1. Quality Adjusted Life Years (QALY)
Quality of life index
1.0 = normal health
0.0 = death (extremely bad health)
Example
Losing sense of sight
Quality of life index is 0.5
Life = 80 years
0.5 x 80 = 40 QALYs
Most debate about the QoL estimates

2. Quality of life measurement
Typically done with questionnaires
Disease specific
International Prostate Symptom Score
Generic
SF-36, NHP
Utility
HUI, EQ-5D, AQoL, 15D, Rosser index
Utility assessment
SG, TTO, PTO, VAS
For QALY we need utility

3. EuroQol EQ-5D: of the shelf QALY value
MOBILITY
I have no problems in walking about
I have some problems in walking about
I am confined to bed
SELF-CARE
I have no problems with self-care
I have some problems washing or dressing myself
I am unable to wash or dress myself
USUAL ACTIVITIES (e.g. work, study, housework family or leisure activities)
I have no problems with performing my usual activities
I have some problems with performing my usual activities
I am unable to perform my usual activities
PAIN/DISCOMFORT
I have no pain or discomfort
I have moderate pain or discomfort
I have extreme pain or discomfort
ANXIETY/DEPRESSION
I am not anxious or depressed
I am moderately anxious or depressed
I am extremely anxious or depressed

4. Calculate QALY Count life years Value (V) quality of life (Q)
1 = Healthy
0 = Dead
Adjusted life years (Y) for value quality of life
QALY = Y * V(Q)
Y: numbers of life years
Q: health state
V(Q): the quality of life value of health state Q

5. Which health care program is the most cost-effective?
A new wheelchair for elderly (iBOT)
Special post natal care

6. Which health care program is the most cost-effective?
A new wheelchair for elderly (iBOT)
Increases quality of life = 0.1
10 years benefit
Extra costs: $ 3,000 per life year
QALY = Y x V(Q) = 10 x 0.1 = 1 QALY
Costs are 10 x $3,000 = $30,000
Cost/QALY = 30,000/QALY

7. Special postnatal care
Quality of life = 0.8
35 year
Costs are $250,000
QALY = 35 x 0.8 = 28 QALY
Cost/QALY = 8,929/QALY

8. QALY league table
The outcome of QALY-analyses can be used to calculated costs per QALY-league tables, like the one I present here.
The table reads as follows: it takes more than dollars to produce one QALY, if one spend that money on CSF in elderly with leukaemia. If one had spent that dollar on heart transplantation, one would have produced 4 QALY: four times as much. This means that spending you money on heart transplantation is four times more cost-effective than spending your money on CSF in elderly with leukaemia. In other words, the lower the cost-effectiveness ratio in terms of costs per QALY, the better it gets.
At iMTA we calculated the cost per QALY for sildenafil. As one can see in the QALY-league table, the cost-effectiveness of sildenafil is very good compared to all kinds of other interventions.
Given that most of the other interventions are indeed reimbursed in most countries, the favourable cost-effectiveness ratio of sildenafil is a strong argument in favour of reimbursement.

9. Disability Adjusted Life Years (DALYs)
DALYs for a disease are the sum of the years of life lost due to premature mortality (YLL) in the population and the years lost due to disability (YLD) for incident cases of the health condition. One DALY represents the loss of one year of equivalent full health.

10. Disability Adjusted Life Years (DALYs)
Measures healthy time lost from specific diseases and injuries in a population
Comparable and additive across diseases
Ex: Broken scapula = .5 DALYs lost
Protein deficiency = 2 DALYs lost
Based on relatively accessible incidence data (ICD codes)

11. DALY Calculation (the easiest way)
Years lost to disability
Inputs
Duration of disease/injury
Disability weight of disease/injury
% long-term cases
Years of lost life
(YLLs)
Inputs
Life expectancy at age of death
Age at death

12. DALY Calculation: an example
A Two-Car Collision
2 people injured
45 y/o woman – SCI
55 y/o man – fractured rib
YLDs from injuries
Duration (36 year LE) * Disability Wt (.725) = 26 YLDs
Duration (.115 years) * Disability Wt (.199) = 0.02 YLDs
1 family dies
10 year old girl
8 year old boy
38 year old mother
42 year old father
YLLs from deaths
70 year life expectancy
73 year life expectancy
46 year life expectancy
33 year life expectancy
222 YLLs
DALYs +
26.02 YLDs =

13. Cost of Illness (COI) Analysis
Estimate the impact of a disease / condition on the overall costs
Include direct as well as indirect costs
Example:The overall costs for cancer in 2002 in the US was $171.6 billion (ACS, 2003), including
$60.9 billion in direct medical costs
$15.5 billion in indirect morbidity costs
$95.2 billion in indirect mortality costs

14. Budget Impact Analysis (BIA)
Estimate the financial effect of an intervention on a health plan or program
BIA is often requested by managed care organizations in the US or national health insurance programs (e.g., Canada, UK)
Example: Treating all stage IV NSCLC patients in Canada with paclitaxel and cisplatin as outpatients would cost $155 million, an additional $15 million per annum compared to best supportive care.

15. BIA (cont.) Most BIA analysis has a one year time frame.
BIA taking a longer time frame need to consider the impact of new interventions on the underlying disease prevalence and make appropriate adjustments in analyses.

16. What are the DIFFERENCES between each type of analysis?

17. Economic Evaluation Methodologies

18. CBA vs. CEA Uses dollar values for outcome measurements
Maximizes benefit of investment/intervention
Assumes limited resources
Compares programs with different objectives
Uses nonmonetary outcome measurements
Minimizes cost of program
Assumes adequate resources
Compares programs with the same objectives

19. More Concepts in Economic Evaluation …

20. Cost Categories
Inclusion and measurement will depend on the study’s perspective and its time frame.
Direct medical: medical care services
Direct non-medical:
Patient time cost for treatment or intervention
Formal and informal caregiver time
Transportation
Productivity (morbidity and mortality)
absenteeism
Presenteeism
Sometimes costs are categorized as "direct" and "indirect" costs, where direct costs refer to change in resource use attributable to the intervention or treatment, and indirect costs refer to productivity gains or losses related to illness or death.
These costs are from Chapter 6 of the Gold book. You will find that the cost terminologies are not always consistent. For our purposes, I don’t care if you use the Gold terms or not, as long as you are consistent. However, this can be a sticky point with some journals (i.e., with some reviewers). Generally, as long as you offer a complete description of the intervention and its costs, that is sufficient. You can do this without having to mention the words direct or indirect.
Types of resource costs that should be considered include those goods, services, and other resources that are consumed in the provision of an intervention or in dealing with the side effects or other current and future consequences linked to it. These include costs that typically involve a monetary transaction, such as tests, drugs, supplies, health care personnel, and medical facilities (direct health care costs). Other costs related to the intervention, but which may or may not include a monetary transaction, may need to be valued, depending on the perspective of the study (direct non-health care costs). Examples are transportation costs, child care costs, uncompensated caregiver time, and/or the time a patient spends seeking care or participating in or undergoing an intervention (patient time costs). These include travel and waiting time as well as the time actually receiving treatment.
Omission of these costs would bias the CEA against treatments that relied on inputs or outputs that were purchased and in favor of ones that relied on family caregivers or volunteers.
Productivity costs are (1) the costs associated with lost or impaired ability to work or to engage in leisure activities due to morbidity and (2) lost economic productivity due to death.

21. Inflation
Inflation is a sustained increase in the average level of prices. The rate of inflation is the percentage change in average prices from one year to the next
For prices that tend to increase at the rate of general prices (e.g., consumer goods), use the Consumer Price Index (CPI)
For items whose prices rise faster than the general rate of inflation, use a component of the CPI, such as the Medical Care component of CPI
For wages, use either an index of hourly wages or earnings
Why can’t we compare prices in two different time periods without correcting for inflation? Inflation has the effect of reducing the value of a dollar.
For example, $100 in 1932 has different value vs. $100 now.
We need to be able to convert prices to constant dollars so that we can compare their relative prices.
There are two cases when we may need to adjust prices for inflation:
(1) When the data on prices used in an economic evaluation come from different time periods, these should be adjusted to bring the past prices into current terms so that they reflect the opportunity costs of the resources in common dollar terms.
(2) When the study is projecting costs for different time periods in the future, these future prices should be adjusted to account for any real increase or decrease in the price of an item (such as from supply and demand shifts). The adjustment should reflect relative prices net of inflation and increases in productivity or effectiveness.

22. Example
Suppose you want to use information from a published manuscript that listed the cost of a severe adverse event of febrile neutropenia in 1983 dollars to be $1,531. How would you adjust that figure to current dollars?
Index:
1983 (base year)=100
1998 = 242.7
C(1998) = $1,531 * / = $3,716

23. Discounting
Many decisions made today will have repercussions next year and in the years thereafter.
We need a method for comparing the desirability of outcomes that include consequences occurring at different times in the future.
"Inflation" reflects the change in the value of dollars over time, whereas "discounting" deals with people's conception of
"time", in other words, time preference.
For example, if someone is going to give you a car, you probably prefer to have it now than later.

24. The Theory of Discounting
The theoretical justification for discounting is based on two facts:
time preference: most people would accept less money to receive it sooner; and
opportunity cost: less money can be invested by society and allowed to grow at a compound rate of interest to yield the money required for future costs.

25. Discounting Process
Given a stream of costs C1, C2, …, CT, the present value is calculated as:
, where 1/(1+r) t is called the discount factor

26. Issues in Discounting
While there is universal acceptance of the need to discount, there is much controversy over
the appropriate discount rate to use,
whether to discount health benefits as well as costs, and
whether to use the same rate to discount costs and benefits.
In UK: 6% for costs and 1.5 for effectiveness

27. Example

28. Example
A comparison of Programs A and B, adjusted for the differential timing of costs would yield:
PVA = 5/(1.05) + 10/(1.05)2 + 15/(1.05)3 = 26.79
PVB = 15/(1.05) + 10/(1.05)2 + 4/(1.05)3 = 26.81
In this calculation, Program B actually costs more (which is why it is difficult to justify the cost of preventive programs, especially since long-run benefits will be discounted even more than the short-run costs).

29. What is the Theoretical Foundation of CEA?

30. Theoretical Foundation of CEA
Theoretical foundation of CEA was established by a landmark article by Garber & Phelps (1997).
Derive ICER in terms of 3-period U function
E(U) = U1(Y1-C1) + P2(C1)*U2(Y2-C2) P2(C1)P3(C2)U3(Y3)
, where Yi = income;
Ci = medical care expenditure
Pi = probability of surviving into period i

31. CEA and Welfare Economics
Use prob. of surviving as “effectiveness” measure
Incremental cost-effectiveness ratio can be derived from the F.O.C.:
Max. E(U) w.r.t. C1
Decision criteria based on CEA is justified in welfare economics
 achieve optimal resource allocation

32. What are the Decision Criteria under CEA?

33. CEA Framework Two treatments (trx): new (A) vs. old (B) Costs:
Pts in the new trx group: Ca1, Ca2, ….CaK 
Pts in the old trx group: Cb1, Cb2, ….CbJ 
Effectiveness:
Examples of effectiveness measures:
Quality-adjusted life years (QALYs)
Life year saved
Pts in the new tx group: Ea1, Ea2, ….EaK 
Pts in the old tx group: Eb1, Eb2, ….EbJ 
For example, if drug A and B have the same costs, but drug A prevents 100 days of disability while drug B saves one life, CEA cannot be used to help choosing drug A or B.

34. Incremental Cost-Effectiveness Ratio (ICER)
Decision Rule: If IĈER < , then the new treatment is cost-effective
Making inference about the true
(but unobservable) population ICER

35. Making Decisions Using ICER
If the ICER doesn’t fall into the quadrant of dominating or dominating strategy, then decision makings based on CE-ratio become a bit tricky.
Rule 1: value judgement specified by an organization
$20,000 per QALY used in Ontario guidelines
Problems?

36. Making Decisions Using ICER (cont.)
Rule 2: comparison with the commonly used medical procedures.
Rationale: Society should be willing to pay as much for new procedures/technologies as it does for procedures that are currently in common use.
League tables

37. League Table Example

38. Statistical Consideration of CEA?

39. Recent Advances in CEA - 1
Estimate confidence interval of ICER
Statistical Methods:
Box method
Delta Method (Taylor Series Method)
Fieller Theorem Method
Nonparametric Bootstrap Method ….

40. Problems with Inferences Based on ICER
Negative ratios are difficult to interpret
C.I. derived from CE ellipses only make sense when E > 0
 Solution: Net Health Benefit approach

41. Recent Advances in CEA Net Health Benefit Approach Decision rule:
NB() =  E - C
Decision rule:
 Choose the new technology if NB()>0
Methods developed from NHB:
Cost-Effectiveness Acceptability Curve
Bayesian Approach
Regression-based Approach

42. Ten Steps of Performing An Economic Evaluation Study
Establish the perspective
Describe or specify the alternatives
For each alternative, specify the possible outcomes and the probability of their occurrence
Specify and monitor the health-care resource consumed in each alternative
Assign dollar values to each resource consumed

43. Ten Steps of Performing An Economic Evaluation Study (cont.)
Specify and monitor nonmedical resources consumed by each alternative
Specify the unit of outcome measurement
Specify other noneconomic attributes of the alternatives, if appropriate
Analyze the data
Conduct a sensitivity analysis

44. CBA Example Cost per flu shot = $10 Treatment cost per flu = $250
Productivity loss from sick leave = $4,000
Employees = 1000
W/o vaccine: 50 have flu, 3 absence
W/ vaccine: 30 have flu, 1 absence
What should the manager do?

45. CBA Example (cont.) Net Benefit = benefit - cost
=(number of flu avoided)*$250
+ (number of absence avoided)*$4000
- $10*1000
=20*$250+2*$4000-$10000
=$3000 > 0

46. CBA Example (cont.) New flu vaccine available Cost = $20
W/ the new vaccine:
5 have flu, no absence from work
Which one should the manager choose?

47. CBA Example (cont.) NB(new) =45*$250+3*$4,000-$20*1,000 =$3,250
NB(new) > NB(old)
 choose the new vaccine
However, if productivity loss = $3000, then NB(old)=$1000, and NB(new)=$250, then the old vaccine will be chosen

48. CEA Example C-E of two mumps vaccines Perspective:
Several possibilities: state government, or other third party payers.
Describe alternatives:
Vaccine A (old): cheaper, not as effective
Vaccine B (new): more expensive, more effective
Do nothing

49. CEA Example (cont.) Possible outcomes and prob.
Outcomes: Mumps infection, death
probability of infections
Vaccine A:3%, Vaccine B: 0.5%, do nothing: 5%
probability of death:
vaccine A: 0.1%, vaccine B: 0%, do nothing: 0.3%
Health care resource consumed:
Vaccine A: vaccine cost + treatment cost
Vaccine B: vaccine cost + treatment cost
Do nothing: treatment cost

50. CEA Example (cont.) Assign $ to each resource consumed
Vaccine A: $10 /shot, $250 per treatment
Vaccine B: $20 /shot, $250 per treatment
Do nothing: $250 per treatment
Nonmedical resources
Vaccine A: None
Vaccine B: None
Do thong: None
Unit of outcome measures
Death avoided from vaccination

51. CEA Example (cont.) Analyze the data
Note: what we want to construct in CEA is ICER
ICER for vaccine A (vs. do nothing)
DC=(differences in cost between A and do nothing)
=(vaccine cost+treatment cost) -(treating cost) = (10* *250)-(50*250)=5000
DE=(death avoided)=(3-1)=2
Therefore,
ICER(A) = $5000/2 = $2500 (per life saved)

52. CEA Example (cont.) Analyze the data (cont.)
ICER for vaccine B (vs. do nothing)
DC=(differences in cost between B and do nothing) =(vaccine cost+treatment cost) -(treating cost) = (20*1000+5*250)-(50*250)=8750
DE=(death avoided)=(3-0)=3
Therefore,
ICER(B) = $8750/3 = $2916 (per life saved)

53. CEA Example (cont.) Results: (Compare ICER) Sensitivity Analysis
ICER(A) < ICER(B)
Choose vaccine A
Sensitivity Analysis
Used to test how robust the previous conclusions are when assumptions vary.
For example, discount rate, probability of infection, ... etc.

54. CUA Example (cont.) Assume Life Expectancy=50 Utility with mumps=0.8
What’s the outcome measure ?
QALY
That is, 10 years with mumps infection = 8 years in good health

55. CUA Example (cont.) Calculate ICER ICER for vaccine A (vs. do nothing)
DC=5000
DE=(quality adjusted life years saved from death avoided)+(QALY saved from mumps avoided) =(2*50)+(50)*(50-30)*0.8=900
Therefore,
ICER (A) = $5000/900 = $5.55 (per QALY saved)

56. CUA Example (cont.) Calculate ICER (cont.)
ICER for vaccine B (vs. do nothing)
DC=8750
DE=(quality adjusted life years saved from death avoided)+(QALY saved from mumps avoided) =(3*50)+(50-5)*0.8*50=1950
Therefore,
ICER (B) = $8750/1950 = $4.48 (per QALY saved)
ICER(A)=$5.55 > ICER(B)=$4.48

57. THANK YOU…