Cost effectiveness analysis

From Wikipedia

Contents 1 Identify the problem, research objective and questions 2 Define the perspective (to determine the costs to be included) 3 State the strategies compared 4 Model selection 4.1 Decision analysis 4.1.1 Construct the decision tree 4.2 Markov model 5 Inputs 5.1 Decision analysis 5.1.1 Probability values 5.1.2 Cost data 5.1.2.1 Discounting costs : cost adjustment considering future variations 5.1.2.2 Inflation adjusted costs 5.1.3 Utility data (Cost utility analysis) 5.2 Markov model 5.2.1 Health states 5.2.2 Cycle length: time elapsed between successive outcome evaluations 5.2.3 Transition probabilities 5.2.4 Costs 5.2.5 Utilities (Cost utility analysis) 5.2.6 CI (if available, for Monte Carlo analysis) 6 Developing a model structure 6.1 Decision analysis tree 6.2 Markov model 7 Analysis 1 OR 2 7.1 Analysis 1 7.1.1 Decision tree 7.1.2 Cost 7.1.3 Cost effectiveness 7.2 Analysis 2 7.2.1 Analysis using Expected value 8 Sensitivity analysis 8.1 One way sensitivity analysis 8.2 Two-way sensitivity analysis 9 Results 10 References

Identify the problem, research objective and questions

Define the perspective (to determine the costs to be included)

• Payer (for example, Insurance company: amount of money paid to the hospital
• Provider (for example, Hospital): true cost of providing service
• Individual: amount that is paid out of pocket
• Societal: total cost to society (service and productivity)-all payers for all persons
• Program: cost and the long term outcomes

State the strategies compared

Model selection

Decision analysis

Simple decision tree are used for diseases where the decision affects only the immediate outcome

Construct the decision tree

Example 1

Example 2

Markov model

Markov modeling is used for long-term analysis example, chronic diseases used when the outcome of the intervention is not simple dichotomous and the chain from intervention to outcome involves several events, requiring estimation of many probabilities (when the outcome is more than one OR when there are many intervening events between intervention and outcome)

Inputs

Decision analysis

Probability values

Record the probabilities of each chance event on the decision tree

• Sources
  • literature review (meta analysis, primary data collection)
  • consultation with experts
  • Medicare CPT/DRG coding (example, Medicare CPT reimbursement for NC or DRG coding using specific software program or assistance from coders
• If probabilities are not available, seek expert judgment to calculate from the available data in the publications
• Identify certain assumptions that will be required for the model
• Derive utility scores for quality of life

Cost data

• Common sources: administrative sources like insurance companies hospital billing records
• Identify different components of cots defined below, to be used in the study
  • Production cost: total cost of resources consumed in the production of a good or a service
  • Direct cost: the cost of materials and labor that go into production of a good or service
    • Healthcare resources: Medications, Hospital, Clinic, Practitioner, Therapy, tests, procedures,Administration, Nurse time, adverse events, Miscellaneous
    • Non-healthcare costs: Other out-of-pocket expenses of patients, Transportation, Telephone
  • Overhead: The cost of all other items needed to produce good or service
  • Induced cost: intervention related cost, which would not have occurred otherwise (example, adverse events)
  • Indirect cost: Unpaid domestic assitance, days lost from work, decreased productivity
  • Opportunity cost: value of all cost in an alternative use
  • Intangible: Pain, sufferings etc.

Discounting costs : cost adjustment considering future variations

• Calculation:
  • Cpresent=C0+ C1/(1+r)1 + C2/(1+r)2 + ..................+ Cn/(1+r)n
    • Cpresent: cost in current dollars
    • r=discount rate (example, 5%)
    • C0,C1,C2, Cn....costs in future years

Inflation adjusted costs

• Year 1: C1=C0 (1 + i)
• Year 2: C2= C0 (1+i)2

• • 'i' is inflation rate

• The total is,
  • C=C1 + C2 +C3 +.........+Cn

[[Image:Example.jpg]]

Utility data (Cost utility analysis)

Utility analysis: attempts to incorporate into decision analysis effects of interventions on outcomes other than life or death and value or preferences of individuals or society for these outcomes

• Source
  • Develop measurement scales
    • definition of health states (every possible distinct outcome of the intervention and its alternatives)
    • describe health states (each health states should be described from 5 to 9 separate aspects in terms of function or behavior (level of physical health, emotional health, everyday function in social and role activities, and general perceptions of the well being)
  • Obtain rating (rater or group of raters provide information on preferences on each of the health states)
    • selection of raters: based on health states to be evaluated, representative sample, either from society or from patients with specific diseases should be selected as raters
  • Assign rating
    • standard gamble
    • time trade off
    • direct scaling
  • Use of information provided by the raters to create a numerical scale

Markov model

Health states

Cycle length: time elapsed between successive outcome evaluations

Transition probabilities

Costs

Utilities (Cost utility analysis)

CI (if available, for Monte Carlo analysis)

Developing a model structure

Decision analysis tree

Identification of treatment alternatives and outcomes-1

Identification of treatment alternatives and outcomes-2

Identification of treatment alternatives and outcomes-3

Markov model

Analysis 1 OR 2

Analysis 1

Decision tree

• Obtain an estimate of the probability of the expected outcome of each of the decision alternatives
• Apply the process of folding back and averaging using specialized computer software or spreadsheet programs

Cost

• Determine the net cost of the each of the decision option
• Compare the decision options in relation to the net cost for each decision option

Cost effectiveness

• Average cost effectiveness ratio

Cost of the intervention/measure of effectiveness

• Incremental cost effectiveness ratio

Cost of intervention-cost of alternative/effectiveness of intervention-effectiveness of alternative

Analysis 2

Analysis using Expected value

Evaluate the decision tree 1

Evaluate the decision tree 2

Folding back Regimen A 1

Folding back Regimen A 2

Folding back Regimen A 2

Costs associated the regimens

Sensitivity analysis

Compare the stability of the conclusion of the analysis to assumptions made in the analysis to identify crucial areas of information deficiency and guide further research

• Variable

Robust or sensitvie

One way sensitivity analysis

To identify decision threshold values

One-Way Sensitivity analysis using Gastroenterology visit and endoscopy

One-Way Sensitivity analysis using cost of Regimen D

One-Way sensitivity analysis

Multiple 1-wayanalyses: the "tornado diagram"

Two-way sensitivity analysis

Two-way sensitivity analysis

Results

• Review of Included Data Sources
• Descriptive statistics: distribution of demographics and other variables
• Cost effectiveness analysis results
• Sensitivity analysis results
• Quality-adjusted life expectancies
• Interpretation of Main Results

References

• Michael Drummond, Andrea Manca, Mark Sculpher. Increasing the generalizability of economic evaluations: Recommendations for the design, analysis, and reporting of studies. International Journal of Technology Assessment in Health Care, 21:2 (2005), 165–171
• Diana B. Petitti. Meta-analysis, Decision Analysis, and Cost-effectiveness Analysis: Methods for Quantitative Synthesis in Medicine. Published by Oxford University Press US, 1994