Brayan Miranda-Chavez

Objectives: To identify predictors of all-cause 30-day mortality in bloodstream infections (BSI) caused by carbapenem-susceptible Escherichia coli and Klebsiella pneumoniae, with emphasis on bacterial species and the extended-spectrum β-lactamase (ESBL) phenotype. Methods: This was a single-centre retrospective cohort study of adults with a first episode of BSI due to carbapenem-susceptible E. coli or K. pneumoniae at a referral hospital in Lima, Peru (June 2020–December 2023). Primary outcome: All-cause 30-day mortality. A multivariable Cox model was built, with covariate selection guided by a directed acyclic graph, and the species × ESBL interaction was tested. Results: Of the 403 patients (median age 66 years, 55.8% men; 58.6% ESBL+), 83 (20.6%) died. Mortality was higher in patients with K. pneumoniae than in those with E. coli (26.8% versus 17.4%; log-rank P = 0.018). In the multivariable analysis, K. pneumoniae was an independent predictor [adjusted HR (aHR) 1.72; 95% CI, 1.09–2.73; P = 0.021], as was a Charlson score ≥6 (aHR 8.43; 95% CI, 3.33–21.3; P < 0.001). ESBL was not an independent predictor (aHR 1.45; 95% CI, 0.92–2.29; P = 0.113), nor was the species × ESBL interaction (P = 0.43). The effect of K. pneumoniae was consistent across all sensitivity analyses (aHR 1.66–1.94). Conclusions: K. pneumoniae nearly doubled mortality compared with E. coli, independently of ESBL, which did not predict mortality in this cohort of carbapenem-susceptible isolates. These findings challenge the assumption of prognostic homogeneity within carbapenem-susceptible Enterobacterales and support species stratification.

Introduction: Accurate cardiovascular risk (CVR) estimation is required in Peru. Objective: To compare the concordance between CVR calculators in patients from a chronic disease program in Peru and to analyze differences in statin indication and prescription according to each model. Methods: An analytical cross-sectional study was conducted in a primary care center in Tacna using the corrected Framingham Risk Score for Cardiovascular Disease (FRSCVD), Predicting Risk of Cardiovascular Disease EVENTs (PREVENT), Systematic COronary Risk Evaluation 2 (SCORE-2), and the PAHO/WHO calculator. Concordance was assessed using the Bland- Altman method and the kappa index. Additionally, the proportion of patients eligible for statin therapy according to each calculator was analyzed. Results: A total of 424 patients were evaluated, with a median age of 66 years; 87.7% had hypertension and 55.2% had type 2 diabetes mellitus. SCORE-2 showed the highest concordance with the corrected FRSCVD, with kappa=0.61 and 95%CI=0.46 to 0.76. The PAHO/WHO and PREVENT calculators showed lower concordance, with kappa=0.43 and 95%CI=0.33 to 0.54, and kappa=0.28 and 95%CI=0.18 to 0.38, respectively. Therefore, a signi cant therapeutic gap was identi ed, as between 47.1% and 72.9% of patients with a clinical indication for statin therapy were not receiving treatment. Conclusions: SCORE-2 showed the highest concordance with the national standard, the corrected FRSCVD, whereas PREVENT showed the greatest divergence. The variability between models and the

To date, no studies have investigated the effect of native language on depressive symptom networks. To address this gap, we examined depression symptom network structure across four major cultural-linguistic groups in Peru (Castellano, Quechua, Aymara, and Amazonian indigenous languages). We conducted a network analysis of PHQ-9 depression symptoms using nationally representative data from 31,276 Peruvian participants. Networks were estimated using ggmModSelect with Spearman correlations. Native language groups were compared using permutation tests to evaluate network differences. Depression networks exhibited distinct architectures across groups, with Castellano showing the highest connectivity (22 active edges), followed by Quechua (20 edges) and Aymara (12 edges, p < 0.019). While worthlessness/guilt-suicidal ideation was the strongest universal pathway (r = 0.377–0.44), groups exhibited distinct centrality patterns: Castellano networks centered on depressed mood, Quechua on fatigue, and Aymara on suicidal ideation. Aymara networks showed absence of anhedonia-depressed mood connections (r = 0.00) and unique psychomotor-suicidal ideation pathways (r = 0.27), which requires further exploration. It is suggested that native language may shape depression architecture, suggesting distinct cultural-linguistic patterns that challenge current depression models.

Introduction: The World Health Organization defines palliative care (PC) as an approach aimed at improving the quality of life of patients and families facing life-threatening illness. In developing countries, gaps persist in coverage, late referrals, and limited integration of PC into hospital care, especially for non-oncologic conditions. Objective: To evaluate the relationship between chronic diseases and the need for PC among hospitalized patients in a hospital in southern Peru. Methods: Analytical cross-sectional study in a public referral hospital. We included adults ⩾18 years admitted to Internal Medicine, General Surgery, Obstetrics and Gynecology, or the Emergency Department; we excluded pregnant/postpartum patients, those managed exclusively in the ICU/Trauma Shock Unit, urgent surgical cases precluding assessment, and patients with cognitive impairment (Pfeiffer ⩾2). The need for PC was assessed using the SPICT-TM-ES (⩾2 general indicators + ⩾1 clinical indicator). Sociodemographic and clinical variables were collected via interview and chart review. Associations were estimated using Poisson regression with robust variance, reporting adjusted prevalence ratios (aPR) and 95% confidence intervals (95% CI). Results: We included 150 patients; 67.3% met criteria for PC need. Multimorbidity (⩾2 comorbidities) and ⩾2 hospitalizations in the past year were associated with a higher need for PC. In adjusted models, dementia (aPR = 1.15; 95% CI: 1.05–1.26) and malignancy (aPR = 1.14; 95% CI: 1.05–1.24) were positively associated with PC need, whereas type 2 diabetes mellitus was inversely associated (aPR = 0.85; 95% CI: 0.77–0.93). Conclusion: The need for PC among hospitalized patients with chronic conditions was high.

Objective: To evaluate ChatGPT's performance onNLMEs worldwide and determine whether it could achieve licensure to practice medicine across different countries. Methods: We searched PubMed, Scopus, andGoogle Scholarforstudies evaluating ChatGPT's performance onNLMEs. Reference lists of included studies were also reviewed. Two reviewers independently screened studies and extracted the accuracy rates(performance) of GPT-3.5 and GPT-4, including those that passed thresholds, human examinee scores, and other study characteristics. The risk of bias was assessed using the JBI Critical Appraisal Checklist for Prevalence Studies. Results: We identified 37 studies evaluating ChatGPT's performance across 18 NLMEs. Most studies assessed the United States, Chinese, and Japanese examinations. While most studies used official datasets, others relied on unofficial third-party sources, and few employed advanced prompting techniques.GPT-4 wassuperiortoGPT-3.5 in allNLMEs, with accuracy rates ranging from 67% to 89%. GPT-4 passed all 18 NLMEs (100%), while GPT-3.5 passed 10 of 15 (67%). Compared to human examinees, GPT-4 outperformed the average score in 6 of 7 NLMEs (86%); the sole exception was Japan, where examinees achieved 84.9% versus 81.5% for GPT-4. Conclusion: Current evidence demonstrates that GPT-4 can pass all 18 NLMEs evaluated, surpassing human examinees in most cases. However, this finding likely reflects low passing thresholds rather than AI superiority over physicians.

BACKGROUND The growth of studies evaluating ChatGPT's performance in exams swamped the medical education community. However, it has been proved from low to high-stakes examination, affecting the reliability and validity of findings. To ensure reliability and bring a final consensus, we opted to synthesize the evidence of ChatGPT's performance under high-stakes examinations, namely, National Licensing Medical Examinations (NLME). OBJECTIVE To evaluate ChatGPT’s NLMEs performance and assess whether it could achieve a license to practice in various countries. METHODS We searched the Pubmed and Scopus databases for studies that evaluated ChatGPT's performance in NLMEs. In addition to the reference list and in Google Scholar. Studies were screened, and the accuracy rate (performance) of ChatGPT was extracted, as well as other study characteristics. RESULTS We identified 37 studies that evaluated ChatGPT's performance across 18 NLMEs. Most studies evaluated the performance of ChatGPT in the NLME of the United States, China, and Japan. While the majority of studies used official datasets, others used unofficial ones from third parties, and a scarce number of studies used prompting techniques. GPT-4 was superior to GPT-3.5 in all NLMEs and could pass all of them. GPT-4 overperformed the average performance of examinees' in most studies, except the Japan NLME. CONCLUSIONS Current evidence suggests that ChatGPT can pass 18 NLMEs, surpassing almost all candidates, and, if possible, receive a "global medical license." Further research should move towards using ChatGPT as GPT-4o in performance assessment and exploring the potential of ChatGPT for NLMEs development and validation. Moreover, our findings represent a call for reimagining assessment in medical education. CLINICALTRIAL Not applicable

What is the educational challenge? The Medical Education Research Study Quality Instrument (MERSQI) is widely used to evaluate the quality of quantitative research in medical education. It has strong evidence of validity and is endorsed by guidelines. However, the manual appraisal process is time-consuming and resource-intensive, highlighting the need for more efficient methods. What are the proposed solutions? We propose to use ChatGPT to evaluate the quality of medical education research with the MERSQI and compare its scoring with those of human evaluators. What are the potential benefits to a broader global audience? Using ChatGPT to evaluate medical education research with the MERSQI can decrease the resources required for quality appraisal. This allows faster summaries of evidence, reducing the workload of researchers, editors, and educators. Furthermore, ChatGPTs' capability to extract supporting excerpts provides transparency and may have the potential for data extraction and training new medical education researchers. What are the next steps? We plan to continue evaluating medical education research with ChatGPT using the MERSQI and other instruments to determine its feasibility in this realm. Moreover, we plan to investigate which types of studies ChatGPT performs best in.