A recent Emergency Medicine digest,
as sent to subscribers

Study Type: Retrospective cohort study (clinical prediction model development and validation)

Key Question: Can a modified Sudbury Vertigo Risk Score — replacing BPPV diagnosis with history-based variables — reliably identify central causes of dizziness in ED patients?

Key Findings:

• Central aetiology confirmed in 5.2% of 2,958 patients; modified model achieved AUROC 0.81 (95% CI 0.77–0.85) vs 0.85 (95% CI 0.82–0.88) for the original Sudbury score
• The modified model's miss rate (false negatives) was 0.9% (95% CI 0.4–2.1%) vs 0.0% for the Sudbury model — a clinically meaningful safety trade-off
• The modified model identified 18.7% of patients as low-risk (vs 26.8%), indicating reduced efficiency alongside the lower safety margin

Clinical Relevance: For UK ED clinicians unable to reliably assign a BPPV diagnosis at triage, this modified tool offers a history-only alternative, though its higher miss rate warrants caution before adoption.

Limitations: Single-centre retrospective design limits generalisability; no external validation cohort was used.

Study Type: Retrospective observational study (pilot)

Key Question: Can modest financial incentives encourage low-acuity ED patients to consent to interfacility transfer, thereby reducing ED overcrowding?

Key Findings:

• Only 4 patients participated; all received travel vouchers ($300–$500) and were transferred safely to critical access hospitals without complication or retransfer
• All 4 patients were ultimately discharged home
• Patient satisfaction was high (Net Promoter Score +75), though this reflects an extremely small sample

Clinical Relevance: ED boarding and overcrowding are significant patient safety issues in NHS emergency departments; this concept of incentivised, consent-based diversion to underutilised facilities could theoretically inform UK system-level thinking, though the US regulatory (EMTALA) and financial framework differs substantially from NHS structures.

Limitations: The sample size of four patients makes no meaningful conclusions possible regarding safety, efficacy, or generalisability; this is hypothesis-generating only.

Study Type: External validation study (secondary analysis of a multicentre prospective observational cohort)

Key Question: Does the ROMEO score accurately identify adult ED syncope patients at very low risk of significant echocardiographic findings, thereby safely avoiding TTE?

Key Findings:

• A ROMEO score of zero achieved sensitivity 98.9% and NPV 98.6% for significant TTE findings, suggesting strong rule-out performance
• Specificity was low at 20.2% (PPV 24.3%), limiting its utility as a rule-in tool
• AUC 0.83 (95% CI 0.79–0.87); 20.6% of patients receiving TTE had significant structural findings

Clinical Relevance: TTE is frequently ordered in UK EDs and AMUs for syncope; a validated low-risk score could safely reduce unnecessary echocardiography, supporting resource stewardship in an NHS context.

Limitations: Significant selection bias risk due to high non-enrolment rates and over-representation of urban academic centres, limiting generalisability to community ED settings.

Study Type: Observational comparative study

Key Question: How well does an AI self-triage programme agree with standard nurse-led triage (Dutch NTS) in adult ED attendees?

Key Findings:

• Agreement between AI and NTS triage was negligible (Cohen's κ = 0.092, 95% CI −0.196 to 0.380)
• AI overtriaged 12.8% and undertriaged 5.4% of patients versus NTS; however, it allocated more patients with serious clinical sequelae to higher-urgency categories than NTS
• AI correctly predicted final ED diagnosis in only 27.1% of cases

Clinical Relevance: As NHS EDs explore AI-assisted self-triage to manage demand and streaming, this study signals that current tools lack the concordance with established triage standards needed for safe independent deployment.

Limitations: Small single-centre Dutch cohort (n = 203) limits generalisability to UK ED populations and triage systems such as the Manchester Triage System.

Study Type: Prospective observational cohort study

Key Question: Are intra-arrest diastolic blood pressure (DBP) and end-tidal CO₂ (EtCO₂) associated with return of spontaneous circulation (ROSC) in adult ICU cardiac arrest?

Key Findings:

• Mean DBP was significantly higher in events achieving ROSC (39 vs 24 mmHg; mean difference 15 mmHg, 95% CI 8–23; p<0.001)
• Mean EtCO₂ was also higher in ROSC events (19 vs 15 mmHg; difference 4 mmHg, 95% CI 1–6; p=0.010)
• DBP showed numerically better discrimination than EtCO₂ (AUC 0.78 vs 0.69) but this difference was not statistically significant

Clinical Relevance: For ICU patients with arterial lines — increasingly common in resuscitation room settings — DBP monitoring during CPR may complement EtCO₂ as a real-time physiological marker of resuscitation quality.

Limitations: Small single-centre study (68 events), limiting generalisability to broader emergency or non-ICU cardiac arrest populations.

Study Type: Three-arm parallel-group RCT

Key Question: Does the site of tactile stimulation (sole, trunk, or back rub) affect oxygenation outcomes in non-crying neonates ≥32 weeks requiring resuscitation at birth?

Key Findings:

• No significant difference in primary outcome (SpO₂ at 5 minutes): sole rub 83.1%, trunk rub 85.4%, back rub 82.4% (p=0.392)
• Minute-by-minute HR, SpO₂, and FiO₂ trends were comparable across all three groups throughout the first 10 minutes
• No differences in need for further resuscitation, cerebral oxygenation at one hour, or in-hospital complications including mortality

Clinical Relevance: Clinicians can use whichever tactile stimulation site is most practical during neonatal resuscitation without compromising physiological outcomes — relevant to UK delivery room practice and NLS training.

Limitations: Single-centre trial of 183 neonates ≥32 weeks limits generalisability to preterm infants below this threshold.

Study Type: Retrospective cohort study

Key Question: Does conventional CPR with rescue breaths improve outcomes compared to compression-only or no CPR in bystander-witnessed suffocation-related OHCA?

Key Findings:

• After IPTW adjustment, no bystander CPR was associated with *better* neurological outcomes than compression-only CPR (aOR 1.20, 95%CI 1.04–1.38), which is counterintuitive and likely reflects confounding
• Conventional CPR (compressions + ventilation) showed no significant benefit over compression-only CPR for any outcome
• Favourable neurological outcomes were poor across all groups (1.5–2.0%), suggesting suffocation-related OHCA carries a very high mortality regardless of bystander CPR type

Clinical Relevance: Challenges the assumption that adding rescue breaths improves outcomes in suffocation OHCA; relevant to UK resuscitation council guidance and public-facing CPR training priorities.

Limitations: Residual confounding is likely substantial — patients receiving no CPR may have had shorter collapse-to-EMS intervals or less severe arrests, undermining causal inference despite IPTW adjustment.

Study Type: Population-based registry study (retrospective cohort)

Key Question: Does the quality of bystander CPR, as assessed by paramedics, affect 30-day survival in out-of-hospital cardiac arrest?

Key Findings:

• Good-quality bystander CPR was associated with a 162% increase in survival odds (aOR 2.62, 95% CI 1.86–3.68), whereas poor-quality bystander CPR showed no survival benefit (aOR 0.99, 95% CI 0.68–1.45)
• Any bystander CPR (regardless of quality) was associated with a 70% increase in survival odds (aOR 1.70, 95% CI 1.24–2.34)
• Good-quality CPR was more likely in younger victims, witnessed arrests, and non-residential settings

Clinical Relevance: For UK ambulance services and EDs, this reinforces that public CPR training should prioritise compression quality, not just initiation — relevant to NHSE community resuscitation initiatives.

Limitations: Paramedic-assessed CPR quality introduces significant inter-rater subjectivity, and quality data was only recorded in 62% of cases.