The Influence of a Vocalized Checklist on Detection of Multiple Abnormalities in Chest Radiography

Area of Science:

• Radiology
• Medical Imaging
• Diagnostic Accuracy

Background:

• Satisfaction of search (SOS) is a cognitive bias that can lead to missed diagnoses in medical imaging.
• Previous studies suggested checklists could prevent SOS errors, but practical implementation challenges, like diverting attention from the image display, remained.
• A vocalized checklist was proposed as a solution to maintain focus on the image while utilizing a structured search protocol.

Purpose of the Study:

• To evaluate the effectiveness of a vocalized checklist in preventing satisfaction of search (SOS) errors during chest radiograph interpretation.
• To determine if a vocalized checklist mitigates the reluctance to report abnormalities when a simulated pulmonary nodule is present.

Main Methods:

• Twenty radiologists interpreted 64 chest computed radiographs twice: once with and once without a simulated pulmonary nodule.
• A vocalized checklist was used during interpretation to direct the search pattern.
• Receiver operating characteristic (ROC) analysis was employed to assess detection accuracy and decision thresholds.

Main Results:

• The presence of a simulated pulmonary nodule significantly increased the reluctance to report other abnormalities (P < 0.001).
• The vocalized checklist did not reduce or eliminate the SOS effect.
• While the checklist aided in organizing search and reporting, it did not prevent the bias.

Conclusions:

• A vocalized checklist, despite its potential for organizing search, does not prevent satisfaction of search (SOS) errors in chest radiography.
• The cognitive bias of SOS persists even with structured search protocols when a salient finding is present.
• Further strategies are needed to effectively combat SOS errors in diagnostic imaging.