Clinical Models in Venous Thromboembolism:

1
Clinical Models in Venous Thromboembolism
or

• How to make the most of your history and physical
  exam

Eddy Lang CCFP(EM) CSPQ McGill Emergency
Rounds October 2001
2
Clinical Scenario 1 Mr. Tremblay

• 63 year old male with severe cough x 1 day
• No significant PMHx, non smoker
• Tachypnic able to speak complete sentences
• VS HR 105 BP 140/90 T 38.1?C - rectal O2 Sat 93
  RA
• No preceding URI Sx
• Central chest discomfort initially -now resolved
• Some decreased A/E left base
• CXR small left-sided pleural effusion, minimal
  airspace disease

3
Clinical Scenario 2 Ms. Jones

• 22 year old female
• 2 hour history of chest pressure and SOB
• Appears apprehensive, wearing oxygen starting to
  feel better
• VS 120/90 RR 22/min Sat 100 HR 90 T 36.9
• Smoker, OCPs
• Remote history of panic attacks similar but not
  as severe
• No associated symptoms
• Unremarkable physical

4
Test and Treatment Thresholds in the Diagnostic
Process
No further testing necessary
Treatment commences
5
Survey
1. Would you order a ventilation/perfusion scan
or helical CT in this patient (I don't need to
know which)? 2. Would you order a D-Dimer test
on this patient (assume non-elisa)? 3. Would
you use the D-Dimer result to determine whether
you would order any imaging i.e. helical CT or
V/Q?
6
Survey Results
7
Survey Results

• Mr. Tremblay
• 5/7 get V/Q
• 2/7 get D-dimer
• 3/7 incorporate D-dimer in decision to image
• Ms Jones
• 2/7 get V/Q
• 5/7 get D-dimer
• 4/7 incorporate D-dimer in decision to image

8
Clinical Question

• In patients who present with symptoms and
  signs suggestive of PE, can elements of the
  clinical examination in combination with simple
  tests allow me to determine which patients can be
  safely discharged without imaging procedures?

9
Clinical Question

• In patients who present with a syndrome
  suggestive of PE, can a clinical prediction rule
  allow me to determine which patients need further
  work-up?

10
Educational Objectives

• Review the rationale for the development of
  clinical prediction rules in venous
  thromboembolism (VTE)
• Conduct structured critical appraisal of the best
  prediction rule / clinical model research in VTE
• Explore issues related to test selection in
  patients suspected of VTE

11
Top 10 List
12
Top Ten Reasons to Dislike Clinical Prediction
Rules

• 6. I dont practice cookbook medicine
• 7. I refuse to do any math during my shift
  especially if I need a calculator
• 8. Im way too busy to use these things
• 9. My gestalt clinical judgement is better
  than any prediction rule could be
• 10. They are just too damned complicated to use

13
Top Ten Reasons to Dislike Clinical Prediction
Rules

• 1. Everyone knows the H and P stinks, lets
  just get an MRI
• 2. Developed by nerdy academics who havent
  examined a patient in 30 years
• 3. I dont believe in fortune telling
• 4. Except for Christian, everything from Ottawa
  annoys me!
• 5. They are a government plot designed to make
  us cut costs

14
Finding the evidence
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25

• Use of a Clinical Model for Safe Management of
  Patients with Suspected Pulmonary Embolism
• Wells, Ginsberg, Anderson et al.
• Annals of Internal Medicine - 1998

26

• Inclusion Criteria
• Consecutive inpatients and outpatients with
  suspected pulmonary embolism whose symptoms had
  lasted less than 30 days were potentially
  eligible
• 5 centers, 16 physicians
• Exclusion Criteria
• Duration of Sx gt 3 days
• Anticoag gt 72hrs.
• Survival lt 3 months
• Suspected upper extremity DVT

Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
27
Probability Model
? 0.86
Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
28
Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
29
1885 eligible patients
Study Flow

• Prolonged
• anticoagulation
• expected survival
• lt 3 mos.

484 ineligible
1401 eligible patients

• 147 declined
• to consent
• 13 lost to
• follow-up

150 lost
1239 evaluable patients
734 low PTP 3.4 PE
403 mod. PTP 28 PE
102 high PTP 78 PE
30
Rates of Pulmonary Embolism According to Pretest
Probability of Pulmonary Embolism and Results of
Ventilation-Perfusion Lung Scanning
Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
31
Rates of venous thromboembolic events during the
3-month follow-up

• normal perfusion scans and normal initial
  ultrasonograms
• 2 of 332 0.6 95 CI, 0.3 to 3.0
• non-high-probability ventilation-perfusion
  scans, low or moderate pretest probability, and
  normal serial ultrasonograms
• 3 of 665 0.5 95 CI, 0.1 to 1.3
  

Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
32
Cause of Death According to Whether Pulmonary
Embolism Was Initially Diagnosed
Wells, Ginsberg, Anderson et al. Annals of
Internal Medicine - 1998
33

• Excluding pulmonary embolism at the bedside
  without diagnostic imaging management of
  patients with suspected pulmonary embolism
  presenting to the emergency department by using a
  simple clinical model and D-dimer.
• Wells, Anderson, Rodger, Stiell et al.
• Annals of Int. Med. 2001

34

• Inclusion Criteria
• Consecutive emergency department patients
  (adults) with suspected pulmonary embolism whose
  symptoms had lasted less than 30 days were
  potentially eligible
• 4 centers, 43 physicians
• Exclusion Criteria
• Suspected upper extremity DVT
• No Sx within 72 hrs
• Anticoag gt 24 hrs.
• Expected survival lt 3 mos.
• Contraindication to contrast
• Pregnancy
• Geographic inaccessibility

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
35

• Interventions
• Application of a clinical model
• SimpliRED whole-blood agglutination D-dimer
• Primary Outcome
• Proportion of patients with VTE during 3-month
  follow-up
• Methodology
• Intention to treat analysis
• Upper range 95 CI 1 VTE rate
• Sample size 930

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
36
Clinical Model

• Hemoptysis - 1.0 pt.
• Malignancy 1.0 pt.
• HR gt 100 - 1.5 pts.
• Immobilization (? 3 consec. Days) or surgery
  within 4 weeks - 1.5 pts.
• Previous DVT/PE - 1.5 pts.
• Clinical signs and Sx of DVT - 3.0 pts.
• PE as likely or more likely than alternate Dx -
  3.0 pts.

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
37
Clinical ModelScoring

• Low PTP
• lt 2.0 points
• Moderate PTP
• ? 2.0 but lt 6.0
• High PTP
• ? 6.0

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
38
Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
39
(No Transcript)
40
Clinical ModelPerformance

• Low PTP
• 527 pts. (57) 7 VTE events (1.3)
• Moderate PTP
• 339 pts. (36) 55 VTE events (16.2)
• High PTP
• 64 pts. (7) 24 VTE events (40.6)

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
41
Follow-up Data on 849 patients with VTE ruled out

• 17 suspected events 6 low PTP, 9 mod PTP, 2 high
  PTP
• VTE confirmed in 5 (0.6 95 CI 0.2 to 1.4)
• Low PTP
• 4 suspected PE, 2 DVT 1 PE confirmed (day 16)
• Mod PTP
• 4 suspected PE, 5 DVT 3 DVT confirmed
• High PTP
• 2 suspected PE 1 PE confirmed

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
42
Compliance with Algorithm

• 92 patients had fewer tests than called for
• 4/5 confirmed events occurred in this group
• Among 81 patients initially diagnosed with PE 7
  patients labeled as a result of extra testing
• 2/7 from low PTP group

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
43
Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
44
Do we believe this?
45
(No Transcript)
46
(No Transcript)
47
Methodologic Standards for Development of a
Clinical Prediction Rule The Researchers
Perspective

• Is there a need for the decision rule?
• Was the rule derived according to methodologic
  standards?
• Has the rule been prospectively validated and
  refined?
• Has the rule been successfully implemented into
  clinical practice?
• Would use of the rule be cost effective?
• How would the rule be disseminated and
  implemented?

Stiell IG and Wells GA APRIL 1999 334
ANNALS OF EMERGENCY MEDICINE
48
(No Transcript)
49
Methodologic Standards for Validation of a
Clinical Prediction Rule The Clinicians
perspective

• Were the patients chosen in an unbiased fashion
  and do they represent a wide spectrum of the
  severity of the disease?
• Was there a blinded assessment of the criterion
  standard for all patients?
• Was there an explicit and accurate interpretation
  of the predictor variables and the actual rule
  without knowledge of the outcome?
• Was there 100 follow-up of the patients
  enrolled?

Users Guide 2000 AMA press
50
Were the patients chosen in an unbiased fashion
and do they represent a wide spectrum of the
severity of the disease?

• consecutive symptomatic patients
• 86 ultimately diagnosed with PE (9.5)
• Rule in rate of 17 in patients undergoing
  imaging

51
Was there a blinded assessment of the criterion
standard for all patients?

• V/Q scans were interpreted by Nuclear Medicine
  physicians who had no knowledge of the clinical
  model or D-dimer
• Pulmonary angiography and venography were
  evaluated by the same criteria
• A committee blinded to all patient outcomes
  adjudicated suspected outcome events

52
Was there an explicit and accurate interpretation
of the predictor variables and the actual rule
without knowledge of the outcome?

• Model and d-Dimer interpreted prior to imaging
  or F/U
• No specific reporting of Kappa in this study

53
Was there 100 follow-up of the patients enrolled?

• Sixteen patients lost to follow-up because of
  relocation
• No sensitivity analysis performed

54
Algorithm Performance

• 47 of patients did not require imaging
• 7 serial ultrasonography
• 1.1 pulmonary angiography
• Overall sensitivity 99.4 (95 CI 98.6 - 99.8)

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
55
Applicability

• Will the reproducibility of the test result and
  the interpretation be satisfactory in my clinical
  setting?
• Are the results applicable to the patients in my
  practice?
• Will the results change my management strategy?
• Will patients be better off as a result of this
  rule?

Wells, Anderson, Rodger, Stiell et al.
Annals of Internal Medicine - 2001
56
(No Transcript)
57
(No Transcript)
58
Levels of Efficacy in Diagnostic Test Evaluation
How many different possible elements might be
incorporated into the decision to select a given
diagnostic test?
59
Levels of Efficacy in Diagnostic Test Evaluation

• Technical Efficacy
• Diagnostic Accuracy Efficacy
• Diagnostic Thinking Efficacy
• Therapeutic Efficacy
• Patient Outcome Efficacy
• Societal Efficacy

60
Technical Efficacy

• Feasibility and acceptability
• Operator dependence/training
• Analytic sensitivity
• Interferences and cross-reactivity of biochemical
  tests
• Measurement inaccuracy (systematic measurement
  error)
• Measurement imprecision (random measurement
  error)

61
Diagnostic Accuracy Efficacy

• Sensitivity and specificity
• Predictive value
• Likelihood ratios
• Measures of area under the ROC curve

62
Diagnostic Thinking Efficacy

• Percentage of cases in which the final diagnosis
  changed after testing
• Difference in clinicians subjectively estimated
  diagnostic probabilities before and after
  receipt of test info
• Certainty or confidence in a diagnosis
• Percentage of cases in a series in which the test
  was judged helpful to making the diagnosis
• Cost/change in clinical diagnosis

63
Therapeutic Efficacy

• Percentage of times that management changed based
  on test information
• Percentage of times that another test was avoided
  because of information from the test under
  investigation
• Total cost of diagnostic strategies, cost/patient
  tested, or cost/change in management decision

64
Patient Outcome Efficacy

• Symptom severity
• Functional outcome
• Patient utility assessment
• Expected value of test information in QUALYs
• Morbidity avoided by testing or not testing
• Mortality rate or life expectancy
• Cost-effectiveness as cost/unit change in outcome
  variable

65
Societal Efficacy

• Benefit-cost analysis from societal viewpoint
• Cost-effectiveness analysis from societal
  viewpoint

66
The PEDS Study Pulmonary Embolism Diagnosis
Study Investigators Dr. S. Kahn, Dr. A. Hirsch,
Dr. E. Lang, Dr.A. Guttman, Dr. M. Afilalo
Start date September 20th

• Funded by CIHR Clinical Trials Program
• Objectives To determine whether spiral CT can be
  relied on as a safe alternative to V/Q lung
  scanning as the initial diagnostic imaging
  procedure for the evaluation of patients with
  suspected PE.
• Inclusion Criteria Patients with symptoms or
  signs suspected to be caused by acute pulmonary
  embolism (acute onset of new or worsening
  shortness of breath, chest pain, hemoptysis,
  presyncope or syncope).
• Note If you have potential study patients,
• BEFORE ORDERING A DIAGNOSTIC TEST FOR P.E.
  please contact
• For ED patients only Chris Tselios pager
  440-4143
• For all in-patients Carla Strulovitch pager
  981-7932

67
(No Transcript)
68
(No Transcript)
69
Conclusions
70
Educational Objectives

• Review the rationale for the development of
  clinical prediction rules in venous
  thromboembolism (VTE)
• Conduct structured critical appraisal of the best
  prediction rule / clinical model research in VTE
• Explore issues related to test selection in
  patients suspected of VTE