Economic evaluation in health care and Pharmacoeconomics
How to provide services efficiently? ◦ Maximum output at a given cost, or ◦ Minimum cost to get a given output How to maximize the benefits or effectiveness from services with a limited resource base? ◦ Which investments will address the most pressing health problems and bring the greatest health gains? ◦ In other words, how to get the most value for money?
More specifically on cost-effectiveness analysis
Costs more Least desirableLess Effect More Effect Most desirable Costs less
In simple terms it is a ratio of two informationCost Effectiveness = What you pay What you get
Cost CBA CEAEffect CUA Measured in terms of some utility index, such as DALY
Type of Analysis Costs Consequences Result Cost Identical in all Least cost Money Minimisation respects. alternative. Different magnitude Cost per unit of of a common Cost consequence eg. Money measure eg., LY’s cost per death Effectiveness gained, blood pressure reduction. averted. Single or multiple Cost per unit of effects not necessarily Cost Utility Money common. Valued as consequence eg. “utility” eg. QALY cost per DALY. As for CUA but valued in money. Net $ Cost Benefit Money eg willingness-to- cost: benefit ratio. pay
CEA Hospital procedure + Public health drug agentsProgram evaluation Pharmacoeconomics
How “costly” the program / approach is? How much “ net effect” the program / approach could generate? What is the cost per unit of “effect”? Is this cost per unit of effect less than other comparable programs / approaches?
Step 1: Define the program in terms of possible “interventions”Step 2: Calculate “incremental” cost for each intervention (money value)Step 3: Calculate “net” effects of each intervention (in terms of health outcome)Step 4: Divide (2) by (3) for each intervention. These are CE ratios.Step 5: The intervention with minimum CE ratio is the most cost-effective one.
Each program can be defined in terms of a “Goal” and a few alternative “interventions” to reach the goal. For example, take a Malaria control program. The goal is: Reduce Malaria deaths among target beneficiaries The “possible” interventions are: Spray insecticides, or Clean environment, or Promote bed nets, or All of the above
Step 1 Classify costs into recurrent and non- recurrent (or, development) for each intervention. Only incremental costs.Step 2 Annualize the development cost by dividing by the expected lifespan of the input.Step 3 Add annualized development cost to recurrent cost and beneficiaries’ cost (if available) for each intervention.Step 4 Subtract „negative‟ costs (or, cost savings)
Recurrent cost: Costs incurred on the same item more than once in a year.For example, salary, rent, supplies, medicines, maintenance, etc.Non-recurrent cost: Costs incurred not more than once in a yearFor example, purchase of equipment, vehicle, building, etc.
Annualization of non-recurrent cost:Suppose you buy a vehicle for the programme at the price $5000.The expected life span of the vehicle = 10 yearsThen the annual cost = $5000 ÷ 10 = $500Add $500 (not $5000) as the non-recurrent cost for one year.
Beneficiaries‟ cost:This is the amount of money the beneficiaries (or, users) spent out of their own pockets for getting benefits of the interventionFor example, travel cost, drugs purchased, etc.
Expenditure in ‘00 dollarsInterventions Development Recurrent Total project Total Total cost cost cost cost beneficiaries (annualized) costInsecticide 3.58 99.01 102.59 646.21 748.80Clean environ 9.48 251.65 261.13 734.09 995.22Bed net 20.89 225 245.89 721.10 966.99All components 35.99 238.43 274.42 806.04 1080.46
Measure effectiveness not efficacy ◦ Efficacy = measure of effect under ideal conditions (can it work?) ◦ Effectiveness = effect under „real life‟ conditions (does it work?) ◦ Efficacy does not imply effectiveness Measure (count) in natural physical units ◦ Number of lives/life years saved / deaths averted ◦ Change in cholesterol levels Measure final not intermediate outcomes ◦ Intermediate outcomes reflect change in clinical indicators ◦ Final outcomes reflect change in health status
Step 1 Calculate the number of deaths “before” or “without” intervention for each type of interventions.Step 2 Calculate the number of deaths “after” or “with” intervention for each type of interventionsStep 3 Subtract (2) from (1). This is the net gain or “effect” of the intervention.
Deaths without (before) intervention: How many people died or expected to before intervention.This can be calculated from the pre-intervention or baseline mortality rate. This information is available from baseline survey data.Deaths with (after) intervention: How many people are still dying after the intervention.This can be calculated from end-line survey. Alternatively, it can be estimated on various assumptions
Item Without intervention With intervention (Base)(1) Covered population 10000 5000(2) Number of deaths 30 10(3) Mortality rate (per 3 21000)(4) Difference in rate 1(5) Deaths averted [(4) × (1) ] ÷ 1000 = 5 deaths
Step 1 Divide the “effect” for each intervention by its cost.Step 2 Find the intervention with minimum CE ratio.
Program 1 Program 2A Target population 10000 100B Number reached 1000 50C Coverage (B/A X 100) 10% 50%D Impact (unit effects) 100 40E Effectiveness (D/A X 100) 1% 40%F Drug program cost ($) 10,000 200G Efficiency (F/B) $10 / person $ 4 / personH Cost-effectiveness (F/D) $100 / unit $5 / unit
Alternative dosage of lovastatin in secondary prevention of heart disease (Goldman et al 1991, JAMA 265: 1145-51) Ages 65-74 Daily dose Cost ($bn) Life years Cost/Life year 20 mg. 3.615 348,272 10,400 40 mg. 7.051 477,204 14,800
Variable Spironolactone Eplerenone Relative risk of death due to 75.2% 86.2% heart failure Per patient cost of drug (36 $50.28 $1,230.00 months) Cost of drug per $440.00 $53,000.00 death preventedPitt B et al. The New England Journal Medicine 1999;341(10):709-717 Pitt B et al. The New England Journal Medicine 2003;348(14):1309- 1321
It helps set health priorities by comparing various health interventions on economic grounds It helps the managers use scarce resources in a more efficient way However, it reflects only the economic aspects of an intervention. For final selection of an intervention, other aspects need to be considered
Problem # 1 Program‟s goal is not “to reduce deaths”, but “to reduce death + suffering”?Solution:In this case, “effect” of intervention also includes reduction of morbidity rate. But how will you add reduction in mortality and reduction in morbidity (they are different!)?The solution is to translate the project‟s “output” into some outcome index, such as Disability Adjusted Life Years or DALY. That is:Number of deaths averted = X number of DALYNumber of disability-free years created = Y number of DALYSo, total effect = X + Y
Problem # 2 So far the discussion is based on the assumption that the program is for less than or equal to one year. What if it spreads over multiple years?Solution:In this case, we have to apply “discounting” process. That is, both costs and “effects” need to be discounted to present value.
Step 1 Calculate the “output” of the project per year (e.g. treated cataract cases).Step 2 Translate “output” into Disability Adjusted Life Year (DALY) on the basis of utility values. Multiply with average life expectancy after intervention. This is DALY with intervention.Step 3 Translate “no output” into DALY. Multiply with average life expectancy without intervention. This is DALY without intervention.Step 4 Discount both DALYs and subtract Step 3 from Step 2.
It is a state of health indicated in 0 - 1 range, where 0 = death, 1 = perfect health. Thus utility value closer to zero implies more severity (as perceived by the patient) than the same closer to 1.
Healthy 1Life with menopausal symptom 0.99Watery diarrhea, severe anemia 0.90Infertility 0.80Physical limitation with some 0.67 pain 0.39Deaf, dumb, or, blindDead 0
Multiply the output (number of patients treated) with respective utility values (i.e., utility without problem). This will be DALY with intervention. Multiply the same number assuming that no intervention is done with respective utility values (i.e. utility with problem). This will be DALY without intervention.
Discount DALY (1) with intervention and (2) without intervention at the same rate (3% or 5%). Subtract (2) from (1). This is net effect of the project.
Interventions Cost per DALY ($)Immunization against Polio plus DPT $ 20in high-mortality environmentARI treatment for young children in $ 20high mortality environmentImmunization against Hepatitis-B $ 25 - $ 50Primary and secondary prevention of $ 150cardio-vascular diseaseMedical and surgical management of $ 1000cancers