Real Life Health Technology Assessment

Peter C. Coyte
Professor of Health Economics &
Former CHSRF/CIHR Health Services Chair
Dalla Lana School of Public Health, University of Toronto

May 4, 2017
Email: peter.coyte@utoronto.ca
Web: http://ihpme.utoronto.ca/faculty/peter-c-coyte/
 Presentation Objectives
• To outline the important methodological choices that are
required to be made in order to embark on an appropriate
approach to a specific health technology assessment;

• To describe the motivation for, the methods used, and the

findings derived from three distinct economic evaluations
conducted in the last decade:
– Cost-Utility of Transcatheter vs. Surgical Aortic Valve Replacements;
– Cost-Effectiveness of Expanding Services for Children with Autism; and
– Does Place of Death Impact End-of-Life Costs?
Choice of an Appropriate Evaluative Approach
• Economics is about choice, where choice concerns
alternative allocations of scarce resources.
• Similarly, as a health technology assessment analyst you
will need to make important choices about the methods
you will use in your work.
• I cannot underestimate the important of this stage of the
work; use of inappropriate methods not only means that
you have wasted your time, but it results in scientifically
flawed findings that may distort the allocation of scarce
resources.
 Create a Marketing Plan for Your HTA
• The marketing plan MUST be outlined BEFORE you start
your HTA.
• This component of your work is essential because you are in
the business to MARKET your findings. You are a
salesperson…..similar to an employee at Central!
• You need to identify conferences where you may present
your findings and journals that may publish your work.
• You need to identify managers within your organization, and
more broadly, policy-decision makers who may been eager
to use your work to inform their decisions.
• If you do not spend time to market your work then no matter
how valid and reliable is the science it will never be known
or used, and therefore may have no impact!
Five Key Methodologic Components to HTA:
(1) Choice of an Appropriate Counterfactual
• The main consideration when undertaking HTA
concerns the focus of the evaluation which in turn
entails the selection of an appropriate counterfactual.
• A ‘counterfactual’ is the benchmark against which an
intervention or health promoting program is assessed.
– Drummond et al (2005, p9) defines economic evaluation as “..the
comparative analysis of alternative courses of action”. Thus, an
alternative(s) to a proposed intervention or policy warranted, and
that alternative is what constitutes the counterfactual.
Five Key Methodologic Components to HTA:
(2) Choice of the Time Horizon
• The second crucial choice concerns the selection of an
appropriate time horizon for the assessment of the costs and
consequences of an intervention or policy shift.
– The time horizon determines how far into the future the costs
and consequences are taken into account.
– The time horizon must be long enough to capture all relevant
differences in costs and consequences associated with the
alternative interventions under review.
– “.[T]he relative importance of future costs will depend on the
disease and specific interventions (acute illness vs prevention
strategies)” being assessed. Tarrida et al (2009).
– The appropriate time horizon to selected depends on the future
trajectory of costs and health outcomes and the extent to which
those trajectories are impacted by the intervention assessed.
Five Key Methodologic Components to HTA:
(3) Choice of the Perspective or Viewpoint
• A new health technology or policy impacts a range of
costs and consequences. Some are borne by various
levels of government, some by private citizens, and some
by 3rd parties, such as insurance companies.
• The perspective or viewpoint adopted in the analysis plays
a significant role in determining which costs and
consequences to include and which to ignore.
• When considering the most efficient use of resources for
society, it is important to ensure a broad perspective so
that the impact of the technology or policy on all
stakeholders is captured. For other purposes, alternative
perspectives to a societal perspective may be appropriate.
Five Key Methodologic Components to HTA:
(4) Choice of the Target Population
• The analyst needs to specify the target population that is
the subject of the analysis.
• For health interventions, the target population is often a
specific group of patients with a diagnosis and severity of
disease.
• For a health policy, the target population may be a group
of citizens with lifestyles that are the focus of the policy.
• For example, for health promoting interventions, the
appropriate target population may be those citizens
engaged in the harm-inducing lifestyles that are subject of
the policy.
 Five Key Methodologic Components to HTA:
(5) Choice of Appropriate Data to Use
• Appropriate data is crucial to use in order to conduct a
health technology assessment.
• Indeed, valid and reliable data and methods must be used
as much as is feasible to measure the effect of an
intervention or policy on the relevant streams of costs and
health outcomes.
 Three Economic Evaluation Publications
Cost-Utility Analysis of Transcatheter vs. Surgical Aortic Valve Replacements:
• Tam DY, Hughes A, Youn S, Hancock-Howard RL, Coyte PC, Wijeysundera HC,
Fremes SE: Cost-utility analysis of transcatheter versus surgical aortic valve
replacement for the treatment of aortic stenosis in the intermediate surgical risk
population. To submit to Journal of Thoracic and Cardiovascular Surgery

Cost-Effectiveness of Expanding Services for Children with Autism:

• Motiwala SS, Gupta S, Lilly M, Ungar WJ, Coyte PC: The Cost-Effectiveness of
Expanding Intensive Behavioural Intervention (IBI) to All Children with Autistic
Children in Ontario. Healthcare Policy, 1:2, 125-141, 2006.

Does Place of Death Impact End-of-Life Costs?

• Yu M, Guerriere DN, Coyte PC: The Societal cost of home and hospital end-of-life
care for palliative care patients in Ontario, Canada. Home and Social Care in the
Community, 23:6, 605-618, 2015.
 Cost-Utility Analysis of Transcatheter vs.
Surgical Aortic Valve Replacements
• Surgical aortic valve replacement (SAVR) was the gold
standard for the treatment of severe aortic stenosis (AS).
• Transcatheter aortic valve implantation (TAVI) has now
become the main treatment for patients with severe
symptomatic AS who are at high surgical risk or where
they are nor surgical candidates.
• The success of TAVI lead to a RCT for AS patients with
intermediate surgical risk that showed non-inferiority
compared to SAVR in mortality at 30-days and at 2-years.
• The purpose of this study was to conduct a lifetime cost-
utility analysis of TAVI compared to SAVR for AS patients
at intermediate surgical risk.
 Methods I
• A public health system payer perspective was adopted.
• A lifetime Markov model with 30-day cycles was
constructed with transition probabilities based on Trial
data from PARTNER 2.
• All patients enter as either TAVI or SAVR patients and
were at risk for both short- and long-term
complications. Following the procedure all patients
moved to one of five health states (alive and well, re-
hospitalization, stroke, dialysis and death).
• Physician fees, inpatient costs based on case-mix
groups and device costs were placed in 2016
Canadian dollars. Complication costs were drawn from
CIHI’s Patient Cost Estimator.
Decision-Tree for TAVI vs. SAVR
Post-Procedure Markov Model
 Methods II
• Effectiveness was measured in terms of quality-
adjusted life years (QALYs).
• The utility scores (to measure quality of life) for
the different health states were obtained from the
literature for patient populations similar to those
eligible for PARTNER 2 Trial.
• Various sensitivity analyses were conducted.
• The Monte Carlo simulation used to perform the
probabilistic sensitivity analysis assessed
uncertainty in probabilities, costs and utilities.
• Probabilities and utilities were assigned to beta
distributions and costs to gamma distributions.
Transition Probabilities
Incremental Lifetime Costs and Quality
Adjusted Life Years for TAVI vs. SAVR

TAVI SAVR TAVI vs. SAVR

Incremental $46,690 $36,646 $10,044

Costs

Incremental 4.612 4.420 0.192

QALYs

ICER TAVI vs SAVR

$52,197/QALY
 Scatter Plot of the Incremental Cost and
Incremental Gain in QALYs of TAVI vs. SAVR

.
Cost-Effectiveness Acceptability
Curve for TAVI vs. SAVR
Cost-Effectiveness of Expanding Services for
Children with Autism
• Autism is an early-onset developmental disability that often
results in a range of impairments.
• Most cases are diagnosed by three years of age and is much
more common among boys than girls.
• Intensive behavioural Intervention (IBI) captures therapies
provided to children with autism to alleviate their disabilities.
• Ontario publicly funds IBI for up to three years but for only
1/3 all autistic children under six years of age.
• This study evaluated the cost-effectiveness of expanding
funding for IBI to all autistic children Ontario from 2 to 5
years of age. A government perspective was taken and
children were assessed until age normal retirement.
 Dependency Outcomes
• Autistic children were assigned to one of
three levels of functioning: normal; semi-
dependent; and very dependent.
• The efficacy of IBI in moving children
between levels of function were based on
the limited literature at the time.
• The distribution of children by functional
status was based on 3 scenarios: No IBI,
the current status quo where 37%
received services; and full expansion.
 Efficacy Rates Stratified by Levels of
Function for Each Scenario
Efficacy Rates:
Functional State: No Intervention Status Quo Expansion
Normal 25% 26.9% 30%
Semi-Dependent 25% 34.3% 50%
Very Dependent 50% 38.9% 20%

Present Value of 9.6 years 11.2 years 14.0 years

Dependency-free
life years gains
 Government Costs
Three sets of age-related costs borne by the
provincial government were assessed for
each of the three scenarios:
1.Intervention costs over ages 2-5 years;
2.Schooling costs over ages 5-18 years; and
3.Adulthood costs over ages 18-65 years.
Each set of costs included therapy costs,
health care services, educational supports,
adult housing supports, and other supports.
Present Value of Average Lifetime Costs
Per Person by Intervention Scenario and
Functional Level
Functional Level
Normal Semi- Very Average
Dependent Dependent Cost by
Scenario
No $79,604 $812,269 $1,582,693 $1,014,315
Intervention
Status Quo $158,909 $891,574 $1,661,998 $995,074
Expansion $293,645 $1,026,310 $1,796,734 $960,595
 Cost-Effectiveness Plane
PV of Costs

No

SQ

PV of Dep. Free Years

Does Place of Death Impact End-of-Life Costs?
• Population aging, combined with health reforms have
dramatically increased the emphasis on home-based care,
and in the last decade, home-based palliative care.

• There is a preference for home environment during the

final stage of life and it is suggested that this setting is
also less costly.

• The purpose of this study was to assess societal costs of

end-of-life care incurred when death took place either at
home or in a hospital for those in receipt of home-based
palliative care.
 The Methodological Challenge
• The main challenge in undertaking this cost analysis was to
address the potential biases that result when the intervention
assessed (place of death) may be affected by caregivers’,
patients’, and physicians’ decisions about hospital admission.
• Patients who die at home may therefore be systematically
different than those who die in a hospital even though they may
be identical when first enrolled in a home-based palliative care.
• A propensity score stratification method was used to adjust for
these potential biases.
 Kaplan-Meier Survival Curves by the
Initial Palliative Performance Scale Score
Patient Characteristics by Place of Death
Continuous variables Home death mean Hospital death mean P-value for between group difference

Days home 73.35 60.65 0.285

Baseline Caregiver burden 25.98 29.50 0.003
Baseline Palliative Performance Score 32.57 28.69 0.001
Patient age 73.15 72.38 0.677
Caregiver age 58.54 56.59 0.304
Comorbidity score 6.81 6.86 0.730
Deprivation score 0.76 0.62 0.020

Binary variables Home death (n) Hospital death (n) P-value for between group difference

Patient sex: Male 42 42 0.107

Female 63 39

Caregiver sex: Male 41 24 0.182

Female 64 57

Preferred for Place Home 92 38 0.000

of Death Other 13 43

Living arrangement Alone 13 21 0.018

With others 92 60
Distribution of Total Costs for Palliative Care Trajectory
Publicly financed costs $ % of Total Cost
Outpatient service costs Public - Medications $619.87 1.81%
Public - Supplies $4451.31 13.02%
Public - Tests $418.79 1.22%
Public - Appointments inside the home $5491.93 16.06%
Public - Appointments outside the home $327.49 0.96%
Public - Hospice $2.64 0.01%

Hospitalization costs Hospitalization $4558.61 13.33%

ER visit costs ER visits $105.36 0.31%

Total public costs $15976.00 46.72%

Privately financed costs $ % of Total Cost
Out-of-pocket costs Out-of-pocket - Medications $103.81 0.30%
Out-of-pocket - Supplies $207.17 0.61%
Out-of-pocket - Appointments inside the home $1429.82 4.18%
Out-of-pocket - Travel costs $277.71 0.81%

Unpaid caregiver time costs Time lost from household work $2085.17 6.10%
Time lost from leisure $13728.42 40.14%
Time lost from employment $84.06 0.25%

Third party costs 3rd party - Medications $241.79 0.71%

3rd party - Supplies $7.34 0.02%
3 party - Appointments inside the home
rd $54.1 0.16%
3rd party - Hospice $2.34 0.01%

Total private costs $18221.73 53.28%

TOTAL COST $34197.73

 Pre-Stratification Costs by Place of Death
6 months prior to death (n=178)

Costs Home % of total Hospital % of total P-value for Pooled % of total

death cost death cost between mean cost cost
mean mean group
difference

Total cost $31910.54 $29116.62 0.467 $30717.63

Outpatient services $12103.93 37.93% $7837.75 26.92% 0.015 $10282.41 33.47%

cost
Hospitalization cost $2111.84 6.62% $7733.86 26.56% 0.000 $4512.25 14.69%

ER visits cost $59.12 0.19% $156.20 0.54% 0.002 $100.57 0.33%

Out-of-pocket cost $1998.19 6.26% $1800.42 6.18% 0.727 $1913.75 6.23%

Unpaid caregiver time $15324.88 48.02% $11416.84 39.21% 0.050 $13656.28 44.46%
cost
3rd party cost $312.59 0.98% $171.55 0.59% 0.258 $252.37 0.82%
Post-Stratification Costs by Place of Death
6 months prior to death (n=178)
Variable Home death mean Hospital death mean Difference: P-value

Quartile 1
Total cost $46760.71 $31215.06 0.124
Outpatient services cost $11962.25 $7142.65 0.121
Hospitalization cost $743.62 $10623.85 0.123
ER visits cost $94.22 $204.14 0.404
Out-of-pocket cost $3008.53 $2143.07 0.593
Unpaid caregiver time cost $29566.90 $10923.28 0.001
3rd party cost $1385.18 $178.08 0.002
Quartile 2
Total cost $26978.36 $25256.43 0.813
Outpatient services cost $8939.72 $6911.28 0.424
Hospitalization cost $3399.35 $5983.04 0.369
ER visits cost $78.52 $83.75 0.928
Out-of-pocket cost $1877.83 $1399.51 0.671
Unpaid caregiver time cost $12353.57 $10655.04 0.623
3rd party cost $329.38 $223.81 0.731
Quartile 3
Total cost $23095.35 $30944.34 0.282
Outpatient services cost $7439.76 $12059.37 0.061
Hospitalization cost $1890.72 $3068.09 0.328
ER visits cost $48.63 $154.18 0.098
Out-of-pocket cost $1374.76 $1175.94 0.759
Unpaid caregiver time cost $12238.74 $14276.22 0.678
3rd party cost $102.74 $210.55 0.401
Quartile 4
Total cost $42224.39 $36152.60 0.649
Outpatient services cost $18609.80 $10432.53 0.288
Hospitalization cost $1811.74 $7106.27 0.003
ER visits cost $53.84 $134.60 0.176
Out-of-pocket cost $2530.30 $2984.76 0.841
Unpaid caregiver time cost $18874.97 $15465.39 0.578
3rd party cost $343.74 $29.06 0.453
 Conclusions
• Devote time and attention to the marketing of your work.
Before starting any evaluation know where you will publish
your findings and plan your knowledge translation activities.

• Make sure you employ valid and reliable methods and that
you have adopted an appropriate evaluation approach.

• There is tremendous scope to apply HTA methods to an

array of heath technologies, clinical practices, managerial
decisions, lifestyle choices, and health policy strategies.
Humanistically-Informed Scientists &
Scientifically-Informed Humanists.

Thank you!

Peter.coyte@utoronto.ca