30-Second Takeaway
- An accurate ML mortality score did not change ED decisions or outcomes without actionable, trusted integration.
- Perfusion index during CPR tracks ROSC and admission and suggests a potential physiologic target around 0.6.
- Routing, tele-neurosurgery, and steroid use highlight major systems- and protocol-level opportunities to improve outcomes.
- Quick assessment of premorbid function in older adults strongly informs admission risk and length of stay.
- Public underrecognition of sepsis continues to delay emergency care despite high stated urgency.
Week ending December 6, 2025
Risk tools, resuscitation signals, and systems-level levers that actually modify ED care
Accurate ML mortality score fails to change ED management or outcomes
In this single-center open-label RCT, 1303 ED adults were randomized to usual care versus access to the RISKINDEX mortality score. RISKINDEX, using labs, age, and sex, achieved AUROC 0.84 for 31-day mortality, outperforming NEWS, APACHE II, SOFA, and clinician judgment. Model predictions conflicted with clinicians’ expectations in roughly half of cases, especially among less experienced physicians. Despite superior discrimination, treatment plans changed in only 1 of 644 patients and clinical outcomes were unchanged. Clinicians reported low added value, underscoring that prognostic accuracy alone is insufficient without trusted, actionable workflow integration.
Perfusion index during OHCA CPR predicts ROSC and survival to admission
This prospective cohort followed 98 out-of-hospital cardiac arrests with perfusion index monitoring during CPR. Events with sustained ROSC had higher event-average perfusion index than non-ROSC events, with AUROC 0.77 for predicting sustained ROSC. An optimal PI threshold around 0.61 yielded sensitivity 0.67 and specificity 0.81 for sustained ROSC. Higher PI independently increased odds of sustained ROSC and survival to hospital admission, but not neurologically intact survival. Cycle-average PI correlated positively with chest compression fraction, linking better mechanics to improved peripheral perfusion.
Ten percent of suspected strokes routed to non–stroke-capable EDs, especially in rural areas
Using a linked EMS–ED capability dataset, investigators analyzed 314,642 suspected stroke encounters from 1781 EMS agencies between 2019 and 2021. Overall, 10% of patients were transported to hospitals lacking verified stroke care capabilities. Rural service area was strongly associated with lower odds of transport to a stroke-capable ED, with odds ratio 0.15 versus urban. There was large between-agency variation, with median odds ratios exceeding 3 even in fully adjusted models. These findings highlight modifiable gaps in destination protocols and stroke systems of care, particularly for rural populations.
Low-dose, short-course steroids probably reduce short-term mortality in severe pneumonia and ARDS
This systematic review and meta-analysis included 20 randomized trials with 3459 adults with severe pneumonia or ARDS. Regimens used systemic corticosteroids ≤3 mg/kg per day (prednisone-equivalent) for ≤15 days, initiated within 7 days of illness onset. In severe pneumonia, steroids probably reduced short-term mortality, with pooled risk ratio 0.73 and low heterogeneity. In ARDS, similar regimens probably reduced short-term mortality with risk ratio 0.77 and modest heterogeneity. Steroids may reduce secondary shock in severe pneumonia and probably cause little to no increase in hospital-acquired or secondary pneumonias.
References
Numbered in order of appearance. Click any reference to view details.
Additional Reads
Optional additional studies from this edition.