{
  "abstract": "Aims and Objectives Timely CT head interpretation is critical in emergency departments (EDs), yet radiology reporting delays are common and the baseline diagnostic accuracy of ED clinicians remains poorly understood. This multicentre randomised controlled trial investigated whether structured online simulation-based training could improve ED clinicians’ diagnostic accuracy and efficiency in interpreting non-contrast CT head scans, and assessed the potential for earlier clinical decision-making compared to formal radiology reports.Method and Design 186 clinicians (ED doctors, advanced nurse practitioners, physician associates, and radiographers) across six NHS hospitals in the Thames Valley Emergency Medicine Research Network were randomised 5:1 to training (n=150) or control (n=30) groups ( figure 1). Both completed a baseline 50-case online assessment. The training group completed a bespoke online training module on the RAIQC platform before a 3-month clinical phase where all participants interpreted 30 CT head scans during routine shifts (figure 2). Sensitivity, specificity, accuracy, and reporting times were assessed against radiologist reports as the gold standard, with analysis by role, seniority, and pathological subgroup.Abstract 4165 Figure 1Results and Conclusion 163 clinicians completed the study (134 training, 29 control), interpreting 4,424 CT head scans. Overall pooled sensitivity was 71.7% (95% CI: 67.8-75.4) and specificity 94.4% (95% CI: 93.6-95.1). No significant difference in sensitivity was observed between training and control groups (72.1% vs 69.9%, p=0.76). Clinician interpretations preceded radiology reports by mean 66 minutes (95% CI: 50-82). Sensitivity varied by pathology ( figure 3): highest for intraparenchymal haemorrhage (82.1%) and lowest for uncal herniation (28.6%). ED clinicians demonstrated high specificity and moderate sensitivity in detecting intracranial pathology. While online training did not significantly improve diagnostic accuracy, clinical interpretation facilitated earlier clinical decision-making, potentially streamlining ED workflow. Future integration of AI-assisted interpretation tools may further enhance diagnostic performance and efficiency in emergency settings.",
  "authors": [
    {
      "affiliations": [
        "Emergency Medicine Research Oxford (EMROx), Oxford University Hospitals NHS Foundation Trust, Oxford"
      ],
      "name": "Alex Novak"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "Bharath Gopinath"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "Tanya Baron"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "Matthew Davies"
    },
    {
      "affiliations": [
        "Milton Keynes University Hospital NHS Foundation Trust"
      ],
      "name": "Divyansh Gulati"
    },
    {
      "affiliations": [
        "Royal Berkshire NHS Foundation Trust"
      ],
      "name": "Simon Triscott"
    },
    {
      "affiliations": [
        "Frimley Health NHS Foundation Trust"
      ],
      "name": "Sarah Wilson"
    },
    {
      "affiliations": [
        "Buckinghamshire Healthcare NHS Trust"
      ],
      "name": "Ravi Shashikala"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "Sally Beer"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "Sarim Ather"
    },
    {
      "affiliations": [
        "Oxford University Hospitals NHS Foundation Trust"
      ],
      "name": "James Ray"
    }
  ],
  "title": "4165 Impact of online training on emergency department clinicians’ diagnostic performance in CT head interpretation: a multicentre randomised controlled trial",
  "uid": "e3c59e0b-d64e-5e82-b941-63a1ba0c28a1"
}