Strategic Management of Hospitals-Pharmacoeconomics

1

Strategic Management of Hospitals: Pharmacoeconomics as a Decision Tool

Thomas R. Einarson, PhDLeslie Dan Faculty of Pharmacy University of Toronto

2

Outline

Healthcare costs and choices Pharmacoeconomics Definitions Methods Applications Advantages and disadvantages

3

Canadian healthcare spending 1975-2005 (millions)

y = 1E-63e0.0782x

R2 = 0.9662

0

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

180,000

1970 1975 1980 1985 1990 1995 2000 2005 2010

4

Healthcare costs by type

0%

20%

40%

60%

80%

100%

1975 1980 1985 1990 1995 2000 2005

Other

Admin

Public Health

Capital

Drugs

Nursing, etc

Physicians

Institutions

Hospital

5

Trends – hospital $ (Source: CIHI)

Hospital share is decreasing From 44.7% in 1975 → 29.9% in 200528.7% decrease in share

Hospital expenditures$5.5 billion in 1975 [Total: $12 billion]$42.4 billion in 2005 [Total: $142 billion]677% increase [Total: 1064%]

6

Trends – drug $ Drug share increasing

8.5% in 198517.5% in 2005 106% in share

Drug costs $1.1 billion in 1975 $24.8 billion in 2005 2200% increase

7

Reasons for increases Increase in utilization

More people taking drugs More population More >65, free drugs (access)

More disease → more drugs Longevity → more disease → more drugs

Increase in cost Research costs money New drugs research intensive

8

Canada: population increase 1966-2006

y = 0.3007x - 571.25

R2 = 0.992

15

20

25

30

35

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

9

Average cost per citizen covered by BC Pharmacare

Year Cost

1987 $192

1991 $354

1995 $406

1999 $478

10

Drug costs on the rise

11

Costs of therapy (cancer drugs)

600

1,620 4,980

7,600

12,000

18,000

26,000

0

5000

10000

15000

20000

25000

30000

$ p

er t

reat

men

t

Drug acquisition costs

12

Selecting drugs in hospital Open formulary

No restriction Pay for all drugs prescribed

Formulary Most standard drugs available Medical staff + pharmacy decide on content

Problem: demand exceeds budget Need a method for judging relative merits

13

Background

Economics = the study of the allocation of resources

Health economics: application of economic principles to healthcare

Basis = value for money Money = common denominator for

comparisons

14

Pharmacoeconomics

Pharmaco = drugs Application of health economics to drugs

and drug services Techniques not different Interpretations parallel

15

Levels of analysis Macro (population)

Most PEA is this type Applies to the average in the population Requires translation to apply to patients

Meso (group) Usually applied at this level

Micro (Individual)Never done at this level in PEA

16

Definition

Pharmacoeconomics:Evaluation of a drug against one or more comparators with respect to both costs and outcomes. Comparator = depends on purpose

Placebo, do nothing Standard treatment

17

Why?

Resources scarce Demands increasing Forced to make choice Pharmacoeconomics assesses

value for money (common metric) A tool for decision making

18

Perspective Whose viewpoint?

Need to identify the audience for the report Need to specify the analytic viewpoint

Determines:Data collected Valuation of resources Interpretation of results

19

Analytic perspectives

Society (Comprehensive, overall)All costs considered, regardless of payer Usually includes “indirect” costs

Payer Insurer, government, individual

Standard: Societal + Ministry of Health (Canada)

20

Types of cost

Direct costs Indirect costs

Economic NOT accounting definitions used in PEA Overhead costs included in hospital

perspective

21

Direct costs

Costs of resources consumed in treating patients Cost centres:

Drugs (preparation, administration, monitoring) Medical care (MD, specialist) Hospital (inpatient, units) Laboratory (blood, urine, x-ray, NMR, CAT scans) Allied healthcare (nursing, physio, massage, social work) Transportation to therapy (sometimes) Cost of managing ADRs

22

Using costs

Use standard lists for cost values The standard price that any person would pay Do NOT use special costs, bargains Exception: hospitals analyzing their own data

Select year of analysis Usually current year or immediate past year Standardize all costs to year of analysis

Use health component of Consumer Price Index or equivalent

23

Indirect costs Lost productivity

Early death Time totally lost forever

Decreased ability to work Absenteeism due to disease

“Presenteeism” On the job, being paid, not inefficient

Caretaker time Problems

Leisure time, retired people Consider lost time equal

Under-employment/unemployment Use Friction method Counts only time to train replacement personnel

24

Time horizon Analysis must be over enough time to

allow for all important outcomes Cancer drugs use lifetime Antibiotics (acute) – 1 month/3 monthsOften use 1 year

Problem Data availability over time Must discount costs if >1 year

25

Discounting

Preference for benefits now and costs later NOT because of inflation Based on stable rates for safe investments

(government bonds, GICs, cash market) Standard = 5%, vary in sensitivity analyses

over range 0-10%

26

Summary: costs

A: Identify resources consumed B: Quantify resource use C: Qualify (cost) that use

Use standard costs Discount if over >1 year

Total cost = Resourcei*Costi

Average across all patients Extrapolate to population using

demographic statistics

27

Outcomes: ECHO model Economic

Dollars Clinical

Positive: Cure, life saved, life-year gainedNegative: Case avoided (disease, death)

HumanisticQuality of life, patient preferences

28

QALYs Quality adjusted life-years

e.g., cancer patient Estimate duration of life remaining

Identify health states involved e.g., Phases II-IV

Measure length of each health state Adjust each by utility value:

Timei*Utilityi

Average over all patients Universal outcome

Can compare across treatments Can compare across diseases

29

Utilities Standard gamble Time tradeoff Visual analog scale Whose?

Patients Accurate estimate of implications Problem: adaptation to the condition

Normals (society) Need to help them understand the disease

30

Preferred outcomes

Ultimate outcomes:Cure, life-year gained, QALYAvoided cases of:

Disease (vaccine) Death

Avoid intermediate outcomes Decrease in BP, serum cholesterol, units on a

scale (pain, health status, etc.)

31

Types of analyses

Cost-minimization

Cost-effectiveness

Cost-utility

Cost-benefit

Comparison of efficacy /

effectiveness

Cost comparisonTwo or more alternatives

Cost-outcome description (Non-comparative CBA)

Outcome assessment (e.g., Quality of Life measurement)

Cost description

(Cost of Treatment, Burden of Illness)

One (No comparison of alternatives)

Costs + Outcomes

Outcomes only

Costs onlyAlternatives

32

Cost analyses (non-PEA)

Burden of sickness Cost of illness Financial feasibility

Implementing a new pharmacy service E.g., cost-revenue model

33

Burden of sickness/illness Macro level (population based) “Top down” analysis Dollars spent on a disease Direct costs

Drugs, MD, hospital Indirect costs

Morbidity, mortality, time lost from work, decreased productivity (presenteeism), etc.

34

Cost of illness/Cost of treatment Micro level “Bottom up” analysis Individual based Itemize resources consumed

Drug, visits to MD, tests, etc. Chart review, database analysis, modelled

35

Approaches to COI analysis Incidence based

Start at onset of disease Follow until resolution/death E.g., Gonzalez JC, Einarson TR. Cost of Type-2 diabetes in Colombia.

PharmacoEconomics 2008 (in press).

Prevalence based Cross sectional All cases in one year Example: Pain due to multiple sclerosis: analysis of

the prevalence and economic burden in Canada. Pain Res Manag 2007;12(4):259-65

36

Contribution

Identifies costs Burden to society Cost to treat individual

Basis for pharmacoeconomic evaluation (Baseline)

BUT, does not address appropriateness or outcomes

37

COI Example: bisphosphonatesResource CostBisphosphonate $5,421Other Drugs $6,290Hospital $18,229Clinic $7,715Laboratory $2,020Imaging $3,355Radiation $2,657Physicians $2,170Total $47,857

Kruk. Supportive Care in Cancer 2004; 12: 844-51.

38

Burden of ADRs (USA)

Resource Events Cost (millions)Physician visits 115,654,949 $7,460Additional prescriptions 76,347,604 $1,933Emergency visits 17,053,602 $5,321Hospital admissions 8,761,861 $47,445LTC admissions 3,149,675 $14,399Deaths 198,815 Total $76,558

Johnson. Arch Intern Med 1995;155(18):1949-56.

39

HOSPITAL CARE EXPENDITURE BY DIAGNOSTIC CATEGORY, CANADA, 1993

DIAGNOSTIC CATEGORY 1993 Cost($1000)

% ofTotal

Cardiovascular Diseases 4,862,322 18.6

Mental Disorders 3,631,788 13.9

Cancer 2,467,408 9.5

Injuries 2,253,487 8.6

Digestive Diseases 2,093,374 8.0

Respiratory Diseases 1,788,230 6.9

Pregnancy 1,649,707 6.3

Well-patient Care 1,349,205 5.2

Musculoskeletal Diseases 1,285,910 4.9

Genitourinary Diseases 1,075,828 4.1

Nervous System and Sense Organ Diseases 792,862 3.0

Endocrine and Related Diseases 526,785 2.0

Perinatal Conditions 518,012 2.0

Infectious and Parasitic Dieseases 345,071 1.3

Birth Defects 231,923 0.9

Skin and Related Diseases 223,015 0.9

Blood Diseases 156,808 0.6

Ill-defined Conditions 844,567 3.2

TOTAL 26,096,300 100.0

SUBCATEGORY

Coronary Heart Disease 1,571,999 6.0

Stroke 1,258,063 4.8

Chronic Bronchitis / Emphysema / Asthma 636,491 2.4

Motor Vehicle Traffic Accidents 283,106 1.1

Diabetes 274,621 1.1

Female Cancers 206,068 0.8

HOSPITAL CARE EXPENDITURE BY DIAGNOSTIC CATEGORY, CANADA, 1993

DIAGNOSTIC CATEGORY 1993 Cost($1000)

% ofTotal

Cardiovascular Diseases 4,862,322 18.6

Mental Disorders 3,631,788 13.9

Cancer 2,467,408 9.5

Injuries 2,253,487 8.6

Digestive Diseases 2,093,374 8.0

Respiratory Diseases 1,788,230 6.9

Pregnancy 1,649,707 6.3

Well-patient Care 1,349,205 5.2

Musculoskeletal Diseases 1,285,910 4.9

Genitourinary Diseases 1,075,828 4.1

Nervous System and Sense Organ Diseases 792,862 3.0

Endocrine and Related Diseases 526,785 2.0

Perinatal Conditions 518,012 2.0

Infectious and Parasitic Dieseases 345,071 1.3

Birth Defects 231,923 0.9

Skin and Related Diseases 223,015 0.9

Blood Diseases 156,808 0.6

Ill-defined Conditions 844,567 3.2

TOTAL 26,096,300 100.0

SUBCATEGORY

Coronary Heart Disease 1,571,999 6.0

Stroke 1,258,063 4.8

Chronic Bronchitis / Emphysema / Asthma 636,491 2.4

Motor Vehicle Traffic Accidents 283,106 1.1

Diabetes 274,621 1.1

Female Cancers 206,068 0.8

Pharmacoeconomic analysis

Properties Models Outcomes Issues

41

Pharmacoeconomic analysis

Compare 2 or more options Need to identify, describe each Usually a new drug trying to enter the market

Requires an estimate for each of: Costs Outcomes produced

Other requirements: Analytic time horizon, perspective, discount rate(s) Clinical consequences and their management

42

Purpose of PEA

Permission to market drug NOC in Canada License in other countries

Acceptance by formulary Provided by system (e.g., ODB) Paid by insurer (in whole or in part)

Purpose model, approach

43

Approaches: licensure Population based model Evidence = Efficacy data

Data source = RCTs Assume full compliance, ideal conditions Per protocol analysis

Defined population Limited to eligible patients who take full treatment

Comparator = placebo, standard Viewpoint = societal

44

Approaches: formulary Meso (Group) model Evidence = Effectiveness data

Data source: Effectiveness trials

Real life experience Unselected patients Compliance matters

Model via RCT + other factors (adherence) Intent-to-treat analysis

Population = all possible candidates

Comparator = standard therapy Viewpoint = payer

45

Product Need to describe drug in detail Identify therapeutic category

ATC system Comparators

What is being used now Utilization rates, costs, burden Justify WHY this drug is being examined

46

Population Indication

Disease How diagnosed, classified, e.g., ICD-10

Severity Mild, moderate, severe

Population Types of patients

Males, females Ages

47

Comparator Placebo

Form identical to drug of interest Do nothing alternative

Standard therapy Guidelines Actual practice Expert panel

Hospital setting: Usual treatment

48

Pharmacoeconomic models Prospective RetrospectivePredictive

49

Prospective models

Pharmacoeconomic RCTDesigned specifically for PEA

Add-on to clinical RCT “Piggy-back” trials

Collect data Identify resources usedCost concurrently or later

50

RCT advantages Prospective data collection

Data are stochastic Can calculate Mean, SD

All factors controlled “Ideal” conditions Provides efficacy data

Valid for NOC, licensing

Disadvantages: Sample not representative

Patients, conditions, drugs taken, age, sex, etc. Cannot extrapolate results to real life

51

Retrospective models

Model existing RCT Chart review Database analysis

52

Modelling RCTs Easy to do

Efficacy from trial Determine costs

Problems: Retrospective bias Artificial conditions (extrapolation) Requires expert input

53

Chart Reviews Easy to do, not expensive (usually) Problems:

Time consuming Missing data

Outcomes Resources used (not planned for) Patients d/c, re-entry, move

54

Database Analyses

Available, easy to use Powerful - huge samples Versatile:

Cohort, case control, matchingTime series, cross sectional

55

Disadvantages

Missing data Diagnosis, outcomes, status (smoking, alcohol, etc)

Confounding Causation = ?

Assumptions (Consumption, indication, compliance)

Cost = large, delays long

56

Predictive decision models Most common model Relatively easy to perform

Software (TreeAge/Data®)

Often require expert panel Applies to “average” patient Population based

57

Decision Types

Decision treeFixed time periodNon-recursive (one-way only) All outcomes included

Markov modelRecursive model (can move backwards)Cumulates values over timeVersatile

58

Markov model

EDSS 6-6.5

EDSS 5-5.5

EDSS 4-4.5

EDSS 3-3.5

EDSS 2-2.5

EDSS 0-1.5

Monosymptomatic

Clinically Defined MS

Second Event

1st Year Transitions

Expanded Disability Status ScaleIskedjian et al. Multiple Sclerosis 2005; 11: 542-51

59

Markov model

Monosymptomatic

EDSS 0-1.5

EDSS 2-2.5

EDSS 3-3.5

EDSS 4-4.5

EDSS 5-5.5

EDSS 6-6.5

Second Event

Multiple Year Transitions

Iskedjian et al. Multiple Sclerosis 2005; 11: 542-51

60

Base case analysis Identify what happens to the typical

patient Number of treatments undergoneComponents of each treatment

Resources consumed

What happens if success, fail, etc.

61

Requirements for PEA

Drug of interest Comparator(s) Measurements of both

Success (and/or other) rates Costs of inputs (in $) Patient outcomes

Outcomes expressed incrementally

62

Clinical efficacy

Source of data RCT

Head-to-head or Indirect Observational studies

Type of data Intention to treat Per protocol (completers)

63

PEA types

CCA = cost consequence analysis CBA = cost benefit analysis CMA = cost minimization CEA = cost effectiveness CUA = cost utility

64

Cost consequence (CCA)

No summary statistic calculated Costs valued in $ Outcomes quantified and listed Reader judges importance Could be in every PEA

65

Cost benefit (CBA)

Measurement: Costs – $ Outcomes – $

Benefit: Cost ratio calculatedB:C >1 socially beneficial

Incremental ratio used

66

CBA example

If a program costs $2000 and produces $5000 in benefits, as compared with the existing program costing $1000 with benefits of $2000,

Ratio1 = $5000/$2000 = 2.5 Ratio2 = $2000/$1000 = 2.0 Since ratio1:2 = 2.5/2.0 =1.25 >1, the new

program produces more benefits and is preferred, if affordable

67

CBA Advantages:

Theory basedCan use willingness to pay

Contingent valuation

DisadvantagesValuations

Intangibles (pain, suffering)

68

Cost minimization (CMA)

All outcomes same Must be demonstrated

Consider only costs Decision rule: lowest cost is preferred

choice Example: gent vs tobra ??

69

Cost-effectiveness (CEA)

Costs = $ Outcomes = natural units

Cures, deaths avoided Intermediate outcomes

Not desirable (need linkage)e.g., Blood level (cholesterol)

70

Cost utility (CUA)

Cost = $ Outcome = patient utility QALY = quality adjusted life year Assess quality of life utility Quantify and adjust by utility Compare incrementally between drugs

71

Interpreting outcomes

Incremental cost-outcome ratio:

Cost1 – Cost2__________________

Benefit1 – Benefit 2

72

Incremental analysis

Compare costs and outcomes Ideal = lower cost + better outcomes Usual = more cost more outcome Problem = assessment

73

Incremental CEA

If cost lower + benefit greater = dominant If incremental cost cost/outcome, then

CEA If lower cost + lower outcome, could be

OK

74

Interpretation of PEA results

Outcome+

Outcome-

Cost+

IncrementalCE

Dominated

Cost-

Dominant ???

75

Interpretation: Suggestion

76

CEA example

New drug cost = $500, cure = 0.7 (70%) Old drug cost $400, cure = 0.5 (50%) ICER = ($500 - $400)/(0.7 – 0.5) = $500 per

additional cure Old drug cost $400/0.5 = $800/cure; therefore,

the new drug is cost-effective and should be adopted, if it can be afforded

77

CEA example 2…

New drug cost = $500, cure = 0.7 Old drug cost $100, cure = 0.5 ICER = ($500 - $100)/(0.7 – 0.5) = $1,000 per additional cure Requires judgment (no real rules)

78

CEA interpretation

Dominant = adopt (if affordable) If the incremental cost cost/outcome of

standard, then the new drug may be considered to be cost effective

Otherwise, judgment required

79

Summary PEA requires at least 1 drug and 1

comparison treatment PEA involves simultaneous analysis of

costs AND outcomes Decision makers prefer dominant

treatments – they provide savings; incremental cost-effectiveness involves increased costs

80

Trucha canadiense