Decision analysis Flashcards
Decision Analysis
An analytical method for systematically comparing different decision options.
Provides an orderly, analytical approach to assist decision maker in identifying the preferred course of action from among competing alternatives.
When to use decision analysis?
When uncertainty about clinical strategy is evident,
When differences in benefits and risk are evident
No head to head trials
Decision analysis characteristics
Explicit- forces you to structure the decision you face as well as identify the consequences of the possible outcomes.
Quantitative- Forces you to assign number to probability estimates and outcome valuations.
Applications of decision analysis
Choosing among treatments
Choosing between treatment and no treatment
Preventing diseases (vaccinations)
Preventing complications
Decision tree
Built to be reflective of clinical reality.
Focuses on most important components, not “real world”
Steps in decision analysis
- ) Identify decision
- ) Specify alternatives
- ) Draw tree
- ) specify branch, outcomes, probabilities
- ) Calculations
- ) Conduct a sensitivity analysis
Identifying specifc decision
Time period of analysis?
Objective of study?
Perspective?
Example- Decision on whether to add new abx A to an institutional formulary
Specify alternatives
Ideally, compare the most effective treatments Comparison: standard treatment vs. new treatment Intervention vs. no intervention More than two competing options.
Example- what should abx A compare to?
What is a choice represented by?
Square
Represents conscious decision
What is a chance represented by?
Circle
Represents probabilities
It is a chance node
What is terminal (final outcome) represented by?
Triangle.
Represents dollars, quality adjusted life years, survival
For each option of decision analysis, what do you need?
Probability of occurrence and consequences of occurrence.
Interpretation of a probability of outcome D of 0.126.
If 100 people went through the model, about 13 people (12.6) would go down pathway D.
Average cost
The cost multiplied by the probability for each terminal node and then added for each option provides the average cost.
Incremental Cost Effectiveness Ratio (ICER)
Change in costs/change in outcomes
Example- if 500: 500 more dollars for each additional success with abx a.
Incremental net benefit
Alternative to ICER
An estimate for health benefits (outcomes) is added into the incremental analysis (lambda, max willingness to pay, estimated value)
If INB >0, then the intervention is considered cost-effective.
What does it mean if the ICER is positive?
One medication is more effective and more costly.
Using a lambda value makes it easier to choose. If the INB value is negative, you lose value.
Sensitivity Analysis
Measures the uncertainty in a model
Does not compensate for poor assumptions
The larger the variability, induced by changing a parameter, the more sensitive the outcome is
Overview of Markov modeling
Some diseases have complex outcomes that require longer f/u periods (>1 year)
Patients my transition between health states
Each f/u interval is called a cycle
Health states in Markov
Must be mutually exclusive. Cannot have the 2 states at the same time. Typically one occurs from complications of another.
Steps in markov modeling
- ) Choose health states
- ) Determine transition states
- ) Choose how long and how many cycles
- ) Estimate probabilities associated with moving in and out of health states
- ) Estimate costs and outcomes.
Absorbing state in markov model
Once a pt is in an absorbing state, he/she cannot transition out of it.
Death
Diabetes
You need an absorbing state to finish the model.
Disadvantages of Markov Modeling
Complexity and transparency
The Markovian assumption- knowing only the present state of health is sufficient to project the trajectory of future states.
Drug policy is maintained through
Formulary system
Formulary
Preferred list of pharmaceuticals to be used in a health system that reflects current clinical judgement.
Open formulary
Can bring in products not on list.
PBM: Payer will provide coverage for formulary and non-formulary drugs
Closed formulary
Health system: Limited/no capability to bring in non-approved products
PBM: Non-formulary drugs not reimbursed by the payer.
Formulary maintenance
P and T committee
Voting members- physicians, pharmacists, nurses, health system administrators, respiratory therapists, social workers.
Subcommittees may report through P&T
P&T Activities
Formulary drug review FDA drug safety alerts Shortage formulary review Medication use evaluations Drug cost savings initiatives Monitor drug use reports Pharm rep code of conduct enforcement
Formulary categories
Formulary
Restricted formulary
Non-formulary
Non-formulary non-stock
Formulary category: Formulary
Readily available for use
Provider should order and expect for patient to receive in 1-2 hours
Formulary category: Restricted formulary
Readily available if certain criteria are met
Drugs may be restricted to disease states, lab values, hospital floors, providers, etc.
Formulary category: Non-formulary
Not readily available
May be able to order for individual patients if clinical need is demonstrated
Formulary category: Non-formulary non-stock
Will not be available for use under any circumstances.
Formulary maintenance decision
Safety data, efficacy data, cost data
Formulary system support tools
Policy
Guidelines
restrictions
Therapeutic interchanges
Policies
Dictate how drugs will be ordered, dispensed, administered, monitored, and assessed
Should not deviate from policy
Can help TJH medication management standards
Examples: look-alike sound-alike, vaccine storage, handling, dofetilide use
Guidelines
Support evidence-based use of drugs or for disease states
May support restrictions
May guide administration practices
Providers can deviate
Restrictions
Criteria that must be met in order to use drug- clinical, provider-based, location-based, financial
If outside that criteria is considered non-formulary and may require approval
Therapeutic Interchanges
Replace a prescribed drug to another chemically different, clinically similar drug
Evidence-based interchange to equivalent doses of preferred agent
Streamline inventory
Example- PPI
Formulary assessment process
Request- from providers or pharmacists. Want to know about drug name, clinical reasoning, evidence, estimated volume of use, and conflicts of interest
Analysis- Performed by non-biased third party (Drug policy specialist). Assess efficacy, safety, cost data. Recommendation
P&T decision- presentation of data, questions from committee members, vote.
Formulary analysis
Safety data, cost data, efficacy data, other
Safety, what are we looking at?
Contraindications Warnings/precautions AE Post-marketing data Look-alike sound-alike Packaging issues
Efficacy, what are we looking at?
Published information
Reviews/expert opinions
Guidelines
Cost, what are we looking at?
Pharmacoeconomic studies if available
Cost per unit, per dose, relative to alternatives
Volume- based analysis to estimate annual impact
Medication access: third party coverage/assistance programs
Presentation of drug
Pharmacist and other non-biased third party presents drug assessment to P&T or subcommittee
Implementation of formulary
Electronic medical record changes Dose limitations, order sets, product selection Educate Procurement Storage
Order sets/protocols
Help drive how drugs are used in formulary system
Example: Order set for amphotericin B liposomal that includes pre-hydration and medication
Procurement
Traditional wholesale
Direct purchase
Limited distribution
White bagging
Can only get through outside specialty pharmacy, sent to hospital
Brown bagging
Specialty drug dispensed by outside pharmacy and sent to patients home
Clear bagging
Specialty drug acquired/dispensed by institutions own specialty pharmacy and stored prior to use
Quality of life in clinical medicine
Represents the functional effect of an illness and its consequent therapy upon a patient, as perceived by the patient
Measuring HRQoL
Utility measures
Health status measures
Utility measures
Also referred to as “preference-based measures” or QALY measures
Estimates the utility or “value” that individuals assign to different health states
Number between 0-1 X length of time in each health state to represent the combined impact on morbidity and mortality in linear fashion
Also used to calculate cost-utility
Standard gamble and time trade off
Health status measures
A patients estimation of their own health at a point in time
Measured based off of patients viewpoint
Measured in multiple domains; often many scores
HRQoL Domains
Physical functioning
Social and Role Functioning
Mental Health
General Health perceptions
Physical functioning
Limitations or disability
Physical abilities
Days in bed
Bodily pain
Social and Role functioning
Interpersonal contacts- frequency of contacts with family/friends
Social resources- social media
Are you able to continue doing the things you want to do in life?
Mental health
Anxiety/depression
Psychological well-being
Behavioral/emotional control
Cognitive function
General health perceptions
Current self-rating of health
Outlook for future
General health status instruments
General health
Disease specific
Generic HRQoL instruments
Relative to any type of situation
Advantages- readily available, reliability and validity known, comparable across studies
Disadvantages- questions may not be applicable, may not focus on area of interest, may not be sensitive to change
Disease specific QOL measures
Advantages- tailor made, focuses on question of interest, more likely to detect change
Disadvantages- time and effort to design, reliability and validity testing necessary, may not be comprehensive , may not be acceptable to others, cannot cross studies
QOL tools
Karnofsky functional performance scale
EuroQol Instrument
Medical Outcomes Study 36-item Short-form health survey
Karnovsky
Objective measure of functional performance.
Completed by healthcare professional with input by patient.
Each category is subdivided into 3 descriptors, each associated with a score. The higher the score, the better overall functioning.
Categorizes on ability to perform daily activities.
EQ-5D
Non-disease specific instrument for measuring of health and health-related qol in patients
Self completion by pt
Comprises 5 dimensions of health- mobility, self-care, usual activities, pain/discomfort, anxiety and depression
Scale is 0-1
MOS SF-36
36 questions measuring 8 health attributes
Subjective health-realted QOL
Self administered
Great insight into health of patient
SF-36 measurement model
Physical health and mental health broken into 8 total domains
Psychometrics
The science of using standardized tests or scales to evaluate attributes of an individual
Used to translate peoples behavior, feelings, and personal evaluations into quantifiable data
Reliability and validity
Key factors in selecting HRQOL tool
Reliability and validity Responsiveness Acceptability for use Frequency of use Culturally valid
Reliability
Measure is consistent and reproducible
Test-retest reliability
Internal consistency
Intra and inter observer reliability
Validity
Does the questionnaire really measure QOL?
Responsiveness
Do scores accurately reflect over time?
What is most widely used QOL tool?
SF-36
In many languages, culturally tested
What is the use of QOL in clinical trials and practice?
Very important to look at whole patient.
