HSCI 310 Quiz 2

Measuring and Valuing Costs: Numerator and denominator

Cost-effectiveness Thresholds

(k) should represent what is expected to be given up (to afford the more costly alternative)

• need this to compare to QALY to know if intervention is affordable

• no explicit threshold but over-time we can see a pattern/range (general guide)

Nominal

no order and no numeral values (ex. categories such as gender, race, or hair color)

• used for labeling variables without quantitative value.

Ordinal

order but still no numeral values (ex. rating scales, rankings)

Cardinal

numerical values with measurable differences, allowing for mathematical operations (ex. height, weight or QALY). + ORDER

Consequences of k too high or low

include inefficient resource allocation and potential neglect of cost-effective interventions.

HIGH: more treatments (more likely available ~ approved) BUT consequence is we would pay more for the same treatments cause no cap

• little reason to maintain a threshold that is higher than the balance of evidence suggests, despite the obvious stakeholder interest in doing so.

LOW: less treatments (which reduce health outcome b/c people miss out on cost-effective treatments — some health care will be restricted when it need not be.

Do threshold change over time?

Yes, thresholds can evolve based on new evidence, economic conditions, and healthcare priorities and of course inflation — BUT THERE IS A GENERAL TRENDLINE

Thresholds: Cost-effectiveness and affordability

rarely both are achieved simultaneously in healthcare decision-making. Often, an intervention may be cost-effective but still unaffordable, leading to difficult choices about resource allocation. (SYSTEM IS MESSED UP)

Other aspects of value for thresholds

would be:

• increase in utility (well-being) but not improving health

• effects treatment has on everyone inolved (caregivers, family, individual)

• productivity (societal, get back to work)

Cost-consequence Analysis (CCA)

is a form of economic analysis that compares the costs and consequences of different health interventions, presenting the outcomes in a disaggregated format.

• “CEAs may present an array of output measures alongside cost and leave it to the decision-makers to form their own view of the relative importance of these… useful approach, even if the analyst then goes on to value the outcomes relative to one another”

3 fundamental assumptions in economic evaluation

people using this approach ignore three key problems:

1) Congruence of objectives: analysis assumes decision makers value what the analysis measures (too complex ex. QALY)

• CCA more closely meet the needs of decision makers (simple but thats the problem academically)

2) Funnelling outcomes into one (QALY): reducing complex outcomes to a single number loses info

• CCA avoids this, keeps outcomes separate

3) Meaningfulness to decision makers: complex techniques are hard for decision makers to use

Does Economist = decision maker

NO! just shows results via analysis (academic)

Multiple Alternatives

• ‘either/or’ decisions: incremental ratio, cost-effectiveness plane

  • cant have more than 2 things compare

• >2 alternatives (which are more realistic): trial-based analyses and decision models, combinations, sequences or options

  • for combinations of drugs etc.

Which treatment is best? (multiple alternatives)

• don’t need to test them all to each other

• C b/c more QALY b/c lower cost use of incremental analysis for cost-effectiveness comparison

Should D be considered cost-effective? (λ = $50K/QALY)

No. D is dominated by C

(D will never be cost-effective b/c it needs to be best under threshold)

Extended Dominance

if a combination of two options is less costly and more effective than others

What does this show

B and D are most north west = they are dominated

Whats the best?

• cancel out C b/c its the most north west from the slope

  • C = extendedly dominated

• keep excluding one until you only get one treatment left —> B lowest cost per QALY

Guidelines and Appraisal

for evaluating health interventions and making informed decisions based on cost-effectiveness and effectiveness.

• includes: critical appraisal, reporting standards (CHEERS II), Guidelines for conduct (CDA, NICE) ~ reference case, rationale?

CHEERS II report standards

its not a quality assessment (study could be terrible)

Reference case

idea is to ensure some form of standardization (for similar methods)

• not saying that this is the best method but it is all for consistency

Why do we need guidelines for studies

• economic evaluations have two main problems: poor data quality and reporting standards & weak evidence supporting methods

• guidlelines for conducting evaluations are unlikely to help develop better methods (may even discourage it)

• guidelines for reporting evaluations are more helpful because they encourage providing detailed information without stifling methodological innovation.

What needs to be considered when measuring/valuing costs: community based exercise program, life-saving drug treatment for infants, intervention for a condition that affects working age adults

must consider:

• time period of treatments we are interested in cost (longterm vs shortterm)

• people affected, staff (wage), productivity

  • how often drugs taken, side effects

• location (rent), technology (buy), cost of drug storage

shouldn’t considered: production cost/resources, shouldn’t fund private sector or include costs not directly related to the intervention.

Economical Perspectives

such as individual, societal, and healthcare provider viewpoints, which influence the assessment of costs and benefits in health economics.

ex. traditions of welfare economics, societal perspective, international guidelines

Societal perspective

estimate all costs (and benefits) associated with the change in practice

• consisders productivity and overall impact on the community, including indirect costs such as lost productivity and long-term health outcomes.

• welfare interlinked

(no best perspective)

Perspective: CDA 2006 vs CDA 2017

Perspective: NICE

Perspective: HIQA

CDA: publicly funded health care system (only public healthcare costs) —> adding additional analyses (decision makers to decide to try to be more broad)

NICE: cost-effectiveness, NHS + social services costs only (dont add patient, carer, or productivity)

HIQA: cost-utility analysis, publicly-funded and societal care system in Ireland (HSE)

One value issue

to simplify factors to make 1 value is possible but very hard to do economically

Evidence-based Health Economics

If you think measuring health outcomes is hard, you might hope costing is easier. But even there, progress has been limited, and the results are still unsatisfactory

• due to variability in methods and data availability, evidence-based health economics aims to provide a rigorous framework for evaluating the economic impact of health interventions using systematic reviews and robust data analysis.

Analysis of Cost Data in RCTs

Even though all 45 trials collected cost data per patient, most failed to report it properly. They didn't show how much the costs varied (80% failed), often didn't run proper statistical tests (44% failed), and over a third (64%) drew conclusions that were not supported by the results they actually presented.

• This highlights the importance of rigorous methodologies in costing data collection within randomized controlled trials to ensure accurate and reliable economic evaluations.

what does this show

that economic evaluations has very complicated outcomes and costs

• highlighted above are the 3 stage processes

Cost Estimation: 3-stage process

1) identification of the range of resource us —- cost-intervention

• what resources are needed, staff, people, technology, etc.

2) determine the quantity of use — how often using resources

3) valuation of health services — now what’s the new cost (total)

• how much will be allocated to that

Cost Estimation Equation

Total cost for patient i = Σj (Resourceij x Unit costj)

• Resourceij is the amount of resource j used by patient i

• Unit costj is the unit cost for resource j

What does this acupuncture paper show

• gives a sense of resources that they know were input

• recognize how they split home visit GP and office GP

  • separate due to cost difference

what does this show

more generalized categories

• healthcare persepctive only top 3

• societal perspective shows (broader) a bigger gap

(note that national guidelines make it clear which perspective to use)

what does this show

very different productivity perspective

• differs from first chunk at the top where they are sorta similar

Perspectives: many alternatives

MoH (healthcare), other ministries, ‘government’, employer (only US), agency, patient (rare — considers out of pocket costs), …, etc

Perspectives: Practical considerations ~ Broad vs Narrow

prudent/wise to be broad (but at what cost?)

• making complex better but is it possible to maintain

Related Costs: how long for? disease-specific or generic?

How long is a big decision — do you expect there to be difference amongst the groups over time?

This relates to whether costs are evaluated in a disease-specific context or generically across populations

• ex. how many times you see GP for a specific illness versus general health issues

Disease-specific vs Generic

← (upper circle) specific: tailored to a particular disease; generic: applicable to a broader population. goof for when you don’t want much variation

← (bottom circle) generic: applicable to multiple diseases or conditions, considers possible side effects (for comparison), good to be aware of

• could make a cloudy estimate — complicates for respondent

Related Costs: non-health care costs?

nothing is straightforward in economic evaluation

Related Costs: possible dilemma

extend life vs money

• not cost-effective if keeping them alive only causes more costs (expensive treatment)

Related Costs: learning effects

impact of prior experience or education on future health-related decisions and costs. This can lead to variations in treatment adherence and utilization of healthcare resources.

Related Costs: market conditions, patents, etc.

Factors influencing healthcare pricing and accessibility

What is the time period for tracking: hospital, other health care agencies, patient/family/friends

important to consider for follow up and evaluation of healthcare outcomes, including the timing of interventions and resource utilization.

hospital = short-term

other health care agencies = medium-term

patient/family/friends = long-term.

what does this show

• allows us to track cost and differences between groups

• helps decide when to stop study

  • shouldn’t stop at 12 months

  • 2 years okay to stop but could still miss info (but change has slowed)

Issue with: Unrelated Future Costs

These are costs that arise due to unintended consequences of healthcare interventions, often resulting in increased spending on long-term care for chronic conditions.

• ex. the costs of caring for people with dementia or kidney failure after preventing their deaths from a heart attack

YES: more accurate but complicated, interest in incremental cost and treatment, probability estimates possible outcomes down the line (inaccurate to assume they don’t have further issues)

NO: disadvantage, life saving treatments, focuses on cost and QALY; intervention becomes more expensive could make the intervention less likely to be cost-effective despite initial benefits

NICE CDA: Unrelated Future Costs

yes add

Where do we draw a line: Unrelated Future Costs

Generations? just gets messy

• ex. The full opportunity cost of life extension includes all the resource consumed by the survivor less the resources produced by the survivor

Methods of Data Collection (2 main)

Decision modelling vs Trial-based evaluations

(Note: can’t just add economists at the end of study — should be involved with design and study collection)

Decision modelling

synthesize data from various sources (primary studies, reviews, opinion, etc.)

Trial-based evaluations (adv and dis)

Can mix and match for study:

self-report methods

• Adv: easily gather individual info

• Dis: recall of past history can be inaccurate

medical charts / case notes

• Adv: could carry some useful info

• Dis: multiple doctors involved so notes can vary (no consistency and might not even be helpful)

routinely-collected data

• Adv: could be very accurate

• Dis: but not always, could be wrong time frame or information might not be relevant to the study

Methods of Data Collection: Issues to consider

• Under-researched area’s are less glamorous and hard to study but equally important

• protocol-driven costs

  • are people using more healthcare b/c they ate in a study

• length of recall period (+biases)

  • inaccurate timeline of treatment that was big in person life

• primary consistency

  • lots of decisions to be made (many differences) making it hard to compare

what is costing

how precise do you need the costs to be in terms of capturing all relevant expenditures and evaluating financial outcomes? (time consuming)

• two extremes but there are layers between

Precise/Accurate: top-down costing

Makes use of routinely available data. This approach utilizes existing financial records and health data to estimate costs effectively. — more crude/simple

Precise/Accurate: Bottom-up (or micro) costing

Involves detailed tracking of all costs associated with healthcare services at the individual level, including direct expenses like personnel time and indirect costs such as overhead. This method provides a comprehensive view of resource utilization. — EXTREMELY PRECISE

• primary data collection, you yourself is responsible for all the data

• complex and expensive

• gold standard? — no b/c need to balance accuracy

How Precise/Accurate: top vs bottom costing

shows there are groups in between

Productivity changes: higher healthcare costs for individuals to lose fewer work days

challenges whether healthcare system perspective (ignores productivity b/c not cost effective) or societal perspective (which include individual productivity) should be used

• Absence from work activities (loss to society) — short-term and future

• Components: consumptions activities (individ) + benefit for other (taxation)

  • wider considerations:

    • fluidity in the labour market

    • opportunity cost (emplyment foccus?)

    • programs for people unlikely to be in work activities (seniors dont contribute)

    • tacitly accepting different priorities?

• W and WE don’t apply

Productivity changes: Approaches to estimation

The methods through which productivity changes due to healthcare interventions are assessed. This may involve analyzing the impact of health improvements on work absenteeism, efficiency, and overall economic output.

• Human capital approach, friction cost method (presenteeism, equity, double counting (us panel) ← national guidelines that say productivity is already considered in utility so don’t do again WRONG )

Human capital approach (productivity)

The Human Capital Approach is an economic theory that evaluates productivity by estimating the potential output increase based on health improvements and education investments in the workforce. It focuses on how health status affects labor productivity and economic growth.

• absence matters 0 or 100 — CRUDE

friction cost method (productivity)

A technique for estimating productivity losses due to health-related absences by calculating the costs associated with decreased work output, taking into account factors like presenteeism and opportunity costs. — REALISTIC

• absence doesn’t matter if replacible (4 days of loss vs all days gone)

presenteeism (productivity changes)

The phenomenon where employees are physically present at work but perform at reduced efficiency due to health issues, leading to lost productivity. This can negatively impact overall workplace performance and economic output.

Discounting

a process that enables a comparison between the value of a commodity consumed in the future with the value of that same commodity consumed now

• things in the future what cost vs benefit in todays society

why is discounting important?

• many treatments have different time profiles for costs and benefits, e.g.,screening vs heart transplant

• allowance needs to be made for differential timing of costs and benefits

what is this trying to show

• varies in cost-benefit but end price is the same

  • “what is todays’s cost of spending # per yr vs spending upfront”

• C already has discount value ~ already today’s value

Key issues with discounting

• applies to costs and effects

• relevant to any study longer than one year

• differential discount rates? (less of a issue now)

• keenly debated issue

• national guidelines

(don’t worry about where the discount value comes from focus on what the guideline want me to do)

CDA: discounting 2006 → 2017 vs NICE: discounting 2022

2006: discount the costs and health outcomes that occur beyond one year to present values at the (real) rate of 5% per year

2017: discount costs and outcomes that occur beyond one year should be discounted to present values at a rate of 1.5% per year

2022: NICE recommends a discount rate of 3.5% for both cost and health benefits.

Choice of study design

“... comparative analysis of alternative courses of action in terms of both their costs and consequences”

• consider circumstance/context: within condition to across conditions (QALY needed ~ generic/standardized)

M&V of Outcomes: single outcome

• intermediate outcomes aren’t ideal

• monetary outcomes? (difficult in practice)

• life years? (pretty crude)

A) better life

B) survival

(which is valued more $$$ or QoL or live longer)

CDA expectations for study guidelines/questions

must be meaningful, worded well

• can’t be vague and not specific

• any difference should be included

(UK values consistency aims for similarities — WE LIKE)

M&V of Outcomes: QALY (quality-adjusted life years)

• mortality and morbidity

• quality-weighting survival

what does this describe?

cardinal scale: shows how predictable QALY can be w/o program/control group

• A: purely quality associated

• B: survival associated

Perfect health not a good term b/c it is misleading and not proper use FULL HEALTH

interval vs ratio

differ in what zero means:

• interval interprets 0 as a point on a scale, while ratio interprets 0 as the absence of the quantity being measured.

life years (LY)

Survival is relatively straightforward (dead or alive)

• trial-based evaluations

• decision models

quality-adjusted (QA)

• quality of life — hard to measure

• health-related quality of life

• preference-based measures

What is: Health status/HRQoL/subjective health

• asking people to self-report (physical, mental, pain)

• which aspects should be included: descriptions

Subjective wellbeing

• happiness, life satisfaction

• reflection that there’s more to being happy than your level of health

does QALY go beyond health economists

yes, it incorporates quality of life and other factors, considering both the quantity and quality of life in evaluating health interventions.

Condition-specific QoL measures

• narrow assessment of a dimension (or dimensions) of health for a condition

• no cross-condition comparison

• generally, not helpful for QALYs

Examples: Asthma Quality of Life Questionnaire (AQLQ), Dermatology Life Quality Index (DLQI), Arthritis Impact Measurement Scale (AIMS), Dementia Quality of Life (DEMQoL)

Generic outcome measures

• broader assessment of QoL/HRQoL — anyone can complete it than generic

• enables cross-condition comparison

• can be insensitive (being less sensitive is to be expected)

Examples: Nottingham Health Profile (NHP), SF-36v2, Quality of Wellbeing Scale (QWB), EQ-5D-3L, EQ-5D-5L, Health Utilities Index (HUI) Mark 3, etc.

Profile vs Index Scores

Profile scores assess multiple dimensions of health e.g., SF-36v2

• captures multiple dimensions/domains — not helpful for QALYs

Index (single value) scores condense this information into a single value, allowing for easier comparison. — summarize ECONOMIST LIKE (easy descion making)

• trade-off: comprehensive coverage and the ability to derive a single value

• can incorporate preferences

Estimating QALYs: Index Scores

Direct measurement of values: ‘direct’ elicitation from an individual, own health or hypothetical scenarios

• This method utilizes various instruments to evaluate health outcomes and convert them into quality-adjusted life years (QALYs), often allowing for more generalized comparisons in healthcare.

Indirect measurement: ‘off the shelf’ scores, separation of description and valuation to derive a single value representing quality of life, typically using a standardized approach.

Direct Preference Measurement

• Methods to derive index scores: standard gamble & time trade-off, scales (rating, category, VAS), discrete choice experiments

• Beyond the QALY: willingness-to-pay (CBA), wellbeing and capabilities?

Desirable Properties

Sacrifice = determining the strength of preference should involve some kind of sacrifice — give something up

Uncertainty = health outcomes & decision-making are uncertain in the real world

what is standard gamble and what does this show

standard gamble = method for measuring preferences by presenting a choice between a certain outcome and a risky one

• tends to produce higher values (compared to TTO) b/c easier to avoid immediate death

• The probability (P) is varied until the individual is indifferent between the two alternatives (where P = P* ). The preference score or health STATE i is P*.

  • props and visual aids

  • temporary health states

  • states WTD

• there is a point of indifference that would indicate end of exercise

Standard Gamble: Adv vs Dis

Adv: strongest theoretical foundation, both desirable properties

Dis: unrealistic uncertain outcome, artificial risk, cognitively challenging

Gold standard? — no

what is a Time Trade-Off

A method used to assess preferences for health outcomes by determining how much time individuals are willing to trade off to avoid a less desirable health state. — chronic state is preferred to being dead

• Here, x is varied until the respondent is indifferent between the alternatives. The preference score for health state i is, hi = x/t.

• there is a point of indifference that would indicate end of exercise (can’t go in past so no negative)

Time Trade-off: Adv vs Dis

Adv: decision context: quality and quantity, no probability

Dis: desirable properties not clearly established (sacrifice but no uncertainty), unrealistic

Standard Gamble vs Time Trade-off

TTO is simpler’ alternative to SG

• theoretically weaker

• avoids probabilities

• ariants (lead time TTO, props, scripts, etc.)

Scales (rating, category, VAS)

Ranking & placing on a scale e.g., consider five health states

• need to focus on intervals (many variants, thermometer is common)

• Respondents could be asked to place health states on the line. The intervals between the states should reflect the strength of the respondent’s preferences.

  • extremes 0→100 no meaningful between health states; intervals represent preferences

  • dead is a state no events allowed (so not death)

what is this?

Example of a visual analogue scale (VAS) — Used to value a respondent’s own health

This version is part of the EQ-5D instrument, developed by the EuroQol Group

Scales (rating, category, VAS): Adv vs Dis

Adv: theoretical base (although not in economics), simple to administer and complete

Dis: choiceless context, meaningful intervals, less accepted, no sacrifice or uncertainty

what are Direct and Indirect (list)

Direct: TTO, SG, VAS

Indirect: MAUI, EQ-5D, HUI

what are Indirect (MAUI)

Standardized questionnaires that measure health-related quality of life using algorithms, often incorporating utility values to estimate health status. — RECOMMENDED

(EuroQol Group (EQ-5D-3L & EQ-5D-5L), Health Utilities Index, SF-6D & SF-6Dv2; VR-12)

• Scoring is based on societal preferences (political)

Components of Indirect (MAUI)

descriptive component + valuation component that together provide a comprehensive assessment of health status.

Can you compare Indirect and Direct approaches

No reason why you can’t explore indirect and direct approaches. But… costly and fundamentally different sets of values

• Endless discussion in the literature about the role and appropriateness of generic preference-based HRQoL measures

Indirect (MAUI): EQ-5D-3L & EQ-5D-5L

• 53 (243) and 55 (3125) health states

• primarily TTO methods

• profile value: 1 2 3 2 1,

• index value: 0.755, most important #

EQ 5D 3L VS 5D 5L

The EQ-5D-3L and EQ-5D-5L are standardized instruments for measuring health-related quality of life (HRQoL), where the 5L version offers greater sensitivity with five levels of health states compared to three in the 3L version, providing more detailed assessments.

• 3 fails survey design b/c it asks two things in one question, very definitive choices that aren’t clear and can lead to interpretation

EQ-5D-5L (predictions): Which dimensions are most valued to you?

• dimensions are not valued equally (kind of the whole point)

  • unweighted scores (min. = 5; max. = 25) used in some contexts but nothing to do with QALY estimation

• utility decrements associated with the levels of severity

  • 11111 = 1.000* full health

    • Canada is the only one that doesn’t go to 1.0

  • dead = 0.000

  • 55555 = ????? (negative values possible — worse than death)

EQ-5D-5L (predictions): Which dimensions do you think are most valued by society?

Pain and discomfort on average most important, and Usual Activities is ranked last on average

• Canada is the outlier to mobility with having it ranked lower — due to accessibility and help provided here

• Uganda ranks anxiety and depression the lowest — due to culture and interpretation of mental health

what does this show?

this is the overall summary

• single resource that shows despite the various surveys used there are key values amongst them

• lots of variability in values -0.59 QALY vs 0.34 lowest (some don’t rank to dead and some get worse than dead)

Indirect (MAUI): SF-6D & SF-6Dv2

• derived from longer questionnaires

• SG & DCE methodology

• more health states (compared to EQ-5D)