Carlos Darcy Alves Bersot

Large language models (LLMs) are increasingly used for data extraction in systematic reviews, but reported performance varies across extraction tasks, models, domains, reference standards, and definitions of correctness. We systematically reviewed empirical studies of automated or AI-assisted data extraction for evidence synthesis and performed an estimand-driven meta-analysis of uncorrected LLM extraction against human or otherwise validated reference standards. Forty-six reports representing 45 independent study units were included in the systematic review. Eight independent studies contributed 10 model-level estimates and 29,643 item-level evaluations to the broad primary quantitative synthesis. Proportions were analyzed on the logit scale using a multilevel random-effects model with model-level estimates nested within study and cluster-robust inference. Pooled item-level accuracy was 86.4% (cluster-robust 95% CI, 62.6%–96.0%), with generalized I² of 99.6% and a model-based 95% prediction interval of 19.2%–99.4%, indicating marked between-setting variability. In a construct-restricted sensitivity analysis excluding an open-ended, recall-intensive extraction setting, seven studies contributed nine model-level estimates and 29,343 item-level evaluations; pooled accuracy was 91.2% (95% CI, 83.1%–95.6%) and the prediction interval narrowed to 66.7%–98.2%. Leave-one-study-out analyses showed that omission of individual studies other than this methodologically distinct setting had comparatively limited influence on the pooled mean, whereas heterogeneity remained substantial. Across the broader evidence base, performance varied according to extraction task, data type, workflow design, and definition of correctness. These findings indicate that average LLM extraction accuracy can be high, but performance cannot be assumed to generalize across evidence-synthesis settings. Task-specific validation and human verification remain important, particularly for ambiguous, missing, derived, numerical, or otherwise consequential information

Background and objectives: Perioperative intravenous lidocaine infusion has been proposed as an opioid-sparing analgesic adjunct in spine surgery, but previous meta-analyses pooled heterogeneous procedures and may have obscured procedure-specific treatment effects. We evaluated the analgesic and opioid-sparing effects of perioperative intravenous lidocaine in adult spine surgery and whether efficacy differs by procedure type. Methods: We searched PubMed, Embase, Web of Science, and the Cochrane Library from inception through June 2026 for randomized controlled trials comparing perioperative intravenous lidocaine infusion with placebo or standard care in adults undergoing spine surgery. The primary outcome was postoperative pain intensity at 24 hours. Secondary outcomes were opioid consumption, postoperative nausea and vomiting (PONV), and hospital length of stay (LOS). Random-effects meta-analyses, a post hoc subgroup analysis by procedure type, and meta-regression were performed. Results: 10 randomized trials were included. Intravenous lidocaine reduced postoperative pain at 24 hours (9 trials, 655 patients; mean difference (MD), −0.83; 95% CI −1.36 to −0.30; p=0.002; I²=89%, 95% CI 81% to 93%; prediction interval, −2.69 to 1.02; moderate certainty) and opioid consumption (9 trials; MD, −11.64 mg intravenous morphine equivalents; 95% CI −16.14 to −7.14; p<0.001). In a post hoc exploratory analysis, the analgesic effect differed by procedure type (test for subgroup differences, p=0.005): the reduction was clinically meaningful after instrumented fusion or complex spine surgery (MD, −1.23; 95% CI −1.81 to −0.64), exceeding the minimal clinically important difference of 1.0 point, but minimal after decompression (MD, −0.20; 95% CI −0.61 to 0.21). Baseline pain severity explained approximately 53% of between-study heterogeneity, whereas lidocaine infusion rate did not modify the treatment effect (p=0.97). No significant effects were observed for PONV or LOS. Conclusions: There is moderate-certainty evidence that perioperative intravenous lidocaine reduces postoperative pain and opioid consumption after adult spine surgery. Its analgesic benefit appeared procedure-specific, with the greatest benefit after instrumented fusion and complex procedures; this difference emerged from a post hoc, exploratory analysis and should be regarded as hypothesis-generating.

Objective: We aimed to evaluate whether deep neuromuscular blockade (DNMB) improves surgical conditions and facilitates low-pressure pneumoperitoneum compared with moderate neuromuscular blockade (MNMB) during minimally invasive surgery. Data Sources: PubMed/MEDLINE, EMBASE, Cochrane CENTRAL, Scopus, Web of Science, and LILACS were searched from inception through May 2026. Study Selection: Randomized controlled trials comparing DNMB versus MNMB in adults undergoing laparoscopic or robot-assisted surgery were included. Data Extraction: Two reviewers independently screened studies, extracted data, and assessed risk of bias using the revised Cochrane Risk of Bias tool (RoB 2). Certainty of evidence was evaluated using the GRADE approach. Primary Outcomes: Primary outcomes were surgical workspace quality and intra-abdominal pressure requirements. Secondary Outcomes: Secondary outcomes included postoperative abdominal pain at 24 h and referred shoulder pain. Data Synthesis: Seventeen randomized controlled trials involving 1512 patients were included. DNMB consistently improved surgical workspace conditions and facilitated lower pneumoperitoneum pressures compared with MNMB. Pooled analyses additionally suggested reductions in postoperative abdominal pain and referred shoulder pain in selected studies, although postoperative outcomes demonstrated greater heterogeneity across procedures and perioperative protocols. Conclusions: Current evidence suggests that the principal clinical value of DNMB during minimally invasive surgery is optimization of surgical exposure and facilitation of low-pressure pneumoperitoneum strategies. Benefits related to postoperative abdominal pain may occur in selected settings but appear less consistent across procedures.

Chronic pain is a leading cause of disability worldwide and remains a challenge for healthcare systems, medical education, and public health. Although advances in pain neuroscience and interventional therapies have improved pain management, biomedical approaches alone are insufficient to address the multidimensional nature of persistent pain or promote adaptive recovery, participation, and long-term functioning. Contemporary pain medicine therefore requires educational models integrating health promotion, resilience, relationship-centered care, and sustainable professional practice. To develop and describe a competency-based salutogenic educational framework for pain medicine integrating contemporary concepts from pain science, health promotion, behavioral science, and medical education. The framework was developed through conceptual integration of complementary theories from pain medicine, psychology, health promotion, and competency-based medical education. It combines total pain, salutogenesis and sense of coherence, resilience, Social Cognitive Theory, therapeutic alliance, and physician professional sustainability within a longitudinal educational structure. Rather than replacing existing biopsychosocial or salutogenic models, it integrates their complementary strengths by explicitly linking adaptive recovery, relationship-centered care, competency development, authentic assessment, and health promotion. Educational implications include longitudinal competency-based curricula, reflective practice, interdisciplinary learning, patient partnership, workplace-based assessment, and validated psychometric instruments alongside competency-based educational assessments to evaluate learner and patient outcomes. The proposed framework reorients pain medicine education beyond symptom management toward adaptive recovery, participation, equity, and person-centered care. By integrating health promotion with competency-based medical education, it provides a conceptual foundation for curriculum development, implementation research, and evaluation of educational strategies to improve outcomes for patients, healthcare professionals, and health systems.

Thoracic surgery is associated with severe acute and chronic pain, high opioid use, and other adverse outcomes. Gabapentin is commonly used perioperatively, but its effectiveness remains uncertain. This systematic review and meta-analysis evaluated the efficacy and safety of perioperative gabapentin versus placebo after thoracic surgery. Cochrane, Embase, PubMed, and Web of Science were searched through 29 August 2025 for randomized controlled trials comparing gabapentin with placebo in adults undergoing thoracic surgery. The primary outcome was pain at 24 hours. Secondary outcomes included pain at other time points, opioid consumption, chronic pain, and adverse events. Data were pooled using random-effects models. Six trials including 748 patients were analyzed. Gabapentin had no significant effect on rest pain at 24 hours (MD -0.13; 95% CI -1.17 to 0.91). A non-significant trend toward worse cough-evoked pain at 24 hours was observed with gabapentin (MD 0.53; 95% CI 0.00 to 1.07). Gabapentin also did not significantly affect opioid consumption, pain chronification, or adverse events such as nausea, sedation, or delirium. Based on low-certainty evidence, perioperative gabapentin does not appear to provide meaningful benefit after thoracic surgery. The www.crd.york.ac.uk/prospero identifier is CRD420251136259.

Objective: To evaluate whether deep neuromuscular blockade (DNMB) improves surgical conditions and facilitates low-pressure pneumoperitoneum compared with moderate neuromuscular blockade (MNMB) during minimally invasive surgery. Data Sources PubMed/MEDLINE, EMBASE, Cochrane CENTRAL, Scopus, Web of Science, and LILACS were searched from inception through May 2026. Study Selection Randomized controlled trials comparing DNMB versus MNMB in adults undergoing laparoscopic or robot-assisted surgery were included. Data Extraction Two reviewers independently screened studies, extracted data, and assessed risk of bias using the revised Cochrane Risk of Bias tool (RoB 2). Certainty of evidence was evaluated using the GRADE approach. Primary Outcomes: Primary outcomes were surgical workspace quality and intra-abdominal pressure requirements. Secondary Outcomes: Secondary outcomes included postoperative pain at 24 hours and referred shoulder pain. Data Synthesis: Seventeen randomized controlled trials involving more than 1700 patients were included. DNMB consistently improved surgical workspace conditions and facilitated lower pneumoperitoneum pressures compared with MNMB. Pooled analyses additionally suggested reductions in postoperative pain and referred shoulder pain in selected studies, although postoperative outcomes demonstrated greater heterogeneity across procedures and perioperative protocols. Conclusions: Current evidence suggests that the principal clinical value of DNMB during minimally invasive surgery is optimization of surgical exposure and facilitation of low-pressure pneumoperitoneum strategies. Benefits related to postoperative pain may occur in selected settings but appear less consistent across procedures.

The coexistence of glucose-6-phosphate dehydrogenase (G6PD) deficiency and methemoglobinemia may produce misleading pulse oximetry readings, creating a significant diagnostic challenge during the perioperative assessment of oxygenation. Methemoglobinemia results from the oxidation of hemoglobin iron to the ferric state, impairing oxygen delivery and altering light absorption, often leading to falsely low oxygen saturation values. A 20-year-old man with confirmed G6PD deficiency and suspected chronic methemoglobinemia underwent an elective electrophysiologic study for recurrent supraventricular tachycardia. On arrival in the operating room, pulse oximetry showed a saturation of 45% on room air, despite preserved clinical status. Arterial blood gas analysis revealed a partial pressure of oxygen (PaO₂) exceeding 400 mmHg, confirming adequate arterial oxygenation and establishing a saturation gap. General anesthesia was induced and maintained with propofol and remifentanil without complications. Due to the risk of hemolysis associated with methylene blue in G6PD deficiency, conservative management was adopted, with hyperbaric oxygen therapy available as a contingency strategy. The intraoperative course was uneventful, with stable hemodynamics and no evidence of tissue hypoxia or hemolysis. This case demonstrates that severe pulse oximetry desaturation does not necessarily reflect true hypoxemia. Recognition of a saturation gap is essential when oximetry findings are discordant with clinical presentation, particularly in patients with G6PD deficiency, in whom standard therapies may be contraindicated. Accurate evaluation requires the integration of clinical assessment with arterial blood gas analysis. This report highlights the diagnostic pitfalls and implications for safe perioperative management.

Closed-loop vasopressor systems integrate real-time blood pressure monitoring with automated decision logic to support hemodynamic stability in perioperative and critical care environments. These technologies sit at the intersection of biomedical sensing, signal processing, and clinician-supervised automation: the quality, latency, and failure behavior of the monitoring input can directly shape controller performance, safety margins, and clinical usability. In this comprehensive review, we synthesize the major closed-loop vasopressor architectures reported in the literature, examine how sensor modality and signal integrity influence algorithm behavior, and summarize recurrent reliability vulnerabilities spanning sensors, control logic, and device integration. We organize the field through an end-to-end information pipeline—monitoring input, signal conditioning and quality assessment, decision and control strategy, actuation via infusion technology, and supervisory safety layers—highlighting common performance metrics used to benchmark control quality. We then discuss clinical validation patterns across settings, emphasizing practical considerations for deployment and the evidence gaps that remain most relevant to high-risk populations. Finally, we propose reporting and validation priorities for future studies, with a focus on sensor robustness, transparency of algorithm design, integration safeguards, and standardized documentation of failures and overrides.