PH 3915 Final
willingness to pay method of valuation
"What are you willing to pay?' -could hinder technical innovation in health care, for "denial of coverage would have a large adverse financial effect on a company and might make it more averse to the risk of investing in potential technical advances." -assumes an ideal income distribution, which doesn't actually exist; people have DIFFERENT incomes
standard gamble
- "gold standard" because it involves risk - consistant with what actually happens; provides true "utilities" - (Chance Board) Each subject is offered 2 alternatives Alternative 1: no risk - stay in your current health, e.g. arthritis Alternative 2: involves risk (%) - Tx with 2 possible outcomes Perfect health or Death - The percent chance of success is decreased until the subject is indifferent or unable to choose between the two (the lowest probability of perfect health the patient is comfortable with) - The probability at this indifference is the utility recorded (i.e. chance of living a normal life after treatment
issues to address before applying values from clinical trials in economic evaluation studies
-Randomized clinical trials have high internal validity but may have low external validity. -They often do not actually take a random sample from a target population because of strict inclusion criteria. -RCTs are often conducted in tertiary settings by experts instead of community settings, also preventing a truly random sample. -A RCT will not include a long time horizon that accounts for long-term effects
problems with ICERs when working with patient-level data from a clinical trial
1. ICER cannot be computed if the incremental effectiveness is 0 (because then the denominator is zero). 2. 95% confidence interval are harder to compute and need nonparametric bootstrapping, which are computationally intensive 3.ICER cannot be used in a regression analysis as a dependent variable because it's a ratio, and its direction of change is not conceptually consistent
extended dominance
1. List all cost in increasing order 2. Calculate ICERs 3. If ICERs are not in increasing order, then you have extended dominance set of all possible mixed therapeutic strategies that dominate a single strategy by having higher effectiveness and lower cost; rules out interventions that have an ICER > that of a more effective intervention
The discount rate currently used by analysts in most OECD (high income) countries is
3-5%; normally 3%
non-rivalrous
A characteristic of some goods where the consumption of the good by one person does not reduce consumption by someone else; it is one of the two characteristics of public goods.
non-excludable
A characteristics of some goods where it is impossible/impractical to exclude someone from using a good, because it is not possible to charge a price. It is one of the characteristics of public goods. examples-common resources like sidewalks, clean water, biodiversity AND public goods like national defense, street lights, public sanitation
pure public good
A good or service that , to a high degree, is both nonrivalrous and nonexcludable ex. air we breathe, national defense
indirect costs and benefits should be included in the numerator of the C/E ratio
According to the 2016 guidelines of the U.S. Panel on Cost-effectiveness in Health and Medicine, indirect costs and benefits should be included in
major differences between the CUA and CEA approaches to evaluation
CUA "quality adjusts" a CEA, so preferences are incorporated. CEA also cannot compare a broad set of interventions while CUA can, because it uses a generic outcome.
appropriate sources for defining the ranges for sensitivity analyses.
Confidence interval around mean values of the effect measure from the primary study Empirical evidence from other studies Current practice in the CEA literature, e.g. discount rates Decision-maker judgments, e.g. the range for overhead costs
reasons for market failure in the healthcare sector
Considerable diagnostic uncertainty exists when a provider evaluates a patient The demand for healthcare services is predominately determined and induced by the providers Lack of a single payer payment system in the U.S.
cost utility analysis
Cost-effectiveness or cost-benefit analysis in which quality-adjusted life years (QALYs) are used as a measure of health benefits. same as CEA but outcome measured is increase in utility of a target population due to health outcomes achieved (QALYs for current/prospective patient) Important difference from CEA: generic outcome so you can compare between different clinical outcomes
EBM vs. HTA
EBM is primarily concerned with improving decisions at the patient level while Health Technology Assessment is concerned about healthcare decisions at the aggregate or societal level. HTA includes economic evaluation whereas EBM typically does not. "Health technology assessment is the primary activity that considers 'is it worth it?' although as we point out, EBM should also address the more limited question 'is it worth it to the patient?' taking into account the costs to the patient.
What method do you use to estimate the annualized cost of facilities?
Estimate the cost of capital items already in the budget. Estimate the building's replacement cost, factor in depreciation and opportunity cost.
In a CEA, issues of internal and external validity in the measurement of program effectiveness are independent.
False, you sacrifice one for the other-e.g. the ideal environment for efficacy can make it difficult to apply the program to another population.
characteristics of the human capital approach to conducting a CBA
It places monetary weights on healthy time using market rates, and the value of program is assessed in terms of present value of future earnings It does not produce CBA estimates that correspond to the welfare or losses or gains used in efficiency criteria It provides data on the productivity & earnings impacts of the interventions that can be useful to policy makers does NOT measure implicit values like reduction in risk
Markov model vs. decision tree
Markov-way of iterating through decision trees; agents move between different states and then you run through the tree from that state; better for a longer time span, must adjust for person dying at the beginning of the year
the cost-effectiveness acceptability curve is a function of
NMB
advantages of NMB to ICERs in economic evaluations of life-saving programs
NMB is a linear expression, the variance can be easily designed, can statistically present CEA uncertainty stat convenience to calculate sample size for trial-based economic evaluation
net present value rule
NPV>0=ACCEPT NPB<0=REJECT benefits-costs discounted to present
Is a RCT classified as a full evaluation according to Drummond?
No, a full Econ evaluation must explicitly consider relative consequences of the alternatives and compare them w/ relative costs
Do you have to annualize costs when doing cost benefit analysis?
No, you can just use rent.
alternative outcome levels to consider for an Econ evaluation
Outputs-immediate goods and services (e.g. healthcare workers trained) Effects-immediate effects on target population: knowledge, attitude, behavior (e.g. people understand how the disease is spread and take proper precautions) Impacts-long-term impact on target population-health, SES (e.g. ebola infection rate declines and deaths from ebola decline)
What are the limitations of using average per diem hospital cost to estimate the cost of hospital stays for diabetic patients in an economic evaluation?
Per diem costs are broad, general averages that miss specifics of individual procedures. To cost based on this, figure out the "hotel cost" per diem, take out everything else, and then incorporate the specific costs for procedures.
research and development and other first copy costs for evaluation of a new health intervention
R&D are costs incurred in developing the first copy of an item, independent of the units provided once the first copy is produced. Inclusion/exclusion decision depends on the analysis perspective (private firm vs. societal) and when the decision is made (before or after investment in development). You should differentiate these costs.
cost base
The criterion used as a basis for allocation of overhead-for example, patient days, labor hours, square feet of space in the department
How does the distribution of income in society affect public expenditure analyses of life saving programs?
The human capital approach looks at lost earnings to value life so people who would suffer with a particular disease that are wealthier are "worth more" and the programs would favor them. The willingness to pay method of valuation is a broader way of assessing losing life years. It is hypothetical because it asks "What would you be willing to pay?" instead of "What are you willing and able to pay?" but still skews toward wealthier people. INDIVIDUAL PREFS ARE NOT STANDARDIZED
Effects should be discounted in cost-effectiveness studies.
True, effects should be discounted.
advantages and disadvantages of disease-specific HRQL instruments
VAS-easiest to use but DOES NOT include uncertainty TTO-easier than standard gamble but harder than VAS, also does not include uncertainty Standard gamble-includes uncertainty but NOT EFFECTIVE w/ temporary states
reasons to prefer WTP to human capital when conducting a CBA
WTP: revealed pref method for computing values; reflects value individuals place on health and life; effort to derive social pref regarding public policy and implicitly assess the burden of pain, suffering, and reductions in quality of life; can measure implicit values like reduction in risk
What do CEAs ask?
What is the relative (incremental) cost/unit of effect (outcome) between 2+ strategies with a single common effect?
when do you use the net present value rule?
When conducting a CBA of public health program alternatives, if the recommendations using different decision criteria are in conflict
What can be done to address market price issues in public expenditure analyses?
You can account for it being wrong-e.g. adjust the price to what it would be in a competitive market. Government interventions are: do nothing (acknowledge that intervention could do more damage than good, private orgs might be better equipped to intervene) intervene (info flow, taxation, p/q regulation, acting as insurers/payers for vulnerable, universal insurance, equity via subsidies, medical training and tech dev subsidies)
cost pool
a grouping of individual indirect cost items to be allocated (e.g. housekeeping costs)
visual analog scale
a pain rating scale using a straight line; the left end of the line represents no pain, the right end represents the worst pain, and patients mark the place on the line that best represents the severity of their pain different health states RELATIVE TO EACH OTHER; more important than the actual scale
market failure
a situation in which a market left on its own fails to allocate resources efficiently caused by: externalities (overproduction/underproduction); public goods (consumed or financed collectively) in healthcare, lack of competitive conditions what can the government do? Intervene (improve info flow, enforce competition; taxation; price regulation; act as insurers/payers for the vulnerable; universal insurance; equity via subsidies, etc.) or do nothing.
cost-effectiveness analysis
a type of evaluation research that compares program costs with actual program outcomes 2+ programs; outcomes measured could be: health inputs achieved (e.g. immunizations given, # prenatal visits to physician; # ppl reached w/ health promotion message); clinical indicators achieved (e.g. reduction in cholesterol; reduction in BP); other health outcomes achieved (# side effects prevented; # cases prevented; # cases cured; # lives saved)
incremental effect
additional cost/benefit one service or program imposes over another The change in total results (such as revenue, expenses, or income) under a new condition in comparison with some given or known condition
determinants of price elasticity of demand for cigarettes
age of consumer, peer group behavior, income level of the consumer, cigarette company marketing
human capital approach
aka traditional CBA method places a monetary value on loss of health as the lost value of economic productivity due to ill health, disability, or premature mortality. For vulnerable populations like the elderly and patients with chronic illness, the incremental benefits are typically small. This encourages insurers not to cover as many life-saving programs for them and strikes fear in this population about the use of cost data.
When making resource allocation decisions using the statistical lives perspective
appropriate if we had a limited budget and wanted to maximize the # of lives or # of years of living saved once an individual patient is identified, intervention can become an ethical issue and treatment for an individual can result in a disproportionate use of resources for the identified individual typically used for population health planning resource allocation recommendations can directly conflict with identified lives perspective
interventions w/ CEA results in the southwest quadrant to the right of the 45 degree line
are undertaken WITH objection or controversy
BIA
assumes that there is a new intervention that is cost-effective, with an ICER compared to traditional care that is within the WTP. The recommendation has been made to implement the intervention so it focuses on conducting an economic assessment to predict the financial consequences of adopting the health care intervention within a specific health care setting given resource constraints main question is the program's affordability.
Cost and effects should be discounted
at the same rate because people are NOT INDIFFERENT to when costs and effects occur
market-determined prices
basic value structure for CBA and CEA of public expenditure programs
past articles that did not discount benefits or costs of alternative programs/interventions being evaluated
bias eval results in favor of programs w/ short-run payoffs implicitly assume people were indifferent between receiving a consumption item today vs. a future time implicitly assume that there was no availability of productive resource growth (e.g. no opportunity cost of capital) did not implicitly assume a specific magnitude of the discount rate
for cost utility analysis, utilities require
both choice and uncertainty
for cost utility analysis, preferences include
both utilities and values (depends on response method-SG uses utilities (uncertainty), TTO and VAS use values (certainty))
The choice of a discount rate in a low income country
can be a reflection of that society's preferences w/ respect to the intergenerational tradeoff
low discount rate
clinical care targeting children public health programs reducing the incidence of childhood infectious diseases
High discount rate
clinical care targeting the elderly
social opportunity cost of capital
conceptual reason to discount future benefits and costs to their present value
reason to argue for a higher discount rate
concern for current well-being of people NOW living in SERIOUS POVERTY
CEAC curve
cost: x-axis probability of cost-effectiveness: y-axis
Nonspuriousness
criterion for establishing a causal relation between a program/treatment and an outcome; when a relationship between two variables is not caused by variation in a third variable.
Empirical Association
criterion for establishing a causal relation between a program/treatment and outcome; variation in one variable is related to variation in another variable aka correlation between two variables
time order
criterion for establishing a causal relation between two variables; cause must occur before its effect aka independent before dependent variable
cost worksheet
crucial in micro cost estimation. You can develop a detailed description and production function of the program; list, describe, categorize, and quantify all program ingredients; determine market unit values of each resource; prospectively collect cost data (time and other resources); and develop accordingly
What data is used in the net benefit framework?
data on health benefit AND cost
Square nodes in a decision tree
decisions
the preferred method for handling uncertainty in economic evaluation
depends on source of the uncertainty
dominance
describes an intervention that is more effective and less costly than the alternative; thus choosing this alternative would IMPROVE health outcomes and REDUCE health care costs
Lower discount rate
discount future less favor long-term benefits e.g. public health programs
higher discount rate
discount future more favor now e.g. care for the elderly
Willingness to pay for CBA does/does not incorporate productivity estimations for lives saved by healthcare interventions
does not
What are the main problems with the CEAC?
does not show net benefit differences between alternatives for each threshold and it can be hard to determine how much of the uncertainty is unacceptable
for the human capital approach to a CBA in many lower-income countries
don't: -include an increase in property taxes necessary to pay for the intervention -include the reduction in unemployment benefit payments in each year due to the program's success do: -include a volunteer in our program who left a position as a volunteer for another service program so he could help in ours -include an expected increase in the labor productivity of workers who lives are saved by our intervention program in the program's expected benefits
in the cost-effectiveness diagram, less money is ? and less outcome is ?
down is less money, left is less outcome
criteria for establishing causation between a program/treatment and outcome
empirical associations, time ordering, difference in differences, nonspuriousness
time trade-off
exchanging time in poor health for briefer time in good health.
What are justifications for investments in activities in the public sector? What are not?
existence of externalities, public goods, private market imperfections-JUSTIFICATION low internal rate of return-NOT
In low income countries, discount rates tend to be
higher
major threats to internal validity of RCTs
history, selection, statistical regression, attrition
direct allocation vs. step-down allocation
ignores interactions of overhead departments. Allocation denominator is the sum of allocation basis for all 'final departments'. Step-down allocation of overhead allows for one-way partial adjustment for interactions. The denominator is the sum of remaining departments in the step-down sequence.
effects of an economic evaluation
immediate effects on target population: knowledge, attitude, behavior (e.g. people understand how disease is spread, take proper precautions)
outputs of an economic evaluation
immediate goods and services (e.g. healthcare workers trained)
areas of agreement by analysts for cost-effectiveness evaluations
importance of societal viewpoint discounting in principal
the "win-win" quadrant in the CEA diagram
improvement in costs and outcomes for the intervention southeast quadrant of CEA diagram
human capital approach is often used
in cost of illness estimations
threats to external validity in a study design
inadequate specification and artificiality
major threats to external validity of RCTs
inadequate specification, reactive effects (artificiality, Hawthorne effects, growth and decay)
Effects of healthcare and/ or public health interventions may
increase/reduce cost incurred by society increase health status of individuals
strong positive income elasticity of demand for improved health
influences the size of discount rate chosen for health effects
You cannot do CUA with
intermediate outcomes, only final outcome health measures
the cost of identifying and recruiting the target population to the interventions
is included in economic evaluations of community cancer screening promotion programs
important considerations for selection of health state utility measurement in an economic evaluation
is the state temporary? how will you ask the patient? does the patient understand the different health states accurately?
If using differential discount rates
it is likely that the discount rate for health outcomes would be lower than the discount rate for costs
What is the need for BIA and what is the role of BIA in economic evaluation?
it is used in conjunction with CEA, it evaluates the affordability from a payer perspective also takes into account the actual number of people affected by an intervention
What are legitimate criticisms of the human capital approach to CBA? What is not?
legit criticisms-omits distributional considerations; not based on consumer preference (utility theory); omits intangibles not legit-omits productivity criticisms
impacts of an economic evaluation
long-term impacts on target population (health, SES) e.g. ebola infection rate declines, deaths from ebola decline
What is sensitivity analysis?
looks at how changes in the assumptions of an economic model affect predictions allows us to see how changes in one variable affect the outcome must identify the uncertain parameters, specify the level of uncertainty, determine the sensitivity analysis' form
preference
measurement using different scales/systems that tell us how we weight things to create a specific quality of life outcome; between health states or the uncertainty of the standard gamble
transfer payment
money transfers that involve costs in administration but are not measured by the amounts themselves because they don't reflect resource consumption; ex. workers' compensation payments cost the government, present a gain to the patient, and are not a cost or gain to society; reason that study perspective is so important
How do cost-effectiveness acceptability curves address the limitations of probabilistic sensitivity analyses presented using scatter plots?
more intuitive to look at, can see extent to which cost-effectiveness for an alternative is uncertain at a given threshold, can consider which alternative could be "second best" in terms of cost-effective probability
for cost utility analysis, values require
neither choice nor uncertainty
Does CBA allow you to compare projects across diverse areas, e.g. tb vs. diabetes interventions?
no, cost utility analysis does this
main limitation of the step down allocation method
not all units are easy to quantify so it is always an approximation
potential issues w/ using market prices in public expenditure analyses
not socially optimal prices due to market failure (b/c of lack of competitive conditions, public goods, externalities)
difference in differences
one of the three criteria for judging proof of causality between a program or treatment and outcome - The mathematical way to describe an interaction of two independent variables. - Interaction = a difference in differences = the effect of one independent variable depends on the level of the other independent variable
A CUA for a health program intervention is not appropriate if
only intermediate outcomes are available extra cost is not cost-effective effectiveness data shows that one intervention dominates
What perspective is BIA best performed from?
payer perspective due to its purpose of determining the program's affordability
formula for price elasticity of demand
percent change in quantity demanded/percent change in price
parametric uncertainty is addressed by
performing sensitivity analysis over a range of values for costs, utilities, and probabilities
incremental net benefit regression vs. ICER
possible to adjust for disease severity and other variables, run regression diagnostic procedures, and explore whether treatment has different effects for different patient groups
find the price elasticity of demand
price of a commodity increases from $2.00 to $2.40, and the quantity purchased falls from 1,000 to 600 units. percent change in quantity/percent change in price -400/1000=0.4 0.4/2=0.2 -0.4/0.2=-2
Oval/circular nodes in a decision tree indicate
probabilities-must sum to 1
strongest research design for most clinical studies
randomized controlled trials
risk attitude
relative values of adverse consequences of risk averse, neutral, or loving averse is unnatural-specific to the test you give
incremental cost-effectiveness ratio
represents the additional cost relative to the additional net-effectiveness
What costs do we incorporate in BIA?
short-term (1-3 yrs); NOT lifetime
conceptual reasons for discounting
social time rate of preference social opportunity cost of capital
probabilities at each sub-tree branch must
sum to one
cost-effectiveness acceptability curve
tells us to what extent an alternative's cost-effectiveness is uncertain for a given threshold, which other alternative could be "second best" in terms of cost-effectiveness probability main issues are that it does not show net benefit differences between alternatives for each threshold and it can be hard to determine how much of the uncertainty is unacceptable
Triangle nodes are _____
terminal
Where do we enter pay-off values in a decision tree?
terminal nodes of each sub tree
efficacy
the ability to produce a desired or intended result; clinical effectiveness of an intervention CAN IT WORK?
If the ICER < threshold
the alternative generates more health for a given amount of resource than the health care likely to be given up
marginal cost
the cost of producing one more unit of a good (e.g. provide service/intervention to one more)
transactions cost
the costs in time and other resources that parties incur in the process of agreeing to and carrying out an exchange of goods or services; incurred in overcoming market imperfections examples include: legal fees, communication charges, informational cost, transportation, etc.
Practical clinical trials should involve
the decision makers in the selection of the program alternatives to be evaluated and the design of the trial -the hypothesis and study design are developed specifically to answer questions faced by decision makers-can determine using a compromise between internal and external validity if an intervention works under normal circumstances
net monetary benefit
the difference in effects between two options being evaluated is rescaled into monetary value using cost-effectiveness threshold as a value for each unit of effect; difference in costs between the options is subtracted from this value -net health benefit in money -can be explained as the monetary value on the effectiveness based on the threshold for the ICER or willingness to pay per unit effectiveness; positive is good and negative is unacceptable; when ICER=threshold NMB is 0
WTP or human capital approach provide data on the productivity and earnings impacts of interventions, which can be useful to policy makers
the human capital approach
BIA cannot be performed from
the patient's perspective
When should the costs of volunteer workers NOT be included in an economic evaluation?
the study is being conducted from a private payer perspective workers were getting some intrinsic value out of the activity
history
threat to RCT internal validity how something else may have changed over time that affected the study's outcome
selection
threat to RCT internal validity the bias associated with differences in the kinds of people in the experimental and comparison groups
testing
threat to RCT internal validity what human subjects learn from their experience when part of it is the device used to measure their performance e.g. follow up questionnaires may unintentionally cue individuals to get screened for breast cancer
inadequate specification
threat to external validity of RCT lack of important program details, thus hindering application in other contexts
artificiality
threat to external validity of RCT program appropriately measures efficacy but is so highly structured and controlled that it is not practical for real-world conditions (effectiveness)
growth and decay
threat to external validity of the RCT experimental studies may not be long enough to account for effects of alternative treatments
hawthorne effect
threat to external validity of the RCT results are influenced by experimental aspects beyond the planned intervention-a change in a subject's behavior caused simply by the awareness of being studied
attrition
threat to internal validity of RCT group profiles at the end of the experiment may be different from those at the beginning due to different dropout patterns
regression to the mean
threat to internal validity of RCT subjects are chosen to represent an "extreme version" of a characteristic so there is a chance that, when performance is observed once again, it will be closer to the mean
discount rate
time preference for health based on rate at which government or health care funders can invest (borrow/save) and whether the threshold will grow; shows the relative value of current vs. future health in a third world country, we prefer a higher discount rate typically b/c the intervention is more important now (e.g. people are starving)
net health benefit
total health gained at a certain price point; what you are saving (QALYs) by comparing to the threshold
parameter uncertainty
uncertainty in the estimates of the inputs or parameters of the type of decision models discussed in Chapter 9 (e.g. decision trees, Markov models, probabilities and expected values) sometimes called "second-order uncertainty"; in regression analogous concept is standard error of the estimate can characterize with one-way or multi-way sensitivity analyses (you can see what happens to model outputs when parameter values are varied)
opportunity cost
value of the benefits achievable in some other program that has been foregone by committing the resources in question to the first program; hard to assess
cost-effectiveness threshold
value represents the amount of health care resources expected to displace one QALY elsewhere in the health care system; essential to determine if something is/is not cost-effective; standard in a health care system can also be used to establish the amount of additional health care resource to generate the same health benefits elsewhere
net monetary benefit characteristics
variance can be easily defined is NOT a ratio statistic can be used to compute cost-effectiveness acceptability curves if +, intervention is cost-effective
price elasticity of demand for cigarettes for adults
vicinity of -0.5
price elasticity of demand for cigarettes for teenagers
vicinity of -1
probabilistic sensitivity analysis
way of creating a distribution for each parameter and sampling from this to account for uncertainty. It is thought to be more thorough than deterministic sensitivity analysis.
Quality-adjusted life years
way of standardizing health outcomes decide on a utility measure for quality; generic; can use to come up w/ equivalencies
When can nonparametric bootstrapping be used to get confidence intervals for ICERs?
when bootstraps fall in SE and NW quadrants
NMB curve intersects the x-axis
where NMB=0 and the threshold value on the x-axis is the ICER
sources of methods uncertainty in cost-effectiveness analyses
whether to discount the measure of effect
long-run profit maximization objective
why tobacco companies sometimes not raise prices enough to cover the amount of an increase in the sales tax on cigarettes
When should be the costs of volunteer workers be included in an economic evaluation?
workers would have been otherwise productively employed the study is being conducted from a societal perspective
cost-effectiveness diagram-area in the southwest quadrant to the right of the 45 degree line
worse outcomes for intervention; proportionally greater reduction in costs
should CEA account for related and unrelated healthcare costs during the additional life-years produced by an intervention?
yes, the new recommendations by the 2016 guidelines of the U.S. Panel on Cost-effectiveness in Health and Medicine "also suggest inclusion of future costs (ie, that cost-effectiveness analyses account for related or unrelated health care costs that occur during the additional life-years produced by an intervention)."