Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 10 appraisals
Critical care (London, England)
Large language models for optimizing clinical trial recruitment in ICUs: application to ventilator-induced diaphragm dysfunction
BACKGROUND: Ventilator-induced diaphragm dysfunction (VIDD) is a frequent and under-recognized consequence of prolonged mechanical ventilation in intensive-care unit (ICU) patients. Identifying eligible candidates for clinical trials targeting VIDD remains a major operational challenge. This study evaluates the use of large language models (LLMs) to automate patient prescreening from ICU discharge summaries and estimate recruitment capacity for a future phase 2 trial. METHODS: We developed an LLM-based prescreening pipeline to assess trial eligibility criteria from ICU discharge summaries, which was deployed to screen all 2024 ICU stays. Stays that were flagged as potentially eligible underwent expert adjudication. An enriched set of 50 ICU stays was independently annotated by six clinicians to define a reference standard, which was used to evaluate criterion-level model performances using F1-scores. RESULTS: The best-performing model was GPT-OSS:120B with a criterion-level F1-score of 0.82. When applied to consecutive 1,342 ICU stays from Montpellier University Hospital in 2024, the selected model identified 532 patients with ≥ 3 days of mechanical ventilation. After applying exclusion criteria, 185 patients remained potentially eligible. Expert review confirmed 133 patients as eligible, resulting in a positive predictive value of 72% (95% CI 65-78). The LLM-assisted workflow resulted in an estimated 86% reduction in clinician review time. The LLM-based prescreening pipelines achieved criterion-level F1 scores ranging from 0.73 to 0.82, with GPT-OSS:120B demonstrating the highest performance. CONCLUSIONS: LLM-based prescreening offers a promising approach for identifying trial candidates in critical care, prioritizing candidates for clinician review. Future deployments should include targeted expert validation and ongoing monitoring to ensure safety and generalizability.
13 June 2026
Read appraisal →Spine
Identifying Predictors of Failed Back Surgery Syndrome Following Lumbar Spine Surgery: A Machine Learning Approach
STUDY DESIGN: Retrospective cohort study from a tertiary academic medical center. OBJECTIVE: To build a prognostic machine learning model to predict 1-year FBSS incidence after lumbar spine surgery. SUMMARY OF BACKGROUND DATA: A minority of patients who undergo degenerative lumbar spine surgery will have persistent postoperative pain, characterized as "Failed Back Surgery Syndrome" (FBSS). Adequate preoperative identification of patients at risk of having an undesirable outcome after surgery is an essential part of a spine surgeon's workflow. Although several studies have proposed mechanisms and risk factors for FBSS, no studies have developed a prognostic machine learning model to quantify and functionalize predictions. METHODS: A cohort of lumbar fusion and lumbar decompression surgeries was queried from a tertiary academic medical center from 2002 to 2022. Patient and operative characteristics were systematically extracted for each surgery. Several machine learning algorithms were used and optimized to predict FBSS occurrence within 1 year of surgery. SHAP feature importance values were computed for the top-performing model. RESULTS: A total of 10,128 unique lumbar decompression surgeries and 2890 unique lumbar fusion surgeries were included. The Random Forest model had the highest performance of tested models (AUROC of 0.715 for lumbar decompression, 0.701 for lumbar fusion). For lumbar decompression, the top three predictors of FBSS were absence of microdiscectomy, lack of preoperative immunosuppressant usage, and preoperative benzodiazepine usage. For lumbar fusion, prior FBSS diagnosis, lack of preoperative immunosuppressant usage, and operating room duration were the most important predictors. Other key variables spanned several domains, including preoperative medication usage, patient demographics, and operative indications and characteristics. CONCLUSION: This study demonstrates the successful creation of a prognostic machine learning model for prediction of FBSS within one year postoperatively. These models, after external validation, have the potential to be instrumental aspects of a spine surgeon's workflow. LEVEL OF EVIDENCE: Level III.
16 May 2026
Read appraisal →BMJ open
Effect of early antiplatelet therapy on 30-day mortality after coronary artery bypass grafting: a retrospective target trial emulation using the MIMIC-IV database
OBJECTIVES: The effect of early antiplatelet therapy on 30-day mortality after coronary artery bypass grafting (CABG) remains uncertain. This study aimed to evaluate the association between early antiplatelet therapy initiation and clinical outcomes in post-CABG patients. DESIGN: A retrospective target trial emulation. SETTING: Single academic medical centre in Boston, Massachusetts, USA. PARTICIPANTS: We included adult patients who underwent CABG and were admitted to the intensive care unit at Beth Israel Deaconess Medical Center between 2008 and 2022. METHODS: Patients were classified into the treatment group if they initiated antiplatelet therapy within 24 hours postoperatively; otherwise, they were assigned to the control group. To address immortal-time bias for the primary outcome of 30-day mortality, we used clone-censor-weighting. For the exploratory outcome of 30-day acute kidney injury (AKI) incidence, we employed inverse probability of treatment weighting to balance baseline covariates. RESULTS: The analysis for 30-day mortality included 6887 patients. The mortality was 2.35% in the control group versus 1.44% in the treatment group, yielding an absolute risk difference of 0.90% (control group minus treatment group, 95% CI 0.09% to 1.72%). The difference in restricted mean time lost was 0.19 days (control group minus treatment group, 95% CI 0.03 to 0.36). For AKI incidence, 5543 patients were analysed. The incidence was 78.61% in the control group and 69.13% in the treatment group, with an absolute difference of 9.48% (control group minus treatment group, 95% CI 6.40% to 12.50%; p<0.001). CONCLUSIONS: In this target trial emulation, early antiplatelet therapy initiated within 24 hours after CABG was associated with lower 30-day mortality and a reduced incidence of postoperative AKI. These findings warrant confirmation in prospective randomised controlled trials.
24 Apr 2026
Read appraisal →Journal of investigative surgery : the official journal of the Academy of Surgical Research
Acupoint Stimulation After Anesthesia Induction Does Not Decrease Nausea and Vomiting After Hepatectomy: A Multi-Center Double-Blinded Clinical Trial
INTRODUCTION: This study evaluated whether adding transcutaneous electrical acupoint stimulation (TEAS) to a double-antiemetic regimen during hepatectomy reduces the incidence of postoperative nausea and vomiting (PONV). METHODS: In this multicenter double-blinded randomized controlled trial conducted across three tertiary hospitals in China, patients scheduled for elective hepatectomy were randomized to either the TEAS group (stimulation at PC6 during surgery) or the control group (no stimulation). The primary outcome was the incidence of PONV within 24 h postoperatively. Key secondary outcomes included the severity of PONV and pain after surgery. RESULTS: Among 161 randomized patients, 159 completed the study. No significant difference was observed in the incidence of PONV within 24 h postoperatively between the TEAS and control groups (23.8% vs. 26.6%; OR, 95% CI: 0.86, 0.42-1.76). However, the TEAS group exhibited less patient controlled analgesia (PCA) attempts (2.7 vs 3.5) or deliveries (2.6 vs 3.5) by 24 h postoperatively. Recovery outcomes, including time to ambulation and to gastrointestinal function recovery, did not differ between groups. CONCLUSION: Intraoperative TEAS did not significantly reduce PONV incidence in hepatectomy patients receiving antiemetic therapy. Nonetheless, its potential to decrease the need for analgesics warrants further investigation to refine its role in perioperative care. CLINICAL TRIAL REGISTRATION: ClinicalTrial.gov (NCT05396716), https://clinicaltrials.gov.
21 Apr 2026
Read appraisal →Acta orthopaedica
Intraoperative single-dose methadone significantly affects postoperative morphine consumption in older patients with a hip fracture: the MetaHip randomized controlled trial
BACKGROUND AND PURPOSE: Effective pain management in older patients with a hip fracture is critical for postoperative recovery. Our primary objective was to compare intraoperative methadone with placebo on postoperative morphine consumption over 72 hours. METHODS: Patients aged ≥ 60 years with hip fractures were randomized to receive methadone (0.10 mg/kg) or placebo intraoperatively. The primary outcome was postoperative morphine consumption measured in 24-hour intervals over 72 hours. Secondary outcomes included pain scores, time to mobilization, and discharge. Harms were assessed as adverse and serious adverse events. RESULTS: 129 patients were included. The primary endpoint analysis demonstrated that postoperative morphine consumption over 72 hours differed significantly between groups (likelihood-ratio test, P = 0.02). Model-based estimates suggested lower morphine consumption in the methadone group at 0-24 hours (least square mean [LSM] 7.1 [SE 1.2] vs placebo 10.1 [SE 1.7] mg) and at 24-48 hours (4.1 [SE 0.8] vs placebo 5.3 [SE 0.9] mg). At 48-72 hours, the model suggested lower morphine consumption in the placebo group (3.2 [SE 0.6] vs methadone 4.6 [SE 0.9] mg). Secondary outcomes were similar between groups, except that time to hospital discharge was longer in the methadone group (LSM 5.6 vs 4.5 days; mean difference -1.3 days, 95% confidence interval -2.3 to -o.4; P < 0.01). Harms appeared comparable, although the low event rate precluded formal analysis. CONCLUSION: A single intraoperative dose of methadone significantly alters postoperative morphine consumption over 72 hours after hip fracture surgery, without major safety concerns.
17 Apr 2026
Read appraisal →JAMA
Restrictive vs Liberal Physical Restraint Strategies in Critically Ill Patients: The R2D2-ICU Randomized Clinical Trial
IMPORTANCE: The effect of wrist-strap physical restraints on outcomes in patients receiving mechanical ventilation in the intensive care unit (ICU) remains uncertain. OBJECTIVE: To investigate the effect of a low-use wrist-strap physical restraint strategy in critically ill patients receiving invasive mechanical ventilation. DESIGN, SETTING, AND PARTICIPANTS: Open-label randomized clinical trial conducted across 10 ICUs in France. Between January 5, 2021, and January 2, 2024, 405 adult patients who had initiated invasive mechanical ventilation within the previous 6 hours and were expected to require ventilation for at least 48 hours were enrolled. Follow-up was completed on May 17, 2024. Statistical analysis was conducted from June 1, 2025, to December 15, 2025. INTERVENTIONS: Patients were randomized to undergo either a restrictive, low-use physical restraint strategy (wrist straps avoided unless necessary because of severe agitation, defined as a Richmond Agitation-Sedation Scale score of ≥3 [on a scale from -5 (unresponsive) to 4 (combative)]; n = 201) or a liberal, high-use strategy (wrist straps applied systematically and reassessed daily; n = 204). Discontinuation of restraints was allowed in patients who were awake or extubated without delirium (measured via the Confusion Assessment Method for the ICU). MAIN OUTCOMES AND MEASURES: The primary outcome was the number of days alive without coma or delirium during the first 14 days after randomization. Secondary outcomes included incidence of self-extubation and day-90 mortality. RESULTS: Among 396 patients with available primary outcome data, the median (IQR) age was 65 (56-73) years, 245 (62%) were male, and the median (IQR) Sequential Organ Failure Assessment score was 7 (4-10). The mean days alive without coma or delirium were 6.67 days (95% CI, 5.69-7.65) in the low-use strategy group and 6.30 days (95% CI, 5.35-7.24) in the high-use strategy group (adjusted mean difference, 0.37 days [95% CI, -0.71 to 1.46]; P = .51). Self-extubation occurred in 18 patients (9.2%) in the low-use strategy group and 17 (8.5%) in the high-use strategy group, and day-90 mortality was 37.2% and 41.0%, respectively. CONCLUSIONS AND RELEVANCE: In this randomized clinical trial, among adult patients receiving mechanical ventilation in the ICU, a low-use wrist-strap physical restraint strategy compared with a high-use strategy did not reduce days free of delirium or coma at 14 days. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT04273360.
15 Apr 2026
Read appraisal →Current pain and headache reports
Stellate Ganglion Block in the Treatment of Long COVID: A Systematic Review
OBJECTIVES: This review evaluates stellate ganglion block as a treatment for long COVID, seeking to evaluate the treatment's efficacy by various symptoms and the limitations of the current literature. STUDY DESIGN: Systematic Review. SETTING: Ambulatory or Outpatient Setting. METHODS, SUBJECTS: A systematic review of the current literature regarding use of stellate ganglion block in patients with long COVID was conducted. 2 databases were searched on August 28th, 2025. Search terms were "long COVID" and "stellate ganglion block", yielding 45 results. Studies examining patient outcomes after stellate ganglion block were included. Case reports, case series, basic science studies and previous reviews were excluded. Seven studies met inclusion criteria. RESULTS: Patients received a single stellate ganglion block in some studies and multiple stellate ganglion blocks in others. All studies reported symptomatic improvement without control groups. Response rates ranged from 55.8% to 100%. The most robust improvements (> 80% patients reporting relief) were seen in cough, dyspnea, headache, joint pain, pain interference/intensity, pins/needles, subjective relief. CONCLUSION: Stellate ganglion block is a promising treatment that appears to generate substantive benefit for many of the symptoms seen in long COVID. However, the current literature has small, uncontrolled studies with heterogenous study designs and follow-up periods. Standardized research with larger sample sizes, control groups, and longer-term follow up is necessary to elucidate the degree of benefit. IRB approval and clinical trial registration not required.
15 Apr 2026
Read appraisal →Supportive care in cancer : official journal of the Multinational Association of Supportive Care in Cancer
Effectiveness of prehabilitation strategies in enhancing recovery and postoperative outcomes in laparoscopic sleeve gastrectomy patients
OBJECTIVE: This study aims to explore the application of a multimodal prehabilitation strategy in patients undergoing laparoscopic sleeve gastrectomy (LSG) and to evaluate its impact on postoperative recovery, weight loss, and quality of life. METHODS: This was a single-center, prospective, randomized controlled trial. A total of 120 eligible patients undergoing elective LSG were enrolled and randomly assigned to either the control group (n = 60), receiving routine perioperative care, or the observation group (n = 60), receiving a structured 5-7-day prehabilitation program. RESULTS: Compared to the control group, the observation group had a significantly shorter postoperative hospital stay (mean difference [MD] -1.23 days; 95% CI -1.67 to -0.79; P < 0.001) and a lower incidence of complications (1.7% vs. 11.7%; odds ratio 0.13, 95% CI 0.02-0.89; P = 0.038). At 6 months, the prehabilitation group showed a greater reduction in BMI (MD -3.78 kg/m2; 95% CI -5.10 to -2.46; P < 0.001) and higher quality of life scores across all eight SF-36 domains (P < 0.001 for all). CONCLUSION: The application of prehabilitation strategies in patients undergoing laparoscopic sleeve gastrectomy can effectively promote faster postoperative recovery, enhance weight loss, improve quality of life, and reduce the incidence of postoperative complications. It is a safe and effective approach, worthy of clinical promotion. Furthermore, the principles of this multimodal prehabilitation framework offer valuable translational insights for optimizing perioperative supportive care in surgical oncology.
12 Apr 2026
Read appraisal →Current opinion in anaesthesiology
Real-time, artificial intelligence-guided intraoperative resuscitation and fluid management in trauma anesthesia
PURPOSE OF REVIEW: To synthesize recent advances in intraoperative resuscitation for trauma surgery, including fluid composition, transfusion thresholds, and coagulopathy management, and to identify emerging directions that address persistent gaps in precision. Particular attention is given to prospective evidence, as forward-looking evaluations better capture real-world performance, workflow integration, and clinical impact, culminating in the integration of artificial intelligence and machine learning as the next stage of innovation. RECENT FINDINGS: Major trials over the past 5 years have refined evidence for balanced transfusion ratios, restrictive transfusion thresholds, and the transition from crystalloid to hemostatic and physiology-guided strategies. However, outcome variability underscores the need for individualized, data-driven resuscitation. Early prospective machine learning evaluations - along with the US Food and Drug Administration-cleared point-of-care systems such as APPRAISE-HRI - demonstrate that multimodal physiologic modeling can predict hemorrhage and fluid responsiveness as accurately as expert clinicians, while offering real-time monitoring and guidance. SUMMARY: The field is shifting from protocol-based to precision-guided trauma resuscitation. Future progress will rely on randomized-controlled-trial-validated, adaptive, human-in-the-loop artificial intelligence workflows that integrate continuous physiologic data with clinical judgment to enhance the timing, composition, and safety of intraoperative fluid and transfusion therapy.
4 Apr 2026
Read appraisal →Critical care clinics
Innovation, Technology, and Telemedicine in Critical Care
Critical care medicine is undergoing a major transformation driven by rapid technological innovation, digital integration, and telemedicine. This article examines how modern intensive care units (ICUs) are evolving into data-rich, precision environments equipped with advanced monitoring, organ support, and predictive analytics. Innovations such as artificial intelligence-enhanced decision support, portable ultrasound, and tele-ICU platforms are expanding diagnostic accuracy, remote access, and timely interventions. By tracing historical milestones to emerging technologies like smart ICUs and wearable biosensors, this article highlights the shift toward anticipatory, personalized care. It underscores the importance of evidence-based adoption to ensure these tools improve outcomes.
4 Apr 2026
Read appraisal →