Errors in the diagnostic process

1
Errors in the diagnostic process

• Hierarchy of Qualities in Medicine
• Frequency of diagnostic Errors
• Judgment under Uncertainty Heuristics and Biases
• The Voytovich Solution

Click to View Presentation
2
Hierarchy of Qualties in MedicineFrom patients
point of view
Patient Satisfaction
Therapeutic Quality
Diagnostic Quality
3
Hierarchy of Qualties in MedicineMedical point
of view
Patient Satisfaction
Therapeutic Quality
Diagnostic Quality
4
Hierarchy of Qualities in Medicine und cognitive
Processes
Patient Satisfaction
Skills
Rules
Therapeutic Quality
Knowledge
Diagnostic Quality
5
Frequency of diagnostic Errors Follow-up Autopsy
Study Medizinische Klinik USZ 1972-1982-1992-2002
Lancet 20003552027-31

• Random Selection of 100 patients in each year
• Autopsy Rate above 90 until 1992, in the year
  2002 Reduction to 53, complete Autopsy
• Classification of diagnostic Errors according to
  Goldman

6
Classification of diagnostic errors Goldman et
al NEJM 1983380 1000-05

• Major diagnostic Errors
• Class I Knowledge of correct Diagnosis would
  have led to Survival
• Class II Knowledge of correct Diagnosis would
  not have affected Survival (too ill, no Treatment
  available)
• Minor diagnostic Errors
• Class III Missed Diagnosis but not cause of
  Death
• Class IV Occult, clinically not diagnosable
  Entity of epidemiological Interest eg Gallstones

7
Major Diagnostic Errors 1972-2002
8
Minor Diagnostic Errors 1972-2002
9
Class III vs. Class IIIIV
10
Correct Diagnosis 1972-2002
11
Autopsy Rate and class I Errors over time
Kaveh G et al, JAMA 2003 2892849-56
12
Frequency of class I Errors

• Klasse I Autopsie- Fehler rate
• Medizin IPS (USZ) 2002 2 53
• Med-IPS (Mayo-Clinic) 1998-2000 4 33
• Med-IPS (Paris) 1995-98 10.2 53
• Med-IPS (Leuven,Belgien) 1996 16 93
• 32 Spitäler in USA 1984 13 30
• Medizin (Boston, USA) 1984 12 40
• Arch Int Med 2004164389 Mayo Clin Proc
  200075562 Ann Thorac Surg 199764380
  JAMA1987258339 Mayo Clin Proc. 200378947-50.
  NEJM 19883181249

13
Autopsy Gold Standard for clinical Diagnosis?
Pelletier et al J Gen Intern Med 19894300-03
14
Autopsy Gold standard for clinical diagnosis?
Pelletier et al J Gen Intern Med 19894300-03
15
Judgment under Uncertainty Heuristics and
BiasesTversky and Kahneman Science
19741851124-31

• Representativness
• Similarity with typical examples stored in
  memory
• Availability
• Recent expierence, painful memory
• Adjustment and Anchoring
• Stick to early hypotheses despite new information

16
(No Transcript)
17
Cognitive Mechanisms of diagnostic Errors
according to A.E. Voytovich

• Omission
• Premature Closure
• Inadequate Synthesis
• Wrong Formulations

J Med Educ 198560302-07
18
Omission

• Most frequent Error
• Decreases with Experience
• ? Consequences
• Delayed or missed Diagnosis

19
Premature Closure

• Independent of Experience
• Correlates with Confidence (ie Overconfidence) in
  Relation to the actual Case
• Reflects estimated Frequency of the diagnosed
  Disease
• ? Consequences
• Delayed or missed Diagnosis
• Unnecessary Therapies
• False Sense of Confidence if Error is not detected

20
Inadequate Synthesis

• Correlates with Experience
• ? Consequences
• Unnecessary Investigations
• Delayed Treatment

21
Cognitive Mechanisms of diagnostic Errors
according to A.E. Voytovich

• Omission
• Premature Closure
• Inadequate Synthesis
• Wrong Formulations

J Med Educ 198560302-07
22
Lancet 20003552027-31
23
Major Diagnostic Errors 1972-2002
24
Sensitivity and Specificity

• 1-Sensitivity Rate of missed Diagnoses (false
  negative rate)
• 1-Specificity Rate of wrong Diagnoses (false
  positive rate)

25
Cardiovascular Diseases Sensitivity and
Specificity
p 0.061
p 0.034
100
97
90
86
85
1972
82
82
80
1982
1992
69
70
60
50
Sensitivität
Spezifität
26
Lancet 20003552027-31
27
Infectious Diseases Sensitivity and Specificity
ns
p 0.036
100
100
99
100
86
80
67
1972
60
1982
1992
40
25
20
0
Sensitivität
Spezifität
28
Neoplastic Diseases Sensitivity and Specificity
ns
ns
100
97
96
96
95
92
89
1972
90
88
1982
85
1992
80
75
70
Sensitivität
Spezifität
29
Difficulties in learning from Experience

• Lack of Search for and use of disconforming
  Evidence
• Lack of outcome Information
• Use of unaided Memory for coding, storing and
  retrieving outcome Information

30
Summary and Proposal

• Major diagnostic Errors occur despite an ever
  increasing repertory of diagnostic Procedures
• 85 of serious diagnostic Errors can only be
  detected by Autopsy
• Analysis of error mechanism can be helpful in the
  Discussion and Prevention of diagnostic Errors
• Minimum Autopsy rate of 30 along with a yearly
  Report on diagnostic Errors should be mandatory
  for Accreditation of medical Clinics