Answers are grounded in this paper and cite their sources.
Background Early abdominal computed tomography (CT) is often obtained in acute pancreatitis, although the frequency of potentially actionable report findings and their association with later care are uncertain. We estimated the positive proportion produced by a rule-based report classifier and examined associations between early CT, subsequent management, and resource use among admissions remaining alive, hospitalized, and free of selected major organ support through 24 h. Methods This retrospective cohort linked MIMIC-IV structured records with MIMIC-IV-Note radiology reports. Eligible admissions remained alive, hospitalized, and free of invasive mechanical ventilation, vasopressors, or renal replacement therapy through 24 h. A prespecified phrase dictionary classified the first early-CT report from each admission without independent clinician validation; unresolved reports remained indeterminate. Stabilized inverse-probability-of-treatment weights were winsorized at the 1st and 99th percentiles. Doubly robust outcome models used patient-clustered standard errors. Results The cohort included 4,949 admissions from 3,216 patients; 1,857 received at least one early CT. The classifier was applied only to the first early-CT report. Among 1,599 determinate first early-CT reports, 193 were rule-positive for a potentially actionable finding (12.1%, 95% confidence interval 10.4% to 13.8%); 258 were indeterminate. This percentage is a classifier output, not a clinically validated diagnostic-yield measure. Early CT was not clearly associated with management change within 24 to 72 h (adjusted risk ratio 1.09, 95% confidence interval 0.92 to 1.31; P = 0.315) or post-landmark length of stay (ratio of means 1.07, 95% confidence interval 0.99 to 1.15; P = 0.069). Early CT was associated with subsequent abdominal CT and modestly higher laboratory use. Conclusions The rule-based classifier identified a potentially actionable finding in 12.1% of determinate first early-CT reports. Early CT was associated with greater diagnostic intensity, without a clear association with recorded management change or length of stay. Independent clinician validation is needed before clinical interpretation of the classifier output.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: