Validation of the sensitivity of the National Emergency X-Radiography Utilization Study (NEXUS) Head computed tomographic (CT) decision instrument for selective imaging of blunt head injury patients: An observational study
Bibliographic record
Abstract
BACKGROUND: Clinicians, afraid of missing intracranial injuries, liberally obtain computed tomographic (CT) head imaging in blunt trauma patients. Prior work suggests that clinical criteria (National Emergency X-Radiography Utilization Study [NEXUS] Head CT decision instrument [DI]) can reliably identify patients with important injuries, while excluding injury, and the need for imaging in many patients. Validating this DI requires confirmation of the hypothesis that the lower 95% confidence limit for its sensitivity in detecting serious injury exceeds 99.0%. A secondary goal of the study was to complete an independent validation and comparison of the Canadian and NEXUS Head CT rules among the subgroup of patients meeting the inclusion and exclusion criteria. METHODS AND FINDINGS: We conducted a prospective observational study of the NEXUS Head CT DI in 4 hospital emergency departments between April 2006 and December 2015. Implementation of the rule requires that patients satisfy 8 criteria to achieve "low-risk" classification. Patients are excluded from "low-risk" classification and assigned "high-risk" status if they fail to meet 1 or more criteria. We examined the instrument's performance in assigning "high-risk" status to patients requiring neurosurgical intervention among a cohort of 11,770 blunt head injury patients. The NEXUS Head CT DI assigned high-risk status to 420 of 420 patients requiring neurosurgical intervention (sensitivity, 100.0% [95% confidence interval [CI]: 99.1%-100.0%]). The instrument assigned low-risk status to 2,823 of 11,350 patients who did not require neurosurgical intervention (specificity, 24.9% [95% CI: 24.1%-25.7%]). None of the 2,823 low-risk patients required neurosurgical intervention (negative predictive value [NPV], 100.0% [95% CI: 99.9%-100.0%]). The DI assigned high-risk status to 759 of 767 patients with significant intracranial injuries (sensitivity, 99.0% [95% CI: 98.0%-99.6%]). The instrument assigned low-risk status to 2,815 of 11,003 patients who did not have significant injuries (specificity, 25.6% [95% CI: 24.8%-26.4%]). Significant injuries were absent in 2,815 of the 2,823 patients assigned low-risk status (NPV, 99.7% [95% CI: 99.4%-99.9%]). Of our patients, 7,759 (65.9%) met the inclusion and exclusion criteria of the Canadian Head CT rule, including 111 patients (1.43%) who required neurosurgical intervention and 306 (3.94%) who had significant intracranial injuries. In our study, the Canadian criteria for neurosurgical intervention identified 108 of 111 patients requiring neurosurgical intervention to yield a sensitivity of 97.3% (95% CI: 92.3%-99.4%) and exhibited a specificity of 58.8% (95% CI: 57.7%-59.9%). The NEXUS rule, when applied to this same cohort, identified all 111 patients requiring neurosurgical intervention, yielding a sensitivity of 100% (95% CI: 96.7%-100.0%) with a specificity of 32.6% (95% CI: 31.5%-33.6%). Our study found that the Canadian medium-risk factors identified 301 of 306 patients with significant injuries (sensitivity = 98.4%; 95% CI: 96.2%-99.5%), while the NEXUS rule identified 299 of these patients (sensitivity = 97.7%; 95% CI: 95.3%-99.1%). In our study, the Canadian medium-risk rule exhibited a specificity of 12.3% (95% CI: 11.6%-13.1%), while the NEXUS rule exhibited a specificity of 33.3% (95% CI: 32.3%-34.4%). Limitations of the study may arise from application of the rule by different clinicians in different environments. Clinicians may vary in their interpretation and application of the instrument's criteria and risk assignment and may also vary in deciding which patients require intervention. The instrument's specificity is also subject to spectrum bias and may change with variations in the proportion of "low-risk" patients seen in other centers. CONCLUSIONS: The NEXUS Head CT DI reliably identifies blunt trauma patients who require head CT imaging and could significantly reduce the use of CT imaging.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".