Evaluation of Head CT Exams Resident

1
Evaluation of Head CT Exams - Resident
Attending Diagnoses

• Elizabeth Krupinski, PhD
• William Berger, MD
• William Erly, MD
• University of Arizona

2
Rationale

• Many medical centers offer 24-hour CT
  interpretation for the ED
• On-call system
• Teleradiology
• Radiology residents often provide preliminary
  interpretations

3
Rationale

• Previous reports indicate overall resident miss
  rates are low (lt 2)
• No one has looked at differences between upper
  lower level residents
• Assessing the types of errors made by residents
  may help focus educational efforts

4
Methods

• 1324 consecutive CT head exams ordered by ED
• Initially interpreted by 1 of 18 radiology
  residents
• Interpretation
• Confidence rating on 6-point scale
• 1 of 5 CAQ neuroradiologists interpreted exams
  same day or next morning
• Determined correlation of diagnoses

5
Correlation of Interpretations

• Agree
• Disagree - Insignificant
• No potential for adverse patient outcome
• Failure to identify ischemic white matter
  degeneration
• There was no error in synthesis

6
Correlation of Interpretations

• Disagree - Significant
• Potential for adverse patient outcome
• Mistaking cytoxic edema of metestatic disease for
  ischemic change
• Gross error in synthesis without adverse outcome
  potential
• Mistaking dysgenesis of the corpus
  callosum for chronic hydrocephalus

7
Diagnostic Correlation
8
Disagreement Cases
X2 44.46, df 2, p 0.0001
9
Year of Residency
X2 13.80, df 6, p 0.032
10
Resident Confidence
X2 37.55, df 4, p 0.0001
11
Year of Residency
F 92.01, p 0.0001
12
Confidence Cases
X2 74.56, df 2, p 0.0001
13
Disagree - Insignificant

• 42 False Negatives
• Chronic ischemic foci
• Fractures
• Atrophy
• 17 False Positives
• Intracranial hemorrhage
• Fractures

14
Disagree - Significant

• 14 False Negative
• Fractures
• Acute right occipital infarct
• Temporal lobe contusion
• Chronic cortical infarct
• Clius mass
• Small thalamic hemorrhage

15
Disagree - Significant

• 3 Misinterpretation
• Metastatic disease as ischemic focus
• Tumor as infarct
• Chiari II malformation with callosal dysgenesis
  as hygroma-hydrocephalus-mass effect
• 2 False Positive
• Cerebellar Hemorrhage
• Cerebral contusion

16
Conclusions

• Resident error is low (lt 2)
• Rate of potential change in patient outcome due
  to resident misinterpretation is low (0.05)
• Year of residency affects performance confidence

17
Conclusions

• When resident confidence is low, immediate
  consultation with neuroradiologist is advisable
• Efforts to reduce resident errors should focus on
  ensuring that all bone window images are reviewed
  and on the signs of cytoxic versus vasogenic edema