Bayesian test-selection OSCE — neonate residual risk and caregiver counselling

Candidate brief

You have 10 minutes. [8] [5]

1. Build a one-sentence problem representation and pre-test risk estimate. [8]
2. Choose discriminating tests (and omit non-discriminating ones) with reasons. [3] [4]
3. Interpret a normal early inflammatory marker for residual risk. [2] [9]
4. Counsel the caregiver without false certainty. [5]
5. Hand over pending results with owner and next action. [5]

Scripted clinical data

• 16-day-old term neonate, day 2 of poor feeding, fewer wet nappies. [4]
• Caregiver: “He is not himself; this is different from yesterday.” [5]
• Mottled peripheries, cool feet, heart rate high for age, interactive but quieter than usual. [4]
• Early CRP normal; blood culture pending; bag urine sent before your arrival. [2] [6]

Expected performance

Representation and pre-test risk

• Age, tempo, physiology and caregiver concern in one sentence.
• High residual risk for serious bacterial infection and other neonatal threats despite a normal early marker. [4] [5] [8]

Test selection

• Prioritise bedside stabilisation and discriminating septic work-up elements per local pathway.
• Do not claim that normal CRP alone completes risk assessment. [2] [4]
• Question bag-urine validity; plan better collection if UTI remains live. [6]
• Avoid non-discriminating “full panel for completeness.” [7]

Bayesian interpretation

• Pre-test high × imperfect negative update → residual risk may remain above discharge threshold. [2] [9] [3]
• State what would raise or lower probability next (trajectory, cultures, repeat assessment). [2] [3]

Communication

• “I am still worried about serious infection even though one blood test is reassuring.”
• Explain what you will do now, what you are watching for, and why discharge is unsafe if residual risk remains high. [5]
• Avoid both false reassurance and unsorted rare-disease catalogues.

Handover

• Working diagnosis, residual risks, actions taken, response, pending results, owner, expected impact, review time. [5]

Examiner scoring anchors

Common fails

• Discharging on a single normal marker. [2] [4]
• Treating bag-urine mixed growth as definite disease without method review. [6]
• Explaining “the test is negative, so we have ruled it out.” [1] [2]
• No owner for pending cultures. [5]