Measuring and Valuing Health Flashcards
(21 cards)
what is valuation
Aim to assess how much individuals prefer different levels of health based on the 0 (dead) to full health (1) utility scale. Health with very low U values show it has the largest burden on our wellbeing. Useful for identifying trade-offs individuals are willing to make between dimensions
Few different methods to value health
what values do we need to be valued
- A description of what is being valued
- In the context of quality or disability adjusted life years, the method should allow us to value the description on 0 (dead) to 1 (full health) utility scale
- A group of people to do the valuation
what 2 ways Descriptions could be summarised in:
Vignettes – brief paragraphs of what is being valued
• Standardised questionnaires (e.g. Patient Reported Outcome Measure)
where do descriptions to value health come from
Patients (could ask them to value there current state without description)
• Experts e.g. clinicians, carers (usually different from patients)
• Combination of patients, clinicians, carers
what are Vignettes
brief paragraphs of what is being valued
Use information from patients and/or experts to provide description of health
But cost of constructung a specfic one is high, and concerns with a lack of comparability between studies in terms of what is being measured, less human in comparison to other methods
Standardised questionnaires
Patient (sometimes) and expert information used to develop questionnaires. The questionnaire is used to describe different levels of health
examples- SF-36
can be generic or spefic
Name some Valuation methods (health valuation)
Visual analogue scale • Time trade off • Standard gamble • Discrete choice • Person trade off • Willingness to pay but problematic in health
Valuation methods (health valuation)
Valuation methods are designed to provide information on what individuals prefer, by how much and this should be on a cardinal scale
Examples
Visual analogue scale • Time trade off • Standard gamble • Discrete choice • Person trade off • Willingness to pay but problematic in health
Visual analogue scale (VAS)
Respondents asked to indicate how good or bad a health state/description is, in their opinion using a scale usually ranging from 0 to 100.
It could be they value: • Their own health state • A number of other health states where they place health states on the line
For the QALY, we need full health (1) and dead (0) on the scale • Ask respondents to value health states and ‘dead’ • Then recalibrate to reflect where dead is • No choices being made that involve opportunity cost – just ranking
Calculating VAS values
NOTE THE SCALE IS 0-100 -0 BEING DEATH, 100 PERFECT HEALTH
Individual ‘A’ values a health state at 80 and puts ‘Dead’ at 0 Value = 80/100=0.8 If ‘Dead’ is not included then some argue this does not meet requirements for a QALY/DALY scale Note dead can be somewhere else not just 0- individual ‘B’ values the same health state at 80 but puts ‘Dead’ at 20. They are using a smaller part of the scale to represent states better than dead. Value= (80-20)/(100-20)= 0.75 If ‘Dead’ is not at 0 then there can be negative values e.g. If individual A placed a health state at 10 the value for this would be: (10-20)/(100-20) = -0.125
Time trade off (TTO)
Individuals choose between two options. A. A length of life in the health state to be valued, B. A shorter period in full health.
Who should value health better
General public-
veil of ignorance • no strategic behaviour • tax payers • patients may be unable to provide values (elderly, very ill, children)
Patient –
better understanding of condition and their own well-being • Give patients a voice
Experts-
patients may be unable to provide values (elderly, very ill, children) • may have better understanding than public where patients are unable to provide values
Public values tend to be lower than patient values for physical health and reversed for mental health
why is there a diffrence in health value
Valuing different states
• How is it described? Vignette vs. standardised measures compared to what patient experiences
• General population and patients may focus on different aspects
Different measuring sticks – patients assessment of symptom severity may be different from general population.
Adaptation – patients may have adapted either by changing expectations or views of what matters in life, adjusting their activities, or even be in denial. General population unlikely to take this into account especially from simple descriptions in generic measures
Preference-based measures (PBM)
Valuation can be expensive and time consuming – conducting such studies for each cost utility study would not be cost effective • Using different groups in the valuation (general population/patients etc.) can produce different values • Solution: standardised questionnaires which already have utility values
Generic preference-based measures
A questionnaire (sometimes referred to as a health classification system) completed by patients (i.e. PROM)
• Standardised description of health with dimensions and levels to distinguish severity based on intensity or frequency
• A generic preference based measure should be broad enough to be applicable (in theory) to all conditions/populations
The tariff is used to score the measure and represents preferences for different health states on a scale of 0 (dead) to 1 (full health) (can be less than 0)
• Tariff obtained from a representative general population sample using mainly time trade off (TTO) but also standard gamble (SG), visual analogue scale (VAS) and discrete choice experiments (DCE)
Incremental analysis
Cost effectiveness based on incremental cost per unit of outcome i.e. incremental cost effectiveness ratio (ICER)
(Cost new – Cost current)
(Outcome new – outcome current)
Causes of differences
Descriptive System • Dimensions – e.g. visual function not covered by EQ-5D • Severity – range (e.g. floor effect in SF-6D) • Sensitivity – number of levels (e.g. EQ-5D-3L is very crude at upper end)
Valuation • Valuation methods are different • There are also variations in how each method can be applied which can produce different results • Valuation samples
Rationale for choosing GPBM
Advantages • ‘Off-the-shelf’ and cheap • Convenient and comparatively easy to use in clinical trials (all self-complete) • Allow comparison across conditions • Accepted by NICE and other agencies in their ‘reference case’ analyses
Disadvantages • May not be relevant or sensitive to the condition or treatment effects • Different measures – different values
Choosing between preference-based measures
Psychometric performance in the population of interest
• Feasibility – length, complexity, method of administration.
• Reliability – stability over time, between raters, between location.
• Validity - extent to which an instrument measures what it is intended to measure: content validity, face validity, construct validity
• Responsiveness – extent to which an instrument reflects change where it is known to have occurred
Using DALYS in decision making
International decision making
• DALYs offer a quick approach to estimating the impact of interventions for different diseases in different regions of the world based on similar populations
• Of more interest to international decision making bodies – WHO, World Bank who recommend policy or allocate loans across countries
QALYS vs DALYS
QALYS
Description usually a preferencebased measure but can also be a vignette
Time trade off most common valuation method
Sample usually general population
Country-specific tariffs e.g. 27 for EQ5D-3L
Mainly used in high and middle income countries
Can be difficult and expensive to generate country-specific weights where research income is low
DALYS
220 common health states described based on conditions to reflect ‘major functional consequences’
Discrete choice experiment with ‘equivalent approach’
Sample general population representative of different regions
Aimed to be usable across the world
Mainly used in middle and low income countries
Can be difficult to link utility values to specific health states that are used when modelling outcomes
