Emergency Medicine Journal

Background Emergency departments (EDs) in high-income countries face rising demand, workforce shortages and crowding. We investigated whether prehospital point-of-care blood gas analysis (BGA), used by emergency physicians, is associated with higher ambulatory treatment rates and improved patient selection for hospital admission. Methods We retrospectively analysed routinely collected data from a pilot implementation of a mobile blood gas analyser in physician-staffed emergency medical services (EMS) in Jena, Germany (July 2023 to May 2024). Adult emergency patients receiving prehospital BGA were compared with propensity score-matched EMS controls without BGA. Primary outcomes were the proportion treated on scene and, among transported patients, the hospital admission rate. Secondary outcomes were 30-day safety among ambulatory patients and associations between BGA parameters and disposition. We used standardised mean differences to assess balance and receiver operating characteristic analysis for lactate thresholds. Results Of 109 patients receiving prehospital BGA, 98 met inclusion criteria after excluding 9 patients with missing NACA scores, 1 on-scene death and 1 invalid age record; these were matched to 390 controls (total n = 488). Baseline demographics, severity and vital signs were well balanced. Ambulatory treatment was markedly higher in the BGA cohort compared with matched controls (27.6% vs 8.7%; OR 3.98, 95% CI 2.26 to 7.01; p<0.001). No ambulatory BGA patient required ED re-attendance or repeat EMS contact within 30 days. Among transported patients, 58% in the BGA cohort were admitted to hospital, compared with an overall regional ED conversion rate of approximately 30%. Lactate [&ge;]2.6 mmol/L was the most influential parameter for disposition decisions, with elevated lactate and acid-base disturbances strongly associated with transport and admission. Conclusion Prehospital BGA was associated with fourfold higher ambulatory treatment rates (27.6%) and a twofold higher ED conversion rate among the patients who were transported (58%), indicating improved risk stratification and resource allocation. These findings suggest that integrating objective biochemical data into prehospital assessment may enhance treat-and-refer decision-making and support more efficient use of limited emergency care capacity.

Objectives: Emergency department (ED) overcrowding is a persistent issue in European healthcare systems. A previous randomized controlled trial (RCT) concerning out of hours care in Antwerp (2019) demonstrated that a nurse-led triage tool, extending the Manchester Triage System (eMTS), could safely redirect low-acuity ED patients to a co-located General Practitioner Cooperative (GPC). This study reports a six-year follow-up assessing long-term efficiency, safety, and sustainability of this intervention. Methods: We performed a retrospective observational analysis of routine clinical data. Patients triaged at the ED and referred to the GPC were identified through electronic health records. Efficiency outcomes included the proportion of ED patients managed at the GPC, the proportion of GPC patients originating from the ED and their clinical characteristics. To assess safety, we analysed rates and characteristics of patients referred back from the GPC to the ED. A detailed case review was conducted for all back-referred patients. Results: Of the 110,941 triaged patients, 6,722 (6.1%) were managed at the GPC, accounting for 11% of all GPC consultations. Diverted patients typically presented with digestive, respiratory, and musculoskeletal complaints and had a clinical urgency which was mostly comparable to the overall GPC population. Only 3% of the patients diverted to the GPC were referred back to the ED, versus 5% of other GPC patients. Most back-referrals (83%) were managed on an outpatient basis; four major and 18 minor triage issues were identified, without evidence of increased morbidity. Conclusions: Six years post-trial, the nurse-led eMTS triage tool remains integrated into routine practice, with increasing efficiency and remaining safety without dedicated research resources nor a post implementation plan. Sustained adoption highlights its clinical feasibility and long-term safety. Future trials on triage and primary care should embed explicit post-trial implementation strategies to promote continuity and scalability of successful healthcare interventions. ClinicalTrials.gov Identifier: NCT03793972

BackgroundPotassium is involved in multiple physiological processes in the body, and hyper-kalemia is a common, potentially life-threatening condition. ObjectiveThe aim of our study was to examine the association between plasma potassium levels, and 30-day mortality in patients presenting to an emergency department with normo- or hyperkalemia. DesignRetrospective Cohort study. SettingEmergency Departments in the Capital region of Denmark ParticipantsPersons attending Emergency Departments in the Capital Region of Denmark from 2017-2021 with a plasma potassium level of at least 3.5 mM measured within 4 hours after arrival. MeasurementsThe study was based on data from Danish National Registries and electronic patient records. We performed Kaplan-Meier survival analyses and unadjusted and adjusted cox regression analyses utilizing plasma [K+] 3.5-4.4 mM as the reference group for 30-day mortality hazard ratios (HRs). ResultsA total of 248,453 patients were included with a median age of 60 years (Q1;Q3 42;75), and 6,959 (2.8%) died within 30 days. Mortality was 2.2% for potassium level 3.5-4.4 mM, 6.9% for 4.5-4.9 mM, 17.1% for 5.0-5.9 mM, and 26.9% for [&ge;]6.0 mM. Unadjusted 30-day HRs were 3.2 (95%CI: 3.0-3.4) for [K+] 4.5-4.9 mM, 8.6 (95%CI: 7.9-9.3) for [K+] 5.0-5.9 mM, and 14.7 (95%CI: 12.5-17.0) for [K+] [&ge;]6.0 mM. Adjusted HRs were 1.4 (1.3-1.5), 2.10 (1.9-2.3), and 2.4 (2.0-2.8), respectively. LimitationsRisk of residual confounding. Missing data. No access to data regarding in-hospital treatment. ConclusionPlasma potassium levels above 4.4 mM were associated with increased 30-day mortality among patients presenting to emergency departments. Primary funding sourceDepartment of Emergency Medicine, Copenhagen University hospital, Bispebjerg and Frederiksberg Hospital.

Objectives People with mental health conditions experience increased rates of diagnostic errors and delays in acute treatment. While causes such as diagnostic overshadowing (misattribution of physical symptoms to mental health conditions) are well documented, less attention has been paid to the organisational and structural conditions that shape diagnostic work. This study examines how physical illness is diagnosed in patients with mental health conditions in emergency departments (EDs), with a focus on the structural conditions that enable or constrain safe diagnostic practice. Method We conducted a multi-site ethnography across three purposively selected EDs in England between April 2023 and April 2024, varying in size, population demographics, and local service configuration. Data were collected through 284 hours of non-participant observation and 20 semi-structured interviews with ED staff. Results Our analysis identified four recurring structural gaps that shaped the conditions under which physical health diagnosis took place for patients with mental health conditions: a design gap, whereby targets and physical layouts constrained diagnostic reasoning; a preparedness gap, reflecting the lack of structural support to allow staff to act on their existing knowledge and skills; a coordination gap, reflecting fragmented ownership and the challenges of joint assessment across mental and physical healthcare teams; and an expectation gap, whereby unmet need elsewhere in the system increased demand for ED services that were beyond its formal scope. These gaps made diagnostic errors and delay more likely for patients with mental health conditions seeking physical healthcare in the ED. Conclusions As new dedicated mental health EDs are introduced in England, there is an opportunity to avoid reproducing these structural gaps in new settings. Our study suggests that improving physical healthcare for patients with mental health conditions requires changes to how EDs are designed, resourced and supported, and how they connect with the wider health and care system. Keywords: mental health, diagnostic inequality, emergency departments

Background Bystander automated external defibrillator (BAED) use improves out-of-hospital cardiac arrest (OHCA) outcomes but remains uncommon globally. This study evaluated the outcomes of Singapore's 11-year public-access AED expansion and volunteer-responder implementation in terms of trends in BAED use, associated factors, and clinical outcomes. Methods This population-based, retrospective cohort study used Singapore Pan-Asian Resuscitation Outcomes Study (SG-PAROS) data (2010-2020) for adult, non-traumatic OHCAs. The primary outcome was bystander AED application. Multivariable logistic regression identified factors associated with use. Secondary outcomes included favorable neurological status (CPC 1-2), survival to discharge, and prehospital return of spontaneous circulation (ROSC). Results Of 21,439 included OHCA cases (median age 70.0 years; 63.8% male), BAED use increased from 1.7% to 9.6% over 11 years, with a corresponding increase in overall survival from 2.4 to 4.0%. Malay ethnicity (aOR 1.25, 1.06-1.49), calendar year (aOR 1.26, 1.22-1.29), and delayed emergency medical services (aOR 1.24, 1.06-1.45) were positive predictors of BAED use. Conversely, BAED use was lower among females (aOR 0.80, 95% CI 0.69-0.94), at night (aOR 0.69, 0.56-0.86), and in residential settings (aOR 0.06, 0.05-0.07). Volunteer arrival strongly increased application (aOR 4.16, 3.41-5.09), with a significant interaction (p<0.001); the effect was greater in residential (aOR 7.38, 5.81-9.38) than non-residential settings (aOR 1.71, 1.22-2.40). AED use predicted favorable neurological outcome (aOR 2.80, 2.24-3.50; NNT 8.7), survival (aOR 2.30, 1.89-2.80), and ROSC (aOR 2.11, 1.81-2.46). Conclusion Over 11 years, we saw a significant increase in BAED application and favorable neurological survival. This success was associated with the implementation of an integrated strategy combining widespread AED deployment, national training, and smartphone-activated volunteer responders. Singapore's experience provides a scalable model for urban centers seeking to expand their AED strategy.

Background: In crises, hospitals must rapidly shift from routine operations to structured crisis management, requiring the activation of an incident command system. However, empirical insight into their operational functioning during activation remains limited. Goal: to identify operational enablers and barriers influencing effective crisis response. Methods: Prospective cross-sectional, qualitative, single-center study conducted after a table-top exercise within a hospital incident command system at a tertiary care university hospital (NCT06913010). Data was collected through semi-structured interviews, participant observation, and document analysis, and analyzed using a narrative-phenomenological approach. Results: Nineteen participants were included. Analysis identified nine thematic clusters shaping operational performance: (1) structure and roles; (2) communication; (3) decision-making and prioritization; (4) information management; (5) infrastructure and technology; (6) personnel and organization; (7) training, exercises, and team dynamics; (8) documentation; and (9) external communication and media. Enablers included clear role definition, structured communication, phased decision-making, and regular training. Barriers included role ambiguity, fragmented communication, insufficient prioritization, infrastructure limitations, and staffing constraints. Conclusion: Preparedness frameworks are necessary but insufficient as stand-alone approaches, as operational execution determines real-world performance. Recurring deficits included unclear roles, inconsistent communication, weak prioritization, and gaps in infrastructure and personnel. A limited set of standardized practices - including a clear separation od roles, leadership intent, closed-loop communication, explicit decision cycles from information gathering to structuring to decision-making, checklists, visualization, central information management, and rapid "80% decisions"-substantially enhanced performance. Mission command (Auftragstaktik) further enabled adaptive, coordinated action. Strengthening hospital incident command is a key lever for achieving system-level resilience in crises.

BackgroundEsophageal pressure (Pes) measurement has been used successfully over the past half-century to delineate the respiratory systems physiology and mechanics. However, there is no information about the importance of Pes monitoring in different scenarios. We aimed to assess the importance and frequency of Pes monitoring in different scenarios according to health professionals and its importance in decision-making. MethodCross-sectional study with an international survey. We included healthcare professionals dedicated to patients receiving invasive and non-invasive ventilation without limits of age, gender, experience and seniority in the position or country of residence. We used non-probabilistic snowball sampling. ResultsWe included 152 participants, with 54.61% (83) males. The response rate to the survey questions ranged from 100% to 71.71%. Of the included participants, 91/139 (65.47%) were respiratory therapists, and 31/139 (22.30%) were Physicians. Most participants worked in mixed ICU. 109/121 (90.08%) participants considered Pes monitoring very important or extremely important for teaching or research. Only 32/112 (28.57%) reported using Pes frequently for these proposals. 49/109 (40.50%) participants considered Pes monitoring very important or extremely important during non-invasive ventilatory support. Only 17/112 (15.18%) reported using Pes frequently for these proposals. Regarding MV individualisation in ARDS during total ventilatory support, 94/121 (77.69%) participants considered Pes monitoring very important or extremely important. Only 33/112 (29.46%) reported using Pes frequently in this scenario. 90/121 (74.38%) also considered it very important or extremely important for MV individualisation in obese patients without ARDS, and 108/121 (89.26%) considered it very important or extremely important for MV individualisation in obese patients with ARDS during total ventilatory support. Only 25/112 (22.32%) and 39/112 (34.82%) reported using Pes frequently in these scenarios, respectively. ConclusionsPes monitoring was considered very important or extremely important for most assessed scenarios. Conversely, most participants rarely or never used it, although it changed therapeutic decisions often when implemented.

Introduction -- Emergency Medical Dispatch Systems (EMDS) can reduce delays in accessing emergency care by providing structured communication, triage, and coordination. However, such systems remain absent or underdeveloped in most low- or middle-income countries (LMICs). This study aimed to establish international consensus on essential EMDS components to inform global guidance. Methods -- We convened a multidisciplinary expert group to draft a preliminary list of essential components for three EMDS levels reflecting resource availability and system maturity. We then conducted a three-round Delphi with international experts to reach consensus on core EMDS components. Components which had [&ge;]75% agreement were included, those with [&ge;]75% disagreement were excluded. Components not achieving consensus by Round 3 were removed. Results were analysed overall and stratified by respondents' country income level. A subsequent online expert meeting resolved inconsistencies and finalised the component list. Results -- The expert group generated 111 components for each of three EMDS levels (Foundational, Emerging, and Established) spanning 11 operational domains. Of the 68 experts invited to the Delphi, 43 participated in Round 1 and 30 in Round 3. Across all Delphi rounds, 289 components reached consensus for inclusion. The consensus resulted in a final list of 227 components (63 Foundational, 84 Emerging, and 80 Established). Consensus agreement clustered around core EMDS domains including communication, structured call-taking and prioritisation, advice-giving, resource dispatch and tracking, and foundational governance and data functions, whereas items showing either non-consensus or consensus disagreement were typically technology-dependent or context-specific. Conclusions -- This international consensus offers guidance for EMDS development across diverse resource settings and provides a scalable roadmap to strengthen emergency care systems.

BackgroundThe COVID-19 pandemic exposed critical shortages of mechanical ventilators, particularly in low-resource settings. Disruptions in global supply chains and dependence on specialized components highlighted the need for scalable, locally manufacturing alternatives for emergency respiratory support. AimTo describe and evaluate a simplified, supply-chain-independent mechanical ventilator assembled from widely available automotive and simple hardware components, and intended as a last-resort solution. MethodsThe ventilator is based on a reciprocating air pump driven by an automotive windshield wiper motor coupled to parallel shaft bellows and readily assembled passive membrane valves, only requiring materials available from standard hardware retailers, minimal tools, and basic manual skills. Ventilator performance was assessed through bench testing using a patient model simulating severe lung disease in an adult (R=20 cmH2O{middle dot}s/L, C=15 mL/cmH2O) and pediatric (R=50 cmH2O{middle dot}s/L, C=10 mL/cmH2O) patients. Realistic proof of concept was performed in four mechanically ventilated 50-kg pigs. ResultsThe device delivered tidal volumes up to 600 mL and respiratory rates up to 45 breaths/min with PEEP up to 10 cmH2O, covering pediatric and adult ventilation ranges. In vivo testing showed that the ventilator maintained arterial blood gases within the targeted range. Technical details for ventilator construction are provided in an open-source video tutorial. DiscussionThis low-cost ventilator demonstrated adequate performance under demanding conditions. Although not a substitute for commercial intensive care ventilators, its simplicity, autonomy, and independence from fragile supply chains provide a potentially life-saving option in resource-constrained emergency scenarios.

Abstract Background: Emergency Department (ED) populations are a high-risk group that are opportune for interventions targeting NCDs and NCD risk factors, like tobacco use. Mobile health (mHealth)interventions such as Text2Quit, a novel text message-based mHealth tool addressing tobacco cessation in the US, have demonstrated effectiveness for tobacco cessation and for ED-based mHealth interventions in High Income Countries (HIC). To successfully adapt and implement such mHealth interventions in limited resource settings like African EDs, it is essential to examine the implementation climate and engage key stakeholders. These implementers provide invaluable insight to understand healthcare system level factors that affect adoption, implementation and maintenance of the interventions. Methods: We conducted 12 semi-structured key informant interviews (KIIs) with ED administrators and staff including 2 departmental heads, 5 medical doctors, 3 nurses, and 2 clinical officers at a national referral hospital in Kenya. This was guided by RE-AIM framework indicators of Adoption, Implementation, and Maintenance (eg feasibility of intervention integration, and suggestions to improve implementation). Interviews were conducted in English, recorded, professionally transcribed and translated, and analyzed using a constant comparative analysis approach, according to grounded theory principles. Findings: Key informants were positive about the adoption of them Health intervention in Kenyan EDs and across different health facility levels in Kenya due to the perceived need for the program, facility and staff receptiveness and existing healthcare infrastructure to leverage. Recommended implementation strategies included follow-up mechanisms for program participants, inclusion of all healthcare cadres in implementation and increased sensitization and the use of champions. Barriers to Implementation in the ED included competing clinical priorities with emergency cases, limited staffing and shame associated with smoking. Conclusion: Implementing a mobile health tobacco cessation program like Text2Quit is feasible and acceptable in Kenyan EDs when supported by targeted strategies.

Background: Stroke is a time-sensitive neurological emergency in which early EMS activation and presentation to definitive care are cornerstones of effective therapy. Large language models (LLMs) are increasingly consulted by the public for medical advice, but the veracity of the guidance provided by commercially available models responding to potential stroke symptoms is not well understood. Methods: We performed a cross-model benchmarking study comparing the triage choices of three frontier LLMs (Claude Sonnet 4.6, GPT-4o, and Llama 3.3-70b-versatile) on first-person vignettes describing a unilateral arm symptom on waking, across 10 symptom descriptors, and two clinical phases (before and after a partially reassuring self-examination), with or without a clinical distractor (n=50 per condition). Results: Claude sought emergency care most often, Llama least, and GPT-4o in between, diverging most sharply in the post-examination phase where Claude called 911 in 100% of runs, Llama called for non-emergency help in 100%, and GPT-4o was symptom-dependent. A distractor shifted behavior away from emergency care in almost all conditions: calling 911 fell from 37.9% to 14.6% and waiting rose from 0% to 45.9% in the post-examination vignette. Responses were also sensitive to symptom word: weak, limp, heavy, and clumsy generated higher alarm, whereas numb, tingly, odd, strange, and weird generated less urgent responses. Conclusions: The increasing use of LLMs for medical advice has significant public health implications. Commercially available LLMs show significant model-to-model variability and framing sensitivity when confronted with potential stroke symptoms, including under-recognition of canonical CDC warning descriptors, underscoring the need for systematic benchmarking as these tools become de facto first points of contact for patients experiencing neurological emergencies.

BackgroundSouth Africas public healthcare system serves most of the population through approximately 3,900 primary healthcare clinics characterised by long waiting times and high volumes of repeat-prescription visits. No published pre-arrival digital triage system operates across all 11 official South African languages while aligning with the South African Triage Scale (SATS). This paper reports the design and preliminary safety validation of BIZUSIZO, a hybrid deterministic-AI WhatsApp triage system. MethodsBIZUSIZO delivers SATS-aligned triage via WhatsApp, combining AI-assisted free-text classification (Claude Haiku 4.5) with a Deterministic Clinical Safety Layer (DCSL) that overrides AI output for 53 clinical discriminator categories (14 RED, 19 ORANGE, 20 YELLOW) coded in all 11 official languages and independent of AI availability. A five-domain risk factor assessment can only upgrade triage level. One hundred and twenty clinical vignettes in patient language (English, isiZulu, isiXhosa, Afrikaans; 30 per language) were scored against a developer-assigned gold standard with independent blinded nurse review. A 121-vignette multilingual DCSL safety consistency check across all 11 languages and a 220-call post-hoc framing sensitivity evaluation (110 paired vignettes) were also conducted. ResultsUnder-triage was 3.3% (4/120; 95% CI: 0.9%-8.3%) with no RED under-triage; exact concordance was 80.0% (96/120) and quadratic weighted kappa 0.891 (95% CI: 0.827-0.932). One two-level under-triage was observed on a non-RED presentation (V072, isiXhosa burns vignette, ORANGEGREEN); one two-level over-triage was observed (V054, isiZulu deep laceration, YELLOWRED). In the framing sensitivity evaluation, AI-only classification achieved 50.9% RED invariance under adversarial framing; full-pipeline classification achieved 95.0% in four validated languages, with the DCSL rescuing 18 of 23 AI drift cases. ConclusionsA hybrid deterministic-AI triage system with DCSL-based emergency detection achieved zero RED under-triage and consistent RED detection across all 11 official languages. The 16.7% over-triage rate falls within published South African SATS ranges (13.1-49%). A single two-level under-triage event was observed on an isiXhosa burns vignette (ORANGEGREEN) and is discussed in Limitations. Findings are preliminary; prospective validation against independent nurse triage is the necessary next step.

ObjectivesFor suspected hip fractures, prehospital protocols directing patients to an orthopaedic centre rather than the nearest emergency department (ED) could reduce time-to-surgery but may impact EMS travel burden. This study evaluates the impact of transfer protocols by quantifying transport to hospitals from long term care (LTC) facilities across Ontario. MethodsA retrospective cross-sectional analysis of all Ontario LTC facilities and hospitals was performed. Two protocols were modeled: standard transfer to the nearest ED with subsequent transfer if required, and selective transfer based on Collingwood Hip Fracture Rule prehospital screening1directly to the nearest orthopaedic services (orthoED). Median one-way travel distances were calculated from Google Maps. ResultsIn Ontario, 15.4% of LTC residents require hospital destination decisions because their nearest ED lacks orthopaedic services; for these facilities, median distances were 2.7km to the ED and 36.0km to the orthoED. Among the 52 LTC facilities where selective transfer was distance-optimal, it substantially reduced travel for patients with hip fracture (31.1km vs 49.6km; P<.01) while only modestly increasing travel for patients without hip fracture. Where standard transfer was distance-optimal, little travel difference was noted for patients with hip fracture, however false positive screened patients traveled significantly further to an orthoED. Greatest negative consequences of selective transfer lie in the 1.3% of residents living farthest (>100km) from an orthoED. ConclusionsEMS direct transportation to hospitals with orthopaedics may improve hip fracture care but can increase EMS burden due to patients identified falsely as having a hip fracture, particularly in remote communities.

Background and aimsWe evaluated the efficacy of the Stroke Riskometer mobile phone App to change the Lifes Simple 7(R) (LS7(R)) risk factor score at 6 months post-randomisation. Methods and designThis Phase III, prospective, outcome assessor-blinded, 2-arm randomised controlled trial (RCT) in Australia and New Zealand recruited participants from August 2021 to January 2024. Inclusion criteria: age [&ge;]35 and [&le;]75 years; [&ge;]2 risk factors; smartphone ownership; no cardiovascular disease history. The intervention group was given access to the App; the usual care group received one e-mail with generic risk factor information. The primary outcome was the mean between group difference in LS7(R) (score 0 [poor] to 14 [ideal] comprising blood pressure, cholesterol, glucose, body mass index, smoking, physical activity and diet) from baseline to 6 months post-randomisation. Secondary outcomes were between group changes in individual LS7 items. Analyses were performed using intention to treat (ITT) principles with ANCOVA and linear mixed models to examine differences between groups, with pre-specified per protocol and subgroup analyses. ResultsWe randomised 862 participants (mean {+/-} SD age 58{+/-}11 years; 63% women; 74% Caucasian). At 6 months post-randomisation in ITT analyses, the mean difference between usual care (n=433) and intervention (n=429) groups in the change in LS7(R) score from baseline was 0.03 (95% CI -0.19, 0.25, p=0.79). Per protocol analyses (n=320 usual care; n=276 intervention) were similar (mean difference in change 0.11 95% CI -0.12, 0.34, p=0.34). Compared to usual care in ITT analyses, the intervention group had a borderline increase in metabolic equivalent of task (MET) minutes/week of physical activity (313.42 95% CI -2.80, 629.65, p=0.05), with no differences in other LS7(R) items. DiscussionAmong a general population aged 35 to 75 years with [&ge;]2 stroke risk factors, there was no evidence that having access to the App changed overall LS7(R) scores at 6-month follow-up. Participants in the intervention group did have a small increase in physical activity, compared to the usual care group after 6 months, but not other individual risk factors.

Importance: Abdominal pain causes roughly 10 million US emergency department (ED) visits annually, most resulting in discharge. Post-discharge courses vary, yet existing risk models predict only whether an ED revisit occurs, not what that revisit outcome will entail. Objective: To evaluate whether Curiosity, a generative medical event foundation model, can predict post-ED-discharge trajectories for adults with abdominal pain, differentiating the timing and severity of expected outcomes. Design: Retrospective cohort study; encounters January 1-December 31, 2022; 30-day follow-up; analysis conducted in 2026. Setting: Epic Cosmos research network (multicenter, population-based, de-identified electronic health record). Participants: Adults ([&ge;]18 years) discharged from the ED with abdominal pain, excluding training-set patients. Random sample of 3,000 drawn from 150,030 eligible patients (65.3% female; median age 47 years [IQR 36-60]). Exposure: ED discharge after evaluation for abdominal pain. Main Outcomes and Measures: Primary: Curiosity model vs. per-task, separately estimated XGBoost models on area under the receiver operating characteristic curve (AUROC) for ED revisit ending in admission (admit-revisit), ED revisit ending in discharge (DC-revisit), and any ED revisit at 72 hours, 7 days, and 30 days. Secondary: trajectory-level accuracy across 36 trajectory classes and edit distance vs XGBoost; calibration of simulated vs observed conditional path probabilities across 45 transitions. Results: Curiosity identified patients at high risk of revisit requiring admission more accurately than XGBoost and differentiated those likely to revisit without admission. Among 3,000 patients, Curiosity's 30-day admit-revisit AUROC was 0.83 (95% CI 0.79-0.87) vs 0.70 (95% CI 0.65-0.75) for XGBoost (DeLong P<.001), and admit-revisit AUC-PR was 0.37 (95% CI 0.29-0.46) against a 4.1% cohort base rate, vs XGBoost 0.13 (95% CI 0.09-0.19). Curiosity identified the most likely trajectory out of 36 possibilities for 45.9% of patients (XGBoost 41.0%; McNemar P<.001), with median edit distance 1.28 vs 1.40 (Wilcoxon P<.001). Median absolute calibration error across 45 transitions was 1.30 percentage points (95% CI 0.32-2.49). Conclusions and Relevance: A generative medical event foundation model produced calibrated trajectory-level predictions and discriminated admit-revisits more effectively than task-specific XGBoost baselines, separating patients that revisited and were admitted from those who revisited and were discharged.

Background High Intensity Use of urgent medical services by patients is widely recognised in urgent and emergency care. Studies of high intensity use of the emergency department have consistently shown features of complex systems behaviour in addition to highly heterogeneous individual patient characteristics. There have been no comparable studies of prehospital care use. Methods We examined the use of prehospital urgent and emergency services (NHS 111 and ambulance dispatch) using routinely collected data from regional service in the UK (population 5 million). We used a complex systems perspective, to examine (1) distribution of contacts per individual; (2) the temporal stability of service use by individuals and at the whole-system level (3) the distribution of bursts of contacts. Results We analysed data from 847555 individuals who contacted NHS111 and 389550 who contacted the ambulance dispatch service. 35120 (4.2%) individuals who contacted NHS111 had 5 or more contacts with the service over the two-year period and accounted for 290625 (20.1%) of contacts. 16755 (4.3%) individuals had 5 or more ambulance dispatch contact days and accounted for 169085 (25.8%) of contacts. The distribution of contacts per individual showed a monotonic distribution between 5 and over 100 contacts that was heavy tailed and compatible with a power law distribution. At any level of use, patients with one or more mental health related contacts had a greater likelihood of further contact than those without. Conclusion Prehospital emergency service use shows multiple statistical features typical of a complex system. Interventions to manage demand need to consider both individual high intensity users (particularly in relation to their mental health) and the behaviour of the whole system.

BackgroundPatient safety remains a global priority, yet adverse events persist due to gaps in communication, information, training and safety culture. Rapid response systems standardise observation models are widely used to recognise deterioration and guide escalation and response for ward patients in hospitals. A notable gap concerns the role of planning for further care, can improve hospital resource prioritisation as healthcare professionals respond to patients deterioration in daily practice. Engaging healthcare professionals as key stakeholders to ensure relevance, we identified unanswered research questions on hospital patient safety and rapid response systems and prioritised the top ten research needs. Aim and methodsWe conducted a hospital-tailored, modified James Lind Alliance Priority Setting Partnership (JLA PSP) with healthcare professionals as key stakeholders to identify and prioritise rapid response system related patient safety research needs and evidence uncertainties. The modified JLA process included five stages: (1) establish the Priority Setting Partnership; (2) identify uncertainties; (3) summarise and refine submissions with evidence checks. (4) priority setting; and (5) verify and finalise a top ten list, with evidence checks and project-group oversight throughout. ResultsA modified JLA PSP resulted in the stakeholders co-producing a list of research priorities. The top three priorities addressed implementation strategies, intervention effectiveness, and optimising hospital patient safety through clinical protocols and rapid response system activation thresholds. Additional priorities addressed ethical, educational, and organisational factors, highlighting evidence gaps which recognised and responded to patient deterioration and the need for safer transitions across levels of hospital care. The modified JLA PSP was feasible for co-producing a clinically relevant, practice-oriented research agenda. ConclusionsA transparent, systematic, stakeholder-driven process generated hospital patient safety research priorities for rapid response systems that reflect stakeholder needs and target key evidence gaps guiding future research and strengthening patient safety practice in hospitals and, in primary care.

Background Complicated appendicitis (CA) increases morbidity and resource use.[1,2] In the emergency setting, risk stratification must rely on rapidly available data. Procalcitonin (PCT) is frequently obtained, but its incremental value beyond basic preoperative indicators remains uncertain.[5] We aimed to quantify PCTs incremental predictive value and develop a practical bedside score with temporal validation. Methods We conducted a retrospective cohort study of consecutive laparoscopic appendectomy patients (January 2023-December 2024). CA was defined by postoperative pathology (gangrene/necrosis, perforation, or peri-appendiceal inflammation/abscess; worst-category rule). We compared a base logistic model (age, WBC, neutrophil percentage, fever, symptom-to-surgery interval, shock index) with an extended model adding log-transformed PCT. Discrimination (AUC) and calibration were assessed. Temporal validation used 2023 for development and 2024 for testing. We also created a simple bedside score using pre-specified cutoffs and evaluated CA risk across score strata in 2024. Results In the overall complete-case cohort (n=1,792), 397 patients (22.2%) had CA. Adding PCT modestly improved discrimination in the full cohort (AUC 0.673 to 0.685). For temporal validation, 2023 included 870 patients (CA 26.9%) and 2024 included 921 patients (CA 17.7%); one otherwise eligible patient lacked a usable admission year. In the 2024 test set, discrimination was 0.662 (base) vs 0.673 (base+PCT) with a non-significant AUC difference (DeLong p=0.116); calibration slopes were near 1.0. A 7-item bedside score stratified 2024 CA risk: 9.1% (score 0-1), 14.7% (2-3), and 34.2% [&ge;]4). Using [&ge;]4 points identified a higher-risk subgroup (PPV 34.2%, NPV 87.5%, sensitivity 46.0%, specificity 81.0%). Conclusions PCT adds modest predictive information beyond simple preoperative indicators in the full cohort, but temporal validation suggests that this incremental gain is smaller and not statistically significant in later patients. A pragmatic bedside score can support CA risk stratification and prioritization in emergency care, whereas the role of routine PCT testing may be best reserved for selected situations in which uncertainty remains after initial assessment.

Background: Large language models (LLMs) demonstrate strong performance in controlled medical environments such as multiple choice exams, but their utility in real-world clinical workflows remains unproven. The NHS Advice & Guidance (A&G) service, where Primary Care clinicians can submit text-based queries to specialists, provides an environment for evaluating the clinical performance of LLMs as a specialist. Methods: We compared responses from MedGemma 4B-IT, an open-weight model deployed locally on hospital infrastructure, against specialist neurologist responses across 50 adult neurology A&G cases from University College London Hospital. Two neurologists and two GPs rated 80 blinded and 20 unblinded responses for outcome, safety, efficacy, and feasibility using standardised criteria; outcome was a binary correct/incorrect, while other domains were scored 1-5. Inter-rater reliability was assessed using intraclass correlation coefficients. Results: Although there were no statistically significant differences between blinded specialist neurologists and LLM responses across any domain (outcome: 84% vs 82%, p=0.67; safety: 3.98 vs 4.02, p=0.85; efficacy: 4.06 vs 3.98, p=0.61; feasibility: 4.39 vs 4.20, p=0.45), 10% of LLM responses received concerning scores ([&le;]2 average score) compared to 0% of human responses, indicating potentially clinically important tail risk. Furthermore, unblinded results showed a preference for human responses, with human ratings being preferred across all domains. Only 51% of binary outcomes had unanimous agreement and inter-rater agreement was moderate across other domains (ICC 0.50-0.52). Conclusions: In this pilot study, aggregate scores between blinded human and LLM responses were similar, and no statistically significant differences were detected in this exploratory sample. However, aggregate metrics masked clinically important edge-case failures in LLM responses. Pronounced inter-rater variability and the potential impact of LLM/human syntax on blinded rater judgements highlight the challenges in establishing robust evaluation frameworks for clinical LLM deployment

Importance: Severe stroke is a leading cause of death and disability worldwide. Survivors and their families face long-term unmet needs, including care that does not reflect patients' values, fragmented care, and high rates of psychological distress among caregivers. Objective: To describe the conceptual framework of the longitudinal transdisciplinary neuropalliative care support (LOTUS) intervention and assess its fidelity in a pilot feasibility study. Design: Pilot feasibility randomized study; fidelity was assessed using weekly checklists completed by the LOTUS nurse and qualitative analysis of weekly LOTUS team meeting transcripts. Setting: Single comprehensive stroke center in Western New York. Participants: Patients hospitalized with severe stroke and their caregivers. Dyads were randomized to usual care or intervention. Intervention: The LOTUS intervention is implemented in a stepped-care fashion using 5 strategies: Awareness, Assistance, Adjustment, Acceptance and Alignment (5As). Led by a specially trained nurse with a chaplain, social worker, psychologist, and neuropalliative care physician, the LOTUS team follows dyads from early in the hospital course through 6 months. Main Outcomes and Measures: Fidelity, the degree to which the intervention was delivered as intended, assessed via (1) utilization of 5A activities from weekly LOTUS checklists; (2) thematic analysis of weekly LOTUS team meeting transcripts. Results: Of 26 patients in the trial, 13 were randomized to intervention. The LOTUS nurse completed 108 checklists, with an average of 619 minutes of direct contact per participant over 6 months. Each component of the 5A's was utilized. Awareness and Assistance predominated early after enrollment and revolved around personhood, support, and self-efficacy. Adjustment was especially relevant during care transitions and was typically supported by the LOTUS social worker. Acceptance and Alignment were more prevalent during later meetings, with the LOTUS psychologist supporting identification and modeling of coping skills and the LOTUS physician guiding prognosis and goals-of-care conversations. The LOTUS nurse served as primary point of contact, providing continuity and a trusting relationship, while other team members functioned in a predominantly advisory role. Conclusions: The LOTUS intervention was delivered with fidelity to the 5A-framework, supporting a future randomized clinical trial to evaluate its efficacy in patients with severe stroke and their caregivers.