RSS 1.0Health Care Services
Recent Submissions
-
Tackling algorithmic bias and promoting transparency in health datasets: the STANDING Together consensus recommendations.Without careful dissection of the ways in which biases can be encoded into artificial intelligence (AI) health technologies, there is a risk of perpetuating existing health inequalities at scale. One major source of bias is the data that underpins such technologies. The STANDING Together recommendations aim to encourage transparency regarding limitations of health datasets and proactive evaluation of their effect across population groups. Draft recommendation items were informed by a systematic review and stakeholder survey. The recommendations were developed using a Delphi approach, supplemented by a public consultation and international interview study. Overall, more than 350 representatives from 58 countries provided input into this initiative. 194 Delphi participants from 25 countries voted and provided comments on 32 candidate items across three electronic survey rounds and one in-person consensus meeting. The 29 STANDING Together consensus recommendations are presented here in two parts. Recommendations for Documentation of Health Datasets provide guidance for dataset curators to enable transparency around data composition and limitations. Recommendations for Use of Health Datasets aim to enable identification and mitigation of algorithmic biases that might exacerbate health inequalities. These recommendations are intended to prompt proactive inquiry rather than acting as a checklist. We hope to raise awareness that no dataset is free of limitations, so transparent communication of data limitations should be perceived as valuable, and absence of this information as a limitation. We hope that adoption of the STANDING Together recommendations by stakeholders across the AI health technology lifecycle will enable everyone in society to benefit from technologies which are safe and effective.
-
Bridging distances and enhancing care: a comprehensive review of telemedicine in surgery.Telemedicine in surgical care has undergone rapid advancements in recent years, leveraging technologies such as telerobotics, artificial intelligence (AI) diagnostics, and wearable devices to facilitate remote evaluation and monitoring of patients. These innovations have improved access to care, reduced costs, and enhanced patient satisfaction. However, significant challenges remain, including technical barriers, limited tactile feedback in telesurgery, and inequities arising from digital literacy and infrastructure gaps. The rapid integration of telemedicine in surgical care necessitates a comprehensive understanding of its advancements, challenges, and implications. This review aims to consolidate existing knowledge, identify gaps, and highlight future research directions. The COVID-19 pandemic underscored telemedicine's potential, accelerating its adoption across healthcare systems worldwide. Despite these advancements, issues such as inconsistent reimbursement policies and challenges in integrating telemedicine into existing healthcare systems hinder its widespread adoption. Future research should prioritize the integration of AI, advancements in telepresence, and solutions to socioeconomic barriers to solidify telemedicine's role in global surgical care and enhance patient safety.
-
Intervention design for artificial intelligence-enabled macular service implementation: a primary qualitative study.Background: Neovascular age-related macular degeneration (nAMD) is one of the largest single-disease contributors to hospital outpatient appointments. Challenges in finding the clinical capacity to meet this demand can lead to sight-threatening delays in the macular services that provide treatment. Clinical artificial intelligence (AI) technologies pose one opportunity to rebalance demand and capacity in macular services. However, there is a lack of evidence to guide early-adopters seeking to use AI as a solution to demand-capacity imbalance. This study aims to provide guidance for these early adopters on how AI-enabled macular services may best be implemented by exploring what will influence the outcome of AI implementation and why. Methods: Thirty-six semi-structured interviews were conducted with participants. Data were analysed with the Nonadoption, Abandonment, Scale-up, Spread and Sustainability (NASSS) framework to identify factors likely to influence implementation outcomes. These factors and the primary data then underwent a secondary analysis using the Fit between Individuals, Technology and Task (FITT) framework to propose an actionable intervention. Results: nAMD treatment should be initiated at face-to-face appointments with clinicians who recommend year-long periods of AI-enabled scheduling of treatments. This aims to maintain or enhance the quality of patient communication, whilst reducing consultation frequency. Appropriately trained photographers should take on the additional roles of inputting retinal imaging into the AI device and overseeing its communication to clinical colleagues, while ophthalmologists assume clinical oversight and consultation roles. Interoperability to facilitate this intervention would best be served by imaging equipment that can send images to the cloud securely for analysis by AI tools. Picture Archiving and Communication Software (PACS) should have the capability to output directly into electronic medical records (EMR) familiar to clinical and administrative staff. Conclusion: There are many enablers to implementation and few of the remaining barriers relate directly to the AI technology itself. The proposed intervention requires local tailoring and prospective evaluation but can support early adopters in optimising the chances of success from initial efforts to implement AI-enabled macular services.
-
Burn center function during the COVID-19 pandemic: an international multi-center report of strategy and experienceThe novel coronavirus, SARS-CO V2 responsible for COVID-19 pandemic is rapidly escalating across the globe. Burn centers gearing for the pandemic must strike a balance between contributing to the pandemic response and preserving ongoing burn care in a safe and ethical fashion. The authors of the present communication represent seven burn centers from China, Singapore, Japan, Italy, Spain, the United Kingdom (UK), and the United States (US). Each center is located at a different point along the pandemic curve and serves different patient populations within their healthcare systems. We review our experience with the virus to date, our strategic approach to burn center function under these circumstances, and lessons learned. The purpose of this communication is to share experiences that will assist with continued preparations to help burn centers advocate for optimum burn care and overcome challenges as this pandemic continues.
-
Prognostic value of strain by speckle tracking echocardiography in patients with arrhythmogenic right ventricular cardiomyopathyBackground Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a rare genetic disorder associated with an elevated risk of life-threatening arrhythmias and progressive ventricular impairment. Risk stratification is essential to prevent major adverse cardiac events (MACE). Our study aimed to investigate the incremental value of strain measured by two-dimensional speckle-tracking echocardiography in predicting MACE in ARVC patients compared to conventional echocardiographic parameters. Methods and Results This was a retrospective, single-centre cohort study of 83 patients with ARVC (51% males, median age 37 years (IQR: 23, 53)) under the care of the Inherited Cardiac Conditions clinic at University Hospital Birmingham. MACE was defined as one of the following: sustained ventricular tachycardia (Sus VT), ventricular fibrillation (VF), appropriate implantable cardio-defibrillator (ICD) therapy [shock/anti-tachycardia pacing (ATP)], heart failure (defined as decompensated heart failure, cardiac index by heart catheter, HF medication, and symptoms), cardiac transplantation, or cardiac death. Echocardiography images were analysed by a single observer for right ventricle (RV) and left ventricular (LV) global longitudinal strain (GLS). Multivariable Cox regression was performed in combination with RV fractional area change and tricuspid annular plane systolic excursion. During three years of follow-up, 12% of patients suffered a MACE. ARVC patients with MACE had significantly reduced RV GLS (-13 ± 6% vs. -23 ± 6%, p < 0.001) and RV free wall longitudinal strain (-15 ± 5% vs. -25 ± 7%, p < 0.001) compared to those without MACE. Conclusions Right ventricular free wall longitudinal strain (RVFWLS) may be a more sensitive predictor of MACE than conventional echocardiographic parameters of RV function. Moreover, RV-free wall longitudinal strain may have superior predictive value compared to RV GLS.
-
Response to Dr DancerNo abstract available
-
Bayesian group sequential designs for phase III emergency medicine trials: a case study using the PARAMEDIC2 trialBackground: Phase III trials often require large sample sizes, leading to high costs and delays in clinical decision-making. Group sequential designs can improve trial efficiency by allowing for early stopping for efficacy and/or futility and thus may decrease the sample size, trial duration and associated costs. Bayesian approaches may offer additional benefits by incorporating previous information into the analyses and using decision criteria that are more practically relevant than those used in frequentist approaches. Frequentist group sequential designs have often been used for phase III studies, but the use of Bayesian group sequential designs is less common. The aim of this work was to explore how Bayesian group sequential designs could be constructed for phase III trials conducted in emergency medicine. Methods: The PARAMEDIC2 trial was a phase III randomised controlled trial that compared the use of adrenaline to placebo in out-of-hospital cardiac arrest patients on 30-day survival rates. It used a frequentist group sequential design to allow early stopping for efficacy or harm. We constructed several alternative Bayesian group sequential designs and studied their operating characteristics via simulation. We then virtually re-executed the trial by applying the Bayesian designs to the PARAMEDIC2 data to demonstrate what might have happened if these designs had been used in practice. Results: We produced three alternative Bayesian group sequential designs, each of which had greater than 90% power to detect the target treatment effect. A Bayesian design which performed interim analyses every 500 patients recruited produced the lowest average sample size. Using the alternative designs, the PARAMEDIC2 trial could have declared adrenaline superior for 30-day survival with approximately 1500 fewer patients. Conclusions: Using the PARAMEDIC2 trial as a case study, we demonstrated how Bayesian group sequential designs can be constructed for phase III emergency medicine trials. The Bayesian framework enabled us to obtain efficient designs using decision criteria based on the probability of benefit or harm. It also enabled us to incorporate information from previous studies on the treatment effect via the prior distributions. We recommend the wider use of Bayesian approaches in phase III clinical trials.
-
Barriers to evidence-based treatment of serious burns: the impact of implicit bias on clinician perceptions of patient adherenceThe underlying assumption of modern evidence-based practice is that treatment decisions made by healthcare providers are based solely on the best available scientific data. However, the connection between evidence informed care guidelines and the provision of care remains ambiguous. In reality, a number of contextual and nonclinical factors can also play a role, among which is the implicit bias that affects the way in which we approach or treat others based on irrelevant, individual characteristics despite conscious efforts to treat everyone equally. Influenced by the social and demographic characteristics of patients, this bias and its associated perceptions have been shown to affect clinical decision making and access to care across multiple conditions and settings. This summary article offers an introduction to how the phenomenon of implicit bias can impact on treatment compliance in multiple care contexts, its potential presence and impact in burns care and describes some of the strategies which offer possible solutions to reducing the disconnect between the conscious attempts to deliver equitable care and the discrepancies in care delivery that remain.
-
Enhancing wellbeing in medical practice: Exploring interventions and effectiveness for improving the work lives of resident (junior) doctors: A systematic review and narrative synthesis.Introduction: Globally, resident doctors face challenges like long work hours, critical decision-making stress, and exposure to death and distress, prompting concern for their wellbeing. This study addresses the need for interventions to improve their working conditions, vital for enhancing quality of life, patient care and retaining a skilled workforce. Methods: Following PRISMA guidelines, a systematic literature review until 3 January 2024 explored interventions for resident Ddoctors pre- and post-COVID-19. It evaluated intervention effectiveness, metrics and feasibility, excluding studies with high bias risk. Results: The review identified diverse interventions, from mentoring to wellness resources, showing significant improvements in job satisfaction, mental health and professional growth among resident doctors. Due to methodological variations, a narrative synthesis was conducted. Conclusion: Effective interventions addressing resident doctors' challenges can notably enhance their wellbeing and job satisfaction. Scaling such interventions is vital for fostering supportive work environments, sustaining the healthcare workforce and improving patient care quality.
-
Exploring patient and public participation in the STANDING Together initiative for AI in healthcarePublic members of STANDING Together reflect on their experience in developing standards to tackle bias in health technologies that use artificial intelligence.
-
Risk factors associated with COVID-19 severity among patients on maintenance haemodialysis: a retrospective multicentre cross-sectional study in the UKObjectives: To assess the applicability of risk factors for severe COVID-19 defined in the general population for patients on haemodialysis. Setting: A retrospective cross-sectional study performed across thirty four haemodialysis units in midlands of the UK. Participants: All 274 patients on maintenance haemodialysis who tested positive for SARS-CoV-2 on PCR testing between March and August 2020, in participating haemodialysis centres. Exposure: The utility of obesity, diabetes status, ethnicity, Charlson Comorbidity Index (CCI) and socioeconomic deprivation scores were investigated as risk factors for severe COVID-19. Main outcomes and measures: Severe COVID-19, defined as requiring supplemental oxygen or respiratory support, or a C reactive protein of ≥75 mg/dL (RECOVERY trial definitions), and its association with obesity, diabetes status, ethnicity, CCI, and socioeconomic deprivation. Results: 63.5% (174/274 patients) developed severe disease. Socioeconomic deprivation associated with severity, being most pronounced between the most and least deprived quartiles (OR 2.81, 95% CI 1.22 to 6.47, p=0.015), after adjusting for age, sex and ethnicity. There was no association between obesity, diabetes status, ethnicity or CCI with COVID-19 severity. We found no evidence of temporal evolution of cases (p=0.209) or clustering that would impact our findings. Conclusion: The incidence of severe COVID-19 is high among patients on haemodialysis; this cohort should be considered high risk. There was strong evidence of an association between socioeconomic deprivation and COVID-19 severity. Other risk factors that apply to the general population may not apply to this cohort.
-
Resuscitation with blood products in patients with trauma-related haemorrhagic shock receiving prehospital care (RePHILL): a multicentre, open-label, randomised, controlled, phase 3 trialBackground: Time to treatment matters in traumatic haemorrhage but the optimal prehospital use of blood in major trauma remains uncertain. We investigated whether use of packed red blood cells (PRBC) and lyophilised plasma (LyoPlas) was superior to use of 0·9% sodium chloride for improving tissue perfusion and reducing mortality in trauma-related haemorrhagic shock. Methods: Resuscitation with pre-hospital blood products (RePHILL) is a multicentre, allocation concealed, open-label, parallel group, randomised, controlled, phase 3 trial done in four civilian prehospital critical care services in the UK. Adults (age ≥16 years) with trauma-related haemorrhagic shock and hypotension (defined as systolic blood pressure <90 mm Hg or absence of palpable radial pulse) were assessed for eligibility by prehospital critial care teams. Eligible participants were randomly assigned to receive either up to two units each of PRBC and LyoPlas or up to 1 L of 0·9% sodium chloride administered through the intravenous or intraosseous route. Sealed treatment packs which were identical in external appearance, containing PRBC-LyoPlas or 0·9% sodium chloride were prepared by blood banks and issued to participating sites according to a randomisation schedule prepared by the co-ordinating centre (1:1 ratio, stratified by site). The primary outcome was a composite of episode mortality or impaired lactate clearance, or both, measured in the intention-to-treat population. This study is completed and registered with ISRCTN.com, ISRCTN62326938. Findings: From Nov 29, 2016 to Jan 2, 2021, prehospital critical care teams randomly assigned 432 participants to PRBC-LyoPlas (n=209) or to 0·9% sodium chloride (n=223). Trial recruitment was stopped before it achieved the intended sample size of 490 participants due to disruption caused by the COVID-19 pandemic. The median follow-up was 9 days (IQR 1 to 34) for participants in the PRBC-LyoPlas group and 7 days (0 to 31) for people in the 0·9% sodium chloride group. Participants were mostly white (62%) and male (82%), had a median age of 38 years (IQR 26 to 58), and were mostly involved in a road traffic collision (62%) with severe injuries (median injury severity score 36, IQR 25 to 50). Before randomisation, participants had received on average 430 mL crystalloid fluids and tranexamic acid (90%). The composite primary outcome occurred in 128 (64%) of 199 participants randomly assigned to PRBC-LyoPlas and 136 (65%) of 210 randomly assigned to 0·9% sodium chloride (adjusted risk difference -0·025% [95% CI -9·0 to 9·0], p=0·996). The rates of transfusion-related complications in the first 24 h after ED arrival were similar across treatment groups (PRBC-LyoPlas 11 [7%] of 148 compared with 0·9% sodium chloride nine [7%] of 137, adjusted relative risk 1·05 [95% CI 0·46-2·42]). Serious adverse events included acute respiratory distress syndrome in nine (6%) of 142 patients in the PRBC-LyoPlas group and three (2%) of 130 in 0·9% sodium chloride group, and two other unexpected serious adverse events, one in the PRBC-LyoPlas (cerebral infarct) and one in the 0·9% sodium chloride group (abnormal liver function test). There were no treatment-related deaths. Interpretation: The trial did not show that prehospital PRBC-LyoPlas resuscitation was superior to 0·9% sodium chloride for adult patients with trauma related haemorrhagic shock. Further research is required to identify the characteristics of patients who might benefit from prehospital transfusion and to identify the optimal outcomes for transfusion trials in major trauma. The decision to commit to routine prehospital transfusion will require careful consideration by all stakeholders.
-
Results of a pilot feasibility randomised controlled trial exploring the use of an electronic patient-reported outcome measure in the management of UK patients with advanced chronic kidney diseaseObjectives: The use of routine remote follow-up of patients with chronic kidney disease (CKD) is increasing exponentially. It has been suggested that online electronic patient-reported outcome measures (ePROMs) could be used in parallel, to facilitate real-time symptom monitoring aimed at improving outcomes. We tested the feasibility of this approach in a pilot trial of ePROM symptom monitoring versus usual care in patients with advanced CKD not on dialysis. Design: A 12-month, parallel, pilot randomised controlled trial (RCT) and qualitative substudy. Setting and participants: Queen Elizabeth Hospital Birmingham, UK. Adult patients with advanced CKD (estimated glomerular filtration rate ≥6 and ≤15 mL/min/1.73 m2, or a projected risk of progression to kidney failure within 2 years ≥20%). Intervention: Monthly online ePROM symptom reporting, including automated feedback of tailored self-management advice and triggered clinical notifications in the advent of severe symptoms. Real-time ePROM data were made available to the clinical team via the electronic medical record. Outcomes: Feasibility (recruitment and retention rates, and acceptability/adherence to the ePROM intervention). Health-related quality of life, clinical data (eg, measures of kidney function, kidney failure, hospitalisation, death) and healthcare utilisation. Results: 52 patients were randomised (31% of approached). Case report form returns were high (99.5%), as was retention (96%). Overall, 73% of expected ePROM questionnaires were received. Intervention adherence was high beyond 90 days (74%) and 180 days (65%); but dropped beyond 270 days (46%). Qualitative interviews supported proof of concept and intervention acceptability, but highlighted necessary changes aimed at enhancing overall functionality/scalability of the ePROM system. Limitations: Small sample size. Conclusions: This pilot trial demonstrates that patients are willing to be randomised to a trial assessing ePROM symptom monitoring. The intervention was considered acceptable; though measures to improve longer-term engagement are needed. A full-scale RCT is considered feasible. Trial registration number: ISRCTN12669006 and the UK NIHR Portfolio (CPMS ID: 36497).
-
Reporting guidelines for artificial intelligence in healthcare researchReporting guidelines are structured tools developed using explicit methodology that specify the minimum information required by researchers when reporting a study. The use of artificial intelligence (AI) reporting guidelines that address potential sources of bias specific to studies involving AI interventions has the potential to improve the quality of AI studies, through improvements in their design and delivery, and the completeness and transparency of their reporting. With a number of guidance documents relating to AI studies emerging from different specialist societies, this Review article provides researchers with some key principles for selecting the most appropriate reporting guidelines for a study involving an AI intervention. As the main determinants of a high-quality study are contained within the methodology of the study design rather than the intervention, researchers are recommended to use reporting guidelines that are specific to the study design, and then supplement them with AI-specific guidance contained within available AI reporting guidelines.
-
Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AIA growing number of artificial intelligence (AI)-based clinical decision support systems are showing promising performance in preclinical, in silico evaluation, but few have yet demonstrated real benefit to patient care. Early-stage clinical evaluation is important to assess an AI system's actual clinical performance at small scale, ensure its safety, evaluate the human factors surrounding its use and pave the way to further large-scale trials. However, the reporting of these early studies remains inadequate. The present statement provides a multi-stakeholder, consensus-based reporting guideline for the Developmental and Exploratory Clinical Investigations of DEcision support systems driven by Artificial Intelligence (DECIDE-AI). We conducted a two-round, modified Delphi process to collect and analyze expert opinion on the reporting of early clinical evaluation of AI systems. Experts were recruited from 20 pre-defined stakeholder categories. The final composition and wording of the guideline was determined at a virtual consensus meeting. The checklist and the Explanation & Elaboration (E&E) sections were refined based on feedback from a qualitative evaluation process. In total, 123 experts participated in the first round of Delphi, 138 in the second round, 16 in the consensus meeting and 16 in the qualitative evaluation. The DECIDE-AI reporting guideline comprises 17 AI-specific reporting items (made of 28 subitems) and ten generic reporting items, with an E&E paragraph provided for each. Through consultation and consensus with a range of stakeholders, we developed a guideline comprising key items that should be reported in early-stage clinical studies of AI-based decision support systems in healthcare. By providing an actionable checklist of minimal reporting items, the DECIDE-AI guideline will facilitate the appraisal of these studies and replicability of their findings.
-
Reporting guideline for the early stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AIA growing number of artificial intelligence (AI)-based clinical decision support systems are showing promising performance in preclinical, in silico, evaluation, but few have yet demonstrated real benefit to patient care. Early stage clinical evaluation is important to assess an AI system’s actual clinical performance at small scale, ensure its safety, evaluate the human factors surrounding its use, and pave the way to further large scale trials. However, the reporting of these early studies remains inadequate. The present statement provides a multistakeholder, consensus-based reporting guideline for the Developmental and Exploratory Clinical Investigations of DEcision support systems driven by Artificial Intelligence (DECIDE-AI). We conducted a two round, modified Delphi process to collect and analyse expert opinion on the reporting of early clinical evaluation of AI systems. Experts were recruited from 20 predefined stakeholder categories. The final composition and wording of the guideline was determined at a virtual consensus meeting. The checklist and the Explanation & Elaboration (E&E) sections were refined based on feedback from a qualitative evaluation process. 123 experts participated in the first round of Delphi, 138 in the second, 16 in the consensus meeting, and 16 in the qualitative evaluation. The DECIDE-AI reporting guideline comprises 17 AI specific reporting items (made of 28 subitems) and 10 generic reporting items, with an E&E paragraph provided for each. Through consultation and consensus with a range of stakeholders, we have developed a guideline comprising key items that should be reported in early stage clinical studies of AI-based decision support systems in healthcare. By providing an actionable checklist of minimal reporting items, the DECIDE-AI guideline will facilitate the appraisal of these studies and replicability of their findings.
-
Repair of acute respiratory distress syndrome by stromal cell administration (REALIST): a structured study protocol for an open-label dose-escalation phase 1 trial followed by a randomised, triple-blind, allocation concealed, placebo-controlled phase 2 trialBackground: Mesenchymal stromal cells (MSCs) may be of benefit in ARDS due to immunomodulatory and reparative properties. This trial investigates a novel CD362 enriched umbilical cord derived MSC product (REALIST ORBCEL-C), produced to Good Manufacturing Practice standards, in patients with moderate to severe ARDS due to COVID-19 and ARDS due to other causes. Methods: Phase 1 is a multicentre open-label dose-escalation pilot trial. Patients will receive a single infusion of REALIST ORBCEL-C (100 × 106 cells, 200 × 106 cells or 400 × 106 cells) in a 3 + 3 design. Phase 2 is a multicentre randomised, triple blind, allocation concealed placebo-controlled trial. Two cohorts of patients, with ARDS due to COVID-19 or ARDS due to other causes, will be recruited and randomised 1:1 to receive either a single infusion of REALIST ORBCEL-C (400 × 106 cells or maximal tolerated dose in phase 1) or placebo. Planned recruitment to each cohort is 60 patients. The primary safety outcome is the incidence of serious adverse events. The primary efficacy outcome is oxygenation index at day 7. The trial will be reported according to the Consolidated Standards for Reporting Trials (CONSORT 2010) statement. Discussion: The development and manufacture of an advanced therapy medicinal product to Good Manufacturing Practice standards within NHS infrastructure are discussed, including challenges encountered during the early stages of trial set up. The rationale to include a separate cohort of patients with ARDS due to COVID-19 in phase 2 of the trial is outlined. Trial registration: ClinicalTrials.gov NCT03042143. Registered on 3 February 2017. EudraCT Number 2017-000584-33.
-
Repair of acute respiratory distress syndrome by stromal cell administration (REALIST) trial: A phase 1 trialBackground: Mesenchymal stromal cells (MSCs) may be of benefit in acute respiratory distress syndrome (ARDS) due to immunomodulatory, reparative, and antimicrobial actions. ORBCEL-C is a population of CD362 enriched umbilical cord-derived MSCs. The REALIST phase 1 trial investigated the safety and feasibility of ORBCEL-C in patients with moderate to severe ARDS. Methods: REALIST phase 1 was an open label, dose escalation trial in which cohorts of mechanically ventilated patients with moderate to severe ARDS received increasing doses (100, 200 or 400 × 106 cells) of a single intravenous infusion of ORBCEL-C in a 3 + 3 design. The primary safety outcome was the incidence of serious adverse events. Dose limiting toxicity was defined as a serious adverse reaction within seven days. Trial registration clinicaltrials.gov NCT03042143. Findings: Nine patients were recruited between the 7th January 2019 and 14th January 2020. Study drug administration was well tolerated and no dose limiting toxicity was reported in any of the three cohorts. Eight adverse events were reported for four patients. Pyrexia within 24 h of study drug administration was reported in two patients as pre-specified adverse events. A further two adverse events (non-sustained ventricular tachycardia and deranged liver enzymes), were reported as adverse reactions. Four serious adverse events were reported (colonic perforation, gastric perforation, bradycardia and myocarditis) but none were deemed related to administration of ORBCEL-C. At day 28 no patients had died in cohort one (100 × 106), three patients had died in cohort two (200 × 106) and one patient had died in cohort three (400 × 106). Overall day 28 mortality was 44% (n = 4/9). Interpretation: A single intravenous infusion of ORBCEL-C was well tolerated in patients with moderate to severe ARDS. No dose limiting toxicity was reported up to 400 × 106 cells.
-
Refining mass casualty plans with simulation-based iterative learningBackground: Preparatory, written plans for mass casualty incidents are designed to help hospitals deliver an effective response. However, addressing the frequently observed mismatch between planning and delivery of effective responses to mass casualty incidents is a key challenge. We aimed to use simulation-based iterative learning to bridge this gap. Methods: We used Normalisation Process Theory as the framework for iterative learning from mass casualty incident simulations. Five small-scale 'focused response' simulations generated learning points that were fed into two large-scale whole-hospital response simulations. Debrief notes were used to improve the written plans iteratively. Anonymised individual online staff surveys tracked learning. The primary outcome was system safety and latent errors identified from group debriefs. The secondary outcomes were the proportion of completed surveys, confirmation of reporting location, and respective roles for mass casualty incidents. Results: Seven simulation exercises involving more than 700 staff and multidisciplinary responses were completed with debriefs. Usual emergency care was not affected by simulations. Each simulation identified latent errors and system safety issues, including overly complex processes, utilisation of space, and the need for clarifying roles. After the second whole hospital simulation, participants were more likely to return completed surveys (odds ratio=2.7; 95% confidence interval [CI], 1.7-4.3). Repeated exercises resulted in respondents being more likely to know where to report (odds ratio=4.3; 95% CI, 2.5-7.3) and their respective roles (odds ratio=3.7; 95% CI, 2.2-6.1) after a simulated mass casualty incident was declared. Conclusion: Simulation exercises are a useful tool to improve mass casualty incident plans iteratively and continuously through hospital-wide engagement of staff.
-
Real-world evaluation of rapid and laboratory-free COVID-19 triage for emergency care: external validation and pilot deployment of artificial intelligence driven screeningBackground: Uncertainty in patients' COVID-19 status contributes to treatment delays, nosocomial transmission, and operational pressures in hospitals. However, the typical turnaround time for laboratory PCR remains 12-24 h and lateral flow devices (LFDs) have limited sensitivity. Previously, we have shown that artificial intelligence-driven triage (CURIAL-1.0) can provide rapid COVID-19 screening using clinical data routinely available within 1 h of arrival to hospital. Here, we aimed to improve the time from arrival to the emergency department to the availability of a result, do external and prospective validation, and deploy a novel laboratory-free screening tool in a UK emergency department. Methods: We optimised our previous model, removing less informative predictors to improve generalisability and speed, developing the CURIAL-Lab model with vital signs and readily available blood tests (full blood count [FBC]; urea, creatinine, and electrolytes; liver function tests; and C-reactive protein) and the CURIAL-Rapide model with vital signs and FBC alone. Models were validated externally for emergency admissions to University Hospitals Birmingham, Bedfordshire Hospitals, and Portsmouth Hospitals University National Health Service (NHS) trusts, and prospectively at Oxford University Hospitals, by comparison with PCR testing. Next, we compared model performance directly against LFDs and evaluated a combined pathway that triaged patients who had either a positive CURIAL model result or a positive LFD to a COVID-19-suspected clinical area. Lastly, we deployed CURIAL-Rapide alongside an approved point-of-care FBC analyser to provide laboratory-free COVID-19 screening at the John Radcliffe Hospital (Oxford, UK). Our primary improvement outcome was time-to-result, and our performance measures were sensitivity, specificity, positive and negative predictive values, and area under receiver operating characteristic curve (AUROC). Findings: 72 223 patients met eligibility criteria across the four validating hospital groups, in a total validation period spanning Dec 1, 2019, to March 31, 2021. CURIAL-Lab and CURIAL-Rapide performed consistently across trusts (AUROC range 0·858-0·881, 95% CI 0·838-0·912, for CURIAL-Lab and 0·836-0·854, 0·814-0·889, for CURIAL-Rapide), achieving highest sensitivity at Portsmouth Hospitals (84·1%, Wilson's 95% CI 82·5-85·7, for CURIAL-Lab and 83·5%, 81·8-85·1, for CURIAL-Rapide) at specificities of 71·3% (70·9-71·8) for CURIAL-Lab and 63·6% (63·1-64·1) for CURIAL-Rapide. When combined with LFDs, model predictions improved triage sensitivity from 56·9% (51·7-62·0) for LFDs alone to 85·6% with CURIAL-Lab (81·6-88·9; AUROC 0·925) and 88·2% with CURIAL-Rapide (84·4-91·1; AUROC 0·919), thereby reducing missed COVID-19 cases by 65% with CURIAL-Lab and 72% with CURIAL-Rapide. For the prospective deployment of CURIAL-Rapide, 520 patients were enrolled for point-of-care FBC analysis between Feb 18 and May 10, 2021, of whom 436 received confirmatory PCR testing and ten (2·3%) tested positive. Median time from arrival to a CURIAL-Rapide result was 45 min (IQR 32-64), 16 min (26·3%) sooner than with LFDs (61 min, 37-99; log-rank p<0·0001), and 6 h 52 min (90·2%) sooner than with PCR (7 h 37 min, 6 h 5 min to 15 h 39 min; p<0·0001). Classification performance was high, with sensitivity of 87·5% (95% CI 52·9-97·8), specificity of 85·4% (81·3-88·7), and negative predictive value of 99·7% (98·2-99·9). CURIAL-Rapide correctly excluded infection for 31 (58·5%) of 53 patients who were triaged by a physician to a COVID-19-suspected area but went on to test negative by PCR. Interpretation: Our findings show the generalisability, performance, and real-world operational benefits of artificial intelligence-driven screening for COVID-19 over standard-of-care in emergency departments. CURIAL-Rapide provided rapid, laboratory-free screening when used with near-patient FBC analysis, and was able to reduce the number of patients who tested negative for COVID-19 but were triaged to COVID-19-suspected areas.