Faculty Bibliography

BACKGROUND:Febrile seizures occur in 3%-4% of US children aged six months to five years and are considered benign. However, sudden unexplained death in childhood is associated with 10 times increase in febrile seizures. We assessed the characteristics of children with febrile seizure and sudden death to identify factors that confer increased sudden death risk. METHODS:We conducted a case-control analysis of children with febrile seizure and subsequent sudden death versus living controls from December 2021 to June 2023 through an ∼10-minute anonymous online survey. We enrolled parents of children, living or deceased, whose child had experienced a febrile seizure from age six months to six years. Subjects were excluded if the child had an afebrile seizure or parents had not witnessed a febrile seizure. Demographic characteristics, parasomnias, and febrile seizure features were analyzed. RESULTS:A total of 381 completed surveys were received; 53 (14%) cases of febrile seizure with sudden death and 328 (86%) living controls. Cases reported febrile seizure onset >2 months earlier (P = 0.013) and reported developmental concerns (odds ratio [OR] = 2.32, 95% confidence interval [CI] [1.14, 4.71], P = 0.03), less frequent night awakenings (OR = 0.34, 95% CI [0.18, 0.65], P = 0.001), and less restless sleep (OR = 0.37, 95% CI [0.16, 0.85], P = 0.02). Cases were also less likely to drool (OR = 0.442, 95% CI [0.218, 0.900], P = 0.032) or be unresponsive for more than one minute (OR = 0.45, 95% CI [0.238, 0.854], P = 0.021). CONCLUSIONS:We report novel associations of febrile seizure and sudden death related to age, development, sleep, and observed ictal features. Anonymous survey methodology cannot exclude ascertainment bias and any related potential effect on results. Our findings suggest that impaired arousal mechanisms may increase risk of death in subjects with febrile seizure.

The rapid evolution of large-language models such as ChatGPT, Claude, and Gemini is reshaping the methodological landscape of obstetrics and gynecology (OBGYN) research. This narrative review provides a comprehensive account of generative AI capabilities, key use-cases, and recommended safeguards for investigators. First, generative AI expedites hypothesis generation, enabling researchers to interrogate vast corpora and surface plausible, overlooked questions. Second, it streamlines systematic reviews by composing optimized search strings, screening titles and abstracts, and identifying full-text discrepancies. Third, AI assistants can draft reproducible analytic code, perform preliminary descriptive or inferential analyses, and create publication-ready tables and figures. Fourth, the models support scholarly writing by suggesting journal-specific headings, refining prose, harmonizing references, and translating technical content for multidisciplinary audiences. Fifth, they augment peer-review and editorial workflows by delivering evidence-focused critiques. In educational settings, these models can create adaptive curricula and interactive simulations for trainees, fostering digital literacy and evidence-based practice early in professional development among clinicians. Integration into clinical decision-support pipelines is also foreseeable, warranting proactive governance. Notwithstanding these opportunities, responsible use demands vigilant oversight. Large-language models occasionally fabricate citations or misinterpret domain-specific data ("hallucinations"), potentially propagating misinformation. Outputs are highly prompt-dependent, creating a reliance on informed prompt engineering that may disadvantage less technical clinicians. Moreover, uploading protected health information or copyrighted text raises privacy, security, and intellectual-property concerns. We outline best-practice recommendations: maintain human verification of all AI-generated content; cross-validate references with primary databases; employ privacy-preserving, on-premises deployments for sensitive data; document prompts for reproducibility; and disclose AI involvement transparently. In summary, generative AI offers a powerful adjunct for OBGYN scientists by accelerating topic formulation, evidence synthesis, data analysis, manuscript preparation, and peer review. When coupled with rigorous oversight and ethical safeguards, these tools can enhance productivity without compromising scientific integrity. Future studies should quantify accuracy, bias, and downstream patient impact.

BACKGROUND:Acute appendicitis is the most common abdominal surgical emergency worldwide. Researchers from various countries have evaluated the impact of the COVID-19 pandemic on its diagnosis and treatment, as well as a possible change in its management, such as outpatient treatment. The objective of this study was to describe the characteristics of acute appendicitis treated during the COVID-19 State of Emergency (SOE) and to assess whether it could be included in an ambulatory surgery program. METHODS:Retrospective observational study involving patients treated at public hospitals for appendix diseases (K35-K38). We collected sociodemographic and clinical data. The influence of the lockdown on each variable was evaluated with a multivariate analysis. RESULTS:201 patients were included. The SOE period group comprised 78 patients, and the pre-SOE period (control) included 123 patients who met the same inclusion criteria. The risk of complicated acute appendicitis increased by 1.015 times with age (p = 0.000). The longer surgery time was associated with a 7.265 times higher risk of postsurgical complications (p = 0.000). No differences were observed with respect to drain placement (p = 0.281), although the percentage was higher in 2020 (13.9% vs 19.7%). The length of hospital stay decreased significantly during the lockdown (p = 0.017). CONCLUSION/CONCLUSIONS:The lockdown did not influence the outcome of complicated acute appendicitis from the standpoint of morbidity and mortality, reducing the hospital stay, facilitating the path towards outpatient treatment.

BACKGROUND:Online accessible patient education materials (PEMs) are essential for patient empowerment. However, studies have shown that these materials often exceed the recommended sixth-grade reading level, making them difficult for many patients to understand. Large language models (LLMs) have the potential to simplify PEMs into more readable educational content. OBJECTIVE:We sought to evaluate whether 3 LLMs (ChatGPT [OpenAI], Gemini [Google], and Claude [Anthropic PBC]) can optimize the readability of PEMs to the recommended reading level without compromising accuracy. METHODS:This cross-sectional study used 60 randomly selected PEMs available online from 3 websites. We prompted LLMs to simplify the reading level of online PEMs. The primary outcome was the readability of the original online PEMs compared with the LLM-simplified versions. Readability scores were calculated using 4 validated indices Flesch Reading Ease, Flesch-Kincaid Grade Level, Gunning Fog Index, and Simple Measure of Gobbledygook Index. Accuracy and understandability were also assessed as balancing measures, with understandability measured using the Patient Education Materials Assessment Tool-Understandability (PEMAT-U). RESULTS:The original readability scores for the American Heart Association (AHA), American Cancer Society (ACS), and American Stroke Association (ASA) websites were above the recommended sixth-grade level, with mean grade level scores of 10.7,10.0, and 9.6, respectively. After optimization by the LLMs, readability scores significantly improved across all 3 websites when compared with the original text. Compared with the original website, Wilcoxon signed rank test showed ChatGPT improved the readability to 7.6 from 10.1 (P<.001); Gemini, to 6.6 (P<.001); and Claude, to 5.6 (P<.001). Word counts were significantly reduced by all LLMs, with a decrease from a mean range of 410.9-953.9 words to a mean range of 201.9-248.1 words. None of the ChatGPT LLM-simplified PEMs were inaccurate, while 3.3% of Gemini and Claude LLM-simplified PEMs were inaccurate. Baseline understandability scores, as measured by PEMAT-U, were preserved across all LLM-simplified versions. CONCLUSIONS:This cross-sectional study demonstrates that LLMs have the potential to significantly enhance the readability of online PEMs while maintaining accuracy and understandability, making them more accessible to a broader audience. However, variability in model performance and demonstrated inaccuracies underscore the need for human review of LLM output. Further study is needed to explore advanced LLM techniques and models trained for medical content.

Undergraduate medical educators seek to optimize student learning, improve grading transparency and fairness, and provide useful information to residency programs. Recently, the United States Medical Licensing Examination's shift to pass/fail scoring for Step 1 disrupted curricular and assessment operations, and schools' tiered grading practices have been scrutinized. In noting that significant institutional time and energy were being expended in addressing the current levels of student grade appeals, 6 public and private medical schools in the Northeastern United States engaged in an examination of grade appeals via a root cause analysis (RCA). From November 2021 to April 2022, the authors reviewed specific instances of grading challenges that the team of educators encountered previously. From May to June 2022, the authors met for a facilitated discussion of the question, "Why are students challenging grading processes and systems or outcomes?" From July to October 2022, the authors identified root causes by analyzing results from the fishbone diagram (process, equipment, materials, people, and environment) and using the "five whys" technique. Several potential explanations for grade appeals and challenging grading systems across institutions were identified, including variability in the quality or experience of evaluators, lack of clarity about the goals and expectations of clerkships and a lack of transparency about the grading process, having a tiered grading system, technical issues with equipment, and clinical productivity demands of faculty. In proposing solutions to root causes identified in the RCA, factors were mapped to Liaison Committee on Medical Education (LCME) standards to facilitate quality and process improvements in grading. Aiming to support the learning environment and a fair and equivalent assessment process, the authors present a novel RCA and LCME method that can contribute to improving grading systems and has the potential to enhance learning and success.

DESIGN/METHODS:Retrospective database analysis. OBJECTIVE:This study aims to assess the RVU compensation model for the most common spine surgeries. BACKGROUND:Physician work is commonly valued by utilizing the relative value unit (RVU) model for compensation. Our aim in this study was to assess RVU and RVU per minute valuations of the most common spine procedures comparing cervical versus lumbar and fusion versus non-fusion cases. METHODS:The ACS-NSQIP database was utilized to identify the most commonly performed spine procedures. RVU and RVU per minute values were calculated and compared in each group. RESULTS:In all, 18,779 entries encompassing 28 CPT codes/code combinations were identified. Average RVUs were higher in cervical spine cases compared with lumbar cases (29 vs. 20 RVUs, respectively). Furthermore, cervical cases had significantly higher RVUs per minute compensation compared with lumbar cases (0.26 vs. 0.18 RVUs/min; P=0.01). Fusion cases also fared higher average RVUs compared with nonfusion cases (28 vs. 19 RVUs, respectively). However, when corrected for operative time, fusion cases did not differ significantly in RVUs per minute (P=0.13). CONCLUSIONS:Cervical spine surgery provides the best value for the surgeon in terms of RVUs per minute. We highlight key aspects of the compensation model in spine surgery. LEVEL OF EVIDENCE/METHODS:Level III-economic study.

OBJECTIVE:The Diabetes in Children, Adolescents, and Young Adults (DiCAYA) network seeks to create a nationwide electronic health record (EHR)-based diabetes surveillance system. This study aimed to develop a DiCAYA-wide EHR-based computable phenotype (CP) to identify prevalent cases of diabetes. RESEARCH DESIGN AND METHODS/METHODS:We conducted network-wide chart reviews of 2,134 youth (aged <18 years) and 2,466 young adults (aged 18 to <45 years) among people with possible diabetes. Within this population, we compared the performance of three alternative CPs, using diabetes diagnoses determined by chart review as the gold standard. CPs were evaluated based on their accuracy in identifying diabetes and its subtype. RESULTS:The final DiCAYA CP requires at least one diabetes diagnosis code from clinical encounters. Subsequently, diabetes type classification was based on the ratio of type 1 diabetes (T1D) or type 2 diabetes (T2D) diagnosis codes in the EHR. For both youth and young adults, the sensitivity, specificity, and positive and negative predictive values (PPV and NPV, respectively) in finding diabetes cases were >90%, except for the specificity and NPV in young adults, which were slightly lower at 83.8% and 80.6%, respectively. The final DiCAYA CP achieved >90% sensitivity, specificity, PPV, and NPV in classifying T1D, and demonstrated lower but robust performance in identifying T2D, consistently maintaining >80% across metrics. CONCLUSIONS:The DiCAYA CP effectively identifies overall diabetes and T1D in youth and young adults, though T2D misclassification in youth highlights areas for refinement. The simplicity of the DiCAYA CP enables broad deployment across diverse EHR systems for diabetes surveillance.