashion
(2) examining hospital efficiency within and
across European countries,
(3) identifying factors that affect the
relationship between the costs and quality
of inpatient care.
Busse R, Geissler A, Mason A, Or Z, Scheller-Kreinsen D,
Street A (2012)
Diagnosis-Related Groups in Europe (EuroDRG): Do
they explain variation in hospital costs and length of
stay across patients and hospitals?
Health Economics, Volume 21 (Supplement 2)
30. Productivity trends, UK National Health Service,
Office for National Statistics
• Follows recommendations of Sir Tony Atkinson
(2005), Measurement of Government Output and
Productivity for the National Accounts
• Outputs: about 2,500 types of activity
– Hospital, ambulatory, community, primary care,
prescribing
– Activities weighted according to cost (not value)
– Some (only partial) efforts to adjust for quality of care (eg
post-operative mortality)
• Continuing efforts to refine methods, in collaboration
with OECD
Source: Office for National Statistics (2008), Public Service Productivity: health care, London: ONS
31. Public service healthcare output, inputs and productivity indices
and growth rates, 1995 to 2013
Source: ONS (2015), Public Service Productivity Estimates: Healthcare 2013
33. Efficiency as a ‘residual’
Expenditure
Health
X
X
X
X
X
X
X
X
Production frontier
34. Spending and life expectancy
Denmark
Japan
Korea
Mexico
Norway
Spain
Switzerland
Turkey
US
70
72
74
76
78
80
82
84
0 1000 2000 3000 4000 5000 6000
Years
US $ PPP 2006
Source: OECD Health Data 2010
35. World Health Report 2000
• The health system:
“… all the activities whose
primary purpose is to
promote, restore or
maintain health.”
37. OECD Rankings
• Production function of life expectancy
• Uses panel data from OECD countries
• Determinants of life expectancy
– Health care spending
– Education
– GDP
– Pollution
– Alcohol
– Tobacco
– Diet
• Residual is health system efficiency
Joumard, I., C. Andre, C. Nicq and O. Chatal (2008) Health status determinants: lifestyle, environment, health care
resources and efficiency. Economics Department WorkingPaper 627. Paris: OECD.
38. Country-specific effects (life years)
Joumard et al (2008)
-5 -4 -3 -2 -1 0 1 2 3
United States
Hungary
Norway
Denmark
Turkey
Germany
Austria
Switzerland
Netherlands
Czech Republic
United Kingdom
Ireland
Belgium
France
Sweden
Poland
Finland
Canada
Greece
Korea
New Zealand
Australia
Iceland
40. The MACELI (Macro Cost Effectiveness
corrected for Lifestyle) project
• Baseline analyses without standardizing for lifestyle showed
on average more health spending was associated with better
health. This effect was clearest for countries with lower levels
of spending. Standardization towards a better lifestyle meant
an upward shift of the health production function, but did not
much alter the comparative efficiency of countries.
• The study covered the EU-28 Member States, Iceland, and
Norway. Individual-level data were used to describe lifestyle
across age and gender and to analyse its impact on health
outcomes and health care use.
Pieter van Baal et al (2015), Comparative efficiency of health systems, corrected for
selected lifestyle factors, Brussels: European Union.
43. The system level
– Information
• Prospective: cost-effectiveness guidance
• Retrospective: performance measurement
• Public reporting, benchmarking
– Incentives
• Payment mechanisms (P4P?)
• Markets and competition
• Targets
– Capacity
• Improvement agencies
• Audit and inspection
44. The organizational level
– Information
• Service line reporting
• Benchmarking data
• Electronic patient records
– Incentives
• Team incentives
• New ownership models
– Capacity
• Board governance
• Management accounting
• Management science (reconfiguration)
45. The practitioner level
– Information
• High quality patient data
• Disseminating good practice
– Incentives
• Pay rewards
– Capacity
• Professional engagement and development
• Clinical training & reaccreditation
• Leadership skills
46. The individual level
– Information
• Treatment options and implications
• Healthy living
• Comparative provider performance
– Incentives
• User charges and exemptions
• Patient budgets
• Healthy lifestyle
– Capacity
• Help with navigating health system
• Innovative aids to compliance and ‘distance medicine’
47. How can the biggest gains be secured? Some tentative
thoughts on policy priorities for further investigation
• Reconfiguration of services
– Efficient delivery models, especially for chronic disease
– Coordination of care, including long term care
• Information
– Better treatment guidelines, informed by economic criteria, especially for subgroups
– Comparative performance data on insurers and providers, and associated incentives
– Information technology infrastructure
• Funding mechanisms
– Provider payment – P4P?
– Bundling services – eg ‘year of care’
• Health-related behaviour
– Lifestyle (but NB must lead to healthy longer life to reduce expenditure)
– More efficient use of health services
• Workforce
– Professional improvement
– Task shifting
• Technology
– Harnessing potential for improved efficiency of delivery
– Careful assessment of demand-increasing technologies
• Governance
– Assuring good governance at every level
48. Health system efficiency: How to make measurement matter for
policy and management (2016)
Edited by Jonathan Cylus, Irene Papanicolas & Peter C. Smith
• A framework for thinking about
health system efficiency
• Patient classification systems for
efficiency analyses
• Using registry data to compare
health care efficiency
• Management accounting
• Analytic frontier efficiency
measurement methods
• Cost–effectiveness analysis
• Cross-national efficiency
comparisons of efficiency
• Efficiency measurement for policy
formation and evaluation
• Efficiency measurement for
management
49. Our proposed framework for thinking
about efficiency metrics
• five aspects of any efficiency indicator that
should be explicitly considered:
– the entity to be assessed;
– the outputs (or outcomes) under consideration;
– the inputs under consideration;
– the external influences on attainment;
– the links with the rest of the health system.
50. Average length of stay, 2015 or latest
available
0
2
4
6
8
10
12
Turkey
Netherlands
Bulgaria
Denmark
Ireland
Sweden
France
Norway
Cyprus
Iceland
Spain
Belgium
Poland
Slovenia
UnitedKingdom
Estonia
Slovakia
Romania
Lithuania
Portugal
Latvia
Italy
Malta
Austria
Luxembourg
Switzerland
Germany
CzechRepublic
Hungary
Serbia
Croatia
FYRM
Finland
ALOS(days)
51. Framework: Average length of stay
• Are the hospitals in different countries
comparable in terms of the services offered
(entities)?
• Are patient outcomes comparable?
• Are additional hospital days expensive (inputs)?
• Are there external influences such as case-mix
that are not accounted for?
• Are there less intensive parts of the health
system that could treat these patients instead?
52. Final thoughts
• Productivity is a simple concept with a lot of
measurement challenges
• Need to be careful when drawing conclusions from a
small number of metrics
– Metrics should always be used to determine follow-up
questions/metrics
• Lots of metrics about hospitals – is that the area we are
most interested in?
– Should there be new emphasis on questions linked to
allocative efficiency?
• Very little credible evidence on policy priorities for
productivity improvement.
53. Acknowledgements
– Jonathan Cylus and Irene Papanicolas, London School
of Economics
– Josep Figueras, Ellen Nolte and others in the European
Observatory for Health Systems and Policies
– Tamas Evetovits and Sarah Thompson, World Health
Organization
– Chris Murray and others, Institute for Health Metrics
and Evaluation
– Kalipso Chalkidou and others, International Decision
Support Initiative
– Agnes Couffinhal, Gaetan Lafortune, Isabel Joumard,
Mark Pearson and others, OECD