Decision making in situations of uncertainty

1
(No Transcript)
2
(No Transcript)
3
Decision making in situations of uncertainty
4

• 1. 65 yr male presents with UGI bleed. He had
  angioplasty and stenting two months ago. Should
  he continue antiplatelet agents?
• 2. 85 yr male presents with acute inferior MI to
  hospital in Dawson Creek. Should he be given
  thrombolysis?
• 3. At what level should anti retroviral therapy
  be initiated for HIV infection? CD 200? CD 350?
  CD 500?

5
IN EACH SCENARIO THERE IS

• 1. Uncertainty
• 2. Tension between Risks and Benefits
  a trade off
• 3. A specific decision is identified.

6
MEDICAL DECISION ANALYSIS
7
Patient

• 76 yr female with chronic renal failure. She
  presented with volume overload/CHF and had to
  start hemodialysis.
• While getting her hemodialysis she had several
  episodes of atrial fibrillation.
• Hx of DM, Hypertension

8
Anti Coagulate?
9
RISK OF STROKE IN AF

• RR 2.4 in men and 3.0 in women.

10
CHADS SCORE

• CHF 1 point
• Hypertension 1 point
• Agegt75 1 point
• DM 1 point
• Stroke or TIA history 2 points

11

• CHAD score 4
• Her risk of stroke without anticoagulation 6
• Her risk of stroke WITH anticoagulation 2.2
• With ASA somewhere in the middle

12
WHat is her risk of Major Bleeding?

• How do we determine it?

13
HemorR2hages risk index

• Hepatic or renal disease
• ethanol abuse
• malignancy
• older age
• decreased platelets/or asa
• prior bleed (2 points)

14

• hypertension
• anemia
• genetic factors
• excessive fall risk
• stroke
• with greater than 5 points the risk of major
  bleeding was greater than 12.3

15
4 points 10.4 chance
16
Hemodialysis?

• Traditionally anticoagulants felt to be a
  contraindication to hemodialysis.
• Now - not as much of an issue

17

• 2. Risk of Anticoagulation in dialysis patients
• Division of Nephrology, Queen's University,
  Kingston, Ontario, Canada.
• Studies of full-intensity anticoagulation and the
  1 randomized controlled trial of low-intensity
  anticoagulation showed major bleeding episode
  rates ranging from 0.1 to 0.54 events/patient-year
  of warfarin exposure. These rates are
  approximately twice as high as those of HD
  patients receiving either no warfarin or
  subcutaneous heparin. LIMITATIONS This review is
  based largely on data from observational studies
  in which bleeding rates may be confounded by
  comorbidity. Relatively small sample sizes may
  provide imprecise estimates of rates.
• CONCLUSION Low- and full-intensity
  anticoagulation use in HD patients is associated
  with a significant bleeding risk, which has to be
  balanced against any potential benefit of
  therapy. This has to be considered carefully when
  prescribing warfarin to HD patients.

18

• Increase risk of Intracranial Hemorrage with
  anticoagulation?
• Use of Long-Term Anticoagulation is Associated
  With Traumatic Intracranial Hemorrhage and
  Subsequent Mortality in Elderly Patients
  Hospitalized After Falls Analysis of the New
  York State Administrative Database.
• Controlling for age, gender, and comorbidity,
  patients on LTA were 50 more likely to sustain a
  traumatic ICH after a fall (odds ratio 1.50 95
  confidence interval, 1.231.81 p lt 0.0001).
  Furthermore, among patients who sustained an ICH,
  mortality was 1.57-fold greater in patients on
  LTA (odds ratio 1.57 95 confidence interval,
  1.022.45 p 0.04). Conclusions These data
  indicate that use of LTA is independently
  associated with traumatic ICH and subsequent
  mortality in elderly patients hospitalized after
  a fall.
•  
• J Trauma. 200763519 524.

19
benefits of warfarin in This patient ?

• Decreased risk of stroke.

20
risks

• Major Bleeding including ICH

21
What are the strategic options?
22

• 1. Do nothing
• 2. Anticoagulate with warfarin
• 3. Take a middle ground
• 4. Rythm control?

23
Decision factors - patient

• Risk taking behaviour
• Specific aversions
• eg. Aversion to bleed, or to stroke
• Quality of Life factors
• weekly blood tests
• Worry.

24
Decision factors - Physician

• Biases
• Framing

25
Who makes the decision?

• An Authority?
• The Attending physician?
• The Patient Family in consultation?

26
Is there a Best decision?

• This is a situation of uncertainty. We are faced
  with these decisions in circumstances like these
  every day in medicine.
• Is there a way of determining the best
  decision?
• What kinds of approaches can be used?

27
MEDICAL DECISION ANALYSIS

• A quantitative evaluation of outcomes that result
  from a set of choices.
• A formal modeling of the process clinicians go
  through every day.
• Decision Analysis (DA) makes the process
  explicit and amenable to examination, discussion
  and challenge.

28
Problems appropriate for DA

• Focus on a specific decision that must be made.
• A trade off
• Uncertainty

29
USES of DA

• SPECIFIC Exploring strategies for Specific
  patient problems
• GENERAL Exploring Strategies for a class of
  patients.

30
Steps in the Process

• Frame the question and model the problem by
  creating a Decision Tree.
• Estimate the relevant probabilities
• Estimate the value of the outcomes
• Analyze the tree by the calculating expected
  values.
• Test the Models assumptions with Sensitivity
  Analysis.

31
A Simple Decision Tree
32
(No Transcript)
33
ASSUMPTIONS

• Probability of winning
• Black jack - .20
• Slot Machine - .01
• Winnings
• Black Jack - 200
• Slot Machine - 1000

34
Tension

• Black Jack - Greater chance of success but the
  reward is less
• Slot Machine - Less chance of winning but a
  bigger pot.

35
(No Transcript)
36
Expected Value

• P(event) X Value (1-P(event))X value

37
(No Transcript)
38
Some basis principles

• The tree must have balance.
• Only two branches after each chance node
• No embedded decisions
• Symmetry
• Order doesnt matter

39
WHY DA?

• Modeling of the problem sometimes helps to
  identify the best strategy
• Symmetry of the model encourages us to identify
  new strategies we might not have considered
• Challenging our assumptions.
• Identifying essential probabilities we need to
  know.
• Directing research in clinical useful ways.

40

• Helpful technique for explaining a complex
  situation to a family
• Involving the patient and family directly in the
  decision making process through UTILITY ANALYSIS

41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
Sources of data
45

46
THROMBOLYSIS IN THE ELDERLY
47
Base line Probabilities

48
Decision tree for thrombolysis in the elderly

49
ANalysis of decision tree

50
Sensitivity analysis

51
Two Way sensitivity analysis

52
(No Transcript)
53
How do we measure the outcome?
UTILITY ANALYSIS
54
Quality adjusted Life expectancy

• calculate Life expectancy
• identify Utility for the different possible
  outcome states.
• Utility a coefficient for Quality of Life
• QALE LE x Utility

55
DA in the medical literature

• Multiple articles in the major journals here are
  a few over the last few months
• GI Endoscopic surveillance of Barretts
  Esophagus
• Screening for West nile virus in organ
  transplantation

56

• ARTICLE
• Influence of Alternative Thresholds for
  Initiating HIV Treatment on Quality-Adjusted Life
  Expectancy A Decision Model
• R. Scott Braithwaite, MD, MSc Mark S. Roberts,
  MD, MPP Chung Chou H. Chang, PhD Matthew
  Bidwell Goetz, MD Cynthia L. Gibert, MD, MSc
  Maria C. Rodriguez-Barradas, MD Steven Shechter,
  PhD Andrew Schaefer, PhD Kimberly Nucifora, MS
  Robert Koppenhaver, MS and Amy C. Justice, MD,
  PhD
• Background The optimal threshold for initiating
  HIV treatment is unclear.
• Objective To compare different thresholds for
  initiating HIV treatment.
• Design A validated computer simulation was used
  to weigh important harms from earlier initiation
  of antiretroviral therapy (toxicity, side
  effects, and resistance accumulation) against
  important benefits (decreased HIV-related
  mortality).
• Data Sources Veterans Aging Cohort Study (5742
  HIV-infected patients and 11 484 matched
  uninfected controls) and published reports.
• Target Population Individuals with newly
  diagnosed chronic HIV infection and varying viral
  loads (10 000, 30 000, 100 000, and 300 000
  copies/mL) and ages (30, 40, and 50 years).
• Time Horizon Unlimited.
• Perspective Societal.
• Intervention Alternative thresholds for
  initiating antiretroviral therapy (CD4 counts of
  200, 350, and 500 cells/mm3).
• Outcome Measures Life-years and quality-adjusted
  life-years (QALYs).
• Conclusion This simulation suggests that earlier
  initiation of combination antiretroviral therapy
  is often favored compared with current
  recommendations.
• Increases in life expectancy were small, and
  the analysis did not consider costs.
•  

57
Cost Effectiveness

• Need to compare the cost of implementing a new
  strategy to the gold standard
• Measure of the cost per quality adjusted life
  year achieved

58
Usefulness of the Decision tree

• A simple way to model a problem
• identify new strategies
• Identify essential information to focus a
  literature search
• remove physician framing biases
• Excellent communication tool

59
Other advantages of DA

• Gives a way to compare like values (ie life
  expectancy)
• Introducing the concept of QUALITY OF LIFE to the
  equation.
• Patient input to the decision is essential
• DA gives a way to analyze more general questions
  that have not yet been studied with large
  clinical trials.
• a rational approach to Cost Effectiveness

60
(No Transcript)
61
(No Transcript)
62
(No Transcript)
63
(No Transcript)