Generative artificial intelligence (AI) has attracted enormous attention since the release of ChatGPT (1) in late 2022. Other generative AI chatbots used to generate text (including Bard, Claude, and CoPilot (2–4)) and images (including Dall-E, Midjourney, and Stable Diffusion (5–7)) have likewise seen a remarkable explosion in development and uptake. While ChatGPT has been promoted for its potential to assist users in efficiently and easily creating text to serve a wide range of purposes, this editorial focuses on the journal's opinion on the use of generative AI technology in creating articles submitted to this journal. ChatGPT utilizes a technology known as a generative pretrained transformer large language model (8). This is a form of machine learning in which very large language-based data sets are used to train computers to comprehend natural language. While natural language processing is not new, ChatGPT is currently unique in its ability to not only understand queries and information to which it has access, but also to generate new, comprehensive, and fluent language-based content. Models, such as ChatGPT, which generates language, and Dall-E, Midjourney, or Stable Diffusion, which all generate images, are collectively referred to as "generative AI", and can be seen as marking a dramatic shift in both the capabilities and widespread access to AI technology. Since its release on November 30, 2022, ChatGPT achieved the most rapid adoption of a consumer software application in history by amassing over 100 million users by January 2023 (9). Users have leveraged ChatGPT to write software, song lyrics, stories, poems, and letters. With its ability to recall previous prompts within a conversation, users can fine-tune its responses to modify the content or tone of the content it generates. The underlying model used to generate content is continuously improved through feedback from users. While users easily identified flaws in the responses provided by the earlier GPT-3 version, the model has iterated quickly. GPT-4, OpenAI's latest effort in scaling up deep learning, has been described by its creators to exhibit "human-level performance on various professional and academic benchmarks" (10). It has demonstrated impressive abilities in passing standardized examinations including the Law School Admission Test, Scholastic Aptitude Test, a unified bar exam, and even the United States Medical Licensing Exam (10). ChatGPT, and other large language model-based applications, use large data sets consisting of text available on the web. Sources may include articles, books, web-based advertising, and social media posts. ChatGPT offers several exciting and desirable potential benefits, including its potential to make completing written articles more quickly as well as completing literature review summaries. With its ability to help users write in English fluently, it has been touted as a way to help make academic publishing more equitable and diverse (11). The time saved in summarizing data and generating articles could help researchers publish their studies faster, yielding more time to work on new experimental designs, grant applications, and more. This could significantly accelerate innovation and potentially lead to breakthroughs across many disciplines. We think this technology has enormous potential; however, important risks and limitations still exist that must also be acknowledged. Using ChatGPT or other language models may lead to inaccuracies, biases, and unintended plagiarism. It has been repeatedly shown that these models may create well-written text that has little relationship to reality in a phenomenon sometimes referred to as "hallucinations" (12,13). In particular, ChatGPT may generate text complete with appropriate-sounding but completely fictional citations included (14). It is obvious that articles containing factual inaccuracies, invented citations, and plagiarized content do not meet the editorial standards of our journals. Based on a review of the current capabilities and limitations of available generative AI software, we require our contributors to follow responsible and transparent practices and policies. It is necessary that authors ensure that references cited state what is attributed to them and that the overall written document is logical and consistent with the actual findings reported in the submission. In line with the International Committee of Medical Journal Editors, we expect each submission to reflect the expertise of the author or authors, who are all ultimately responsible for every word in their submission (15). As such, generative AI software, including ChatGPT and others, may not be listed as a coauthor on any submitted article. Authors must clearly state in their article the extent to which AI technologies were used in data analysis, literature review, and article preparation. This will aid reviewers and editors in checking potential biases, inaccuracies, and improper source attribution while providing readers with a transparent view of how the article was created. We ask that authors (and other interested parties) state in their letter of submission the extent to which AI was used as well as other related information they deem pertinent. This information will help ensure the integrity of the editorial review and publication process and will serve as a learning tool for the editors. We all have a great deal to learn about generative AI and its place in academic publishing. We need to evolve together along with the technical capabilities of AI. We expect to modify our editorial policies and instructions for authors often in the coming months and years as these tools, and our understanding of their impacts and best uses continue to develop.
A 12-year-old girl presented to the emergency department with a three-day history of worsening, colicky periumbilical and lower abdominal pain. Associated symptoms included anorexia, nausea and nocturnal awakening. This was preceded by two days of upper respiratory tract infection symptoms including malaise, cough and fever.
Vital signs demonstrated fever (38.8°C), tachycardia (heart rate 148 beats/min) and tachypnea (respiratory rate 28 breaths/min), but normal blood pressure and oxygen saturation on room air. Abdominal examination demonstrated tenderness in the mid and left lateral abdomen, with palpable fullness in the left lower quadrant.
Elevated white blood cell count (15.7×109/L), platelet count (426×109/L), C-reactive protein level (17 mg/L [161.9 nmol/L]), erythrocyte sedimentation rate (16 mm/h) and normal serum albumin were noted. Abdominal ultrasound identified marked ascites and isolated bowel wall thickening of the sigmoid colon.
The patient was admitted for observation. Gastroenterology and Pediatric Surgery were consulted regarding the abdominal pain and ultrasound findings. On further history, it was discovered that the patient had a five-year history of stereotypical recurrent abdominal pain. These episodes were much milder than at the time of presentation, and they occurred up to twice per month, resolving spontaneously within two days. Further history suggested the underlying diagnosis.
Abstract The advent of social media has changed numerous aspects of modern life, with users developing and maintaining personal and professional relationships, following and sharing breaking news and importantly, searching for and disseminating health information and medical research. In the present paper, we reviewed available literature to outline the potential uses, pitfalls and impacts of social media for providers, scientists and institutions involved in digestive health in the domains of patient care, research and professional development. We recommend that these groups become more active participants on social media platforms to combat misinformation, advocate for patients, and curate and disseminate valuable research and educational materials. We also recommend that societies such as NASPGHAN assist its members in accessing training on effective social media use and the creation and maintenance of public‐facing profiles and that academic institutions incorporate substantive social media contributions into academic promotion processes.
Abstract Background Competency based medical education has become the new standard for medical education which shifts the focus of training toward a competency, rather than time-based in framework known in Canada as ‘Competence by Design’ (CBD). CBD assesses a physician trainee’s ability to demonstrate competence in CanMEDS roles via entrustable professional activities (EPAs). EPAs utilize the O-SCORE as the metric for assessing competence. This score was developed and validated for surgical/procedural subspecialties; however, CBD currently coopts this scale for both procedural and non-procedural (cognitive) EPAs. Assessor expertise has also been shown to have an important role in performance assessments, but has not been studied in the context of CBD. Aims Our study aims to assess for differences in O-SCORE utilization between cognitive and procedural EPAs, and whether assessor characteristics are associated with trends in assessment. Methods Anonymized data for all Adult GI subspecialty EPAs completed from Jun 2019 to Jan 2023 at the University of Alberta was obtained. Evaluator sex, clinical vs academic practice, advanced training expertise, and EPA score was extracted. Locally a score of 5 denotes competence, while a 1-3 indicates competence was not yet achieved. A score of 4 may be accepted as evidence of competence (neutral score), at the discretion of the local competency committee. Data was analyzed via T-tests and ANOVA with post hoc Games-Howell testing with 95% confidence intervals (CI). A p-value of ampersand:003C0.05 was significant. Results 2264 EPAs were assessed including 1385 cognitive and 879 procedural EPAs. The number of EPAs completed by evaluators ranged from 11 to 165 with a mean of 60 (standard deviation: 40). Results of O-SCORE usage is summarized in Figure 1A-B. The majority of EPAs indicate competence, with 20-25% neutral, and ampersand:003C10% did not achieve competence. Less than one of third of evaluators utilized a score of 1 or 2 across all EPAs, and zero evaluators utilized a score of 1 for cognitive EPAs. Most commonly evaluators to utilized 3/5 options of the O-SCORE. Separated by EPA type, it was most common to utilize 2/5 and 4/5 options for cognitive and procedural EPAs respectively. Results of demographic comparisons are outlined if Figure 1C-E. Male and clinical evaluators submitted higher scores on average. Hepatologists submitted higher scores than all other advanced training areas for total, cognitive, and procedural EPAs. Conclusions Across total, cognitive, and procedural EPAs there are low rates in the utilization of the whole O-SCORE scale, and our study highlights a discrepancy between procedural and cognitive EPAs. In addition, there small but significant differences in the mean EPAs score awarded between different evaluator demographics (male, clinical, hepatologists providing higher scores). Figure 1. A) Number and proportion of Entrustable Professional Activities (EPA) stratified by type and competence evaluation. B) Number and proportion of Entrustable Professional Activities (EPA) stratified by type with scored 1-5 and percent (%) of staff utilizing each score stratified by EPA type C) Number, mean, and mean difference of Entrustable Professional Activities (EPA) stratified by evaluator sex and EPA type. D) Number, mean, and mean difference of Entrustable Professional Activities (EPA) stratified by evaluator academic vs clinical status and EPA type. E) Number, mean, and mean difference of Entrustable Professional Activities (EPA) stratified by evaluator advanced training and EPA type. CI: Confidence interval; SD: standard deviation; *: pampersand:003C0.05 Funding Agencies None
Abstract Background Parenteral nutrition (PN) is essential for survival in infants with intestinal failure (IF). PN-associated cholestasis (PNAC) and liver disease (PNALD) are life-threatening complications of long-term PN use. SMOFlipid (soybean oil, medium-chain triglycerides, olive oil, and fish oil) has recently been approved as an off-label alternative to the conventional soy-based lipid emulsion (Intralipid). It is thought to have anti-cholestatic properties due to its more diverse lipid composition. Due to its’ recent approval in Canada (2013) and the USA (2016), data remains sparse. Aims We aim to determine if infants with IF receiving SMOFlipid had significantly lower rates of PNAC and improved growth compared to those receiving Intralipid. Methods All patients (≤1 year old at start of PN therapy) who received PN of any duration at two tertiary pediatric hospitals in Edmonton (2010–2018) were identified from the shared pharmacy database. Those with IF who received one type of PN continuously for ≥6 weeks total were included. Individuals with an initial serum conjugated bilirubin >50 µmol/L and/or who had PN interruptions >5 days were excluded. Data on liver parameters, growth, and complications were collected. Non-parametric tests (Mann-Whitney U test for continuous variables and χ2 test for categorical variables) were used to compare PNAC/PNALD (serum conjugated bilirubin >34umol/L during PN) and growth (weight/length/head circumference z-scores) between SMOFlipid and Intralipid. Results 1777 patients were reviewed; 40 infants (55% male), median age 4 weeks (range 0–48 weeks) at the time of PN initiation, met the inclusion criteria. Reasons for exclusion (n=1737) were receiving PN <6 weeks total (n=1485), duplicate patients (n=154), receiving multiple types of PN with each less than 6 weeks total (n=62), an initial serum conjugated bilirubin >50umol/L (n=21), more than 5 consecutive days off of PN (n=12), and older than 1 year old at time of PN initiation (n=3). Twenty-one patients (53%) received SMOFlipid, 15 (38%) Intralipid, and 4 (10%) Omegaven for ≥6 weeks. The majority (92%) were in an intensive care unit (neonatal or pediatric). No patients were septic when starting PN. Individuals received PN over a median of 7.9 weeks (range 6–27 weeks). Conclusions As expected, neonatal onset intestinal failure is rare in Edmonton. In our tertiary pediatric institutions, 2010–2018, SMOFLipid was the predominant lipid choice for infants with intestinal failure, followed by Intralipid. Omegaven was used rarely. This dataset will now allow us to compare the rates of PNAC at six weeks post-PN initiation and differences in growth between infants with IF receiving SMOFlipid versus the traditional Intralipid in our Canadian setting. Analysis is currently underway. Funding Agencies Women and Children’s Health Research Institute (WCHRI) at the University of Alberta
Abstract Background Infants with small bowel stomas (SBstoma) frequently struggle with absorption and rely on parenteral nutrition (PN). Intestinal absorption is difficult to predict based solely on intestinal anatomy. The purpose of this study was to characterize the microbiota and metabolic by‐products within stoma effluent and correlate with clinical features and intestinal absorption. Methods Prospective cohort study collecting stoma samples from neonates with SBstoma ( N = 23) or colostomy control ( N = 6) at initial enteral feed (first sample) and before stoma closure (last sample). Gut bacteriome (16S ribosomal RNA [rRNA] sequencing), short‐chain fatty acids (SCFAs) and bile acids (BAs) were characterized along with volume and energy content of a 48 h collection via bomb calorimetry (last sample). Hierarchical clustering and linear regression were used to compare the bacteriome and BAs/SCFAs, to bowel length, PN, and growth. Results Infants with ≤50% small bowel lost more fluid on average than those with >50% and controls (22, 18, 16 mL/kg/day, p = 0.013), but had similar energy losses (7, 10, 9 kcal/kg/day, p = 0.147). Infants growing poorly had enrichment of Proteobacteria compared to infants growing well (90% vs. 15%, p = 0.004). An increase in the ratio of secondary BAs within the small bowel over time, correlated with poor prognostic factors (≤50% small bowel, >50% of calories from PN, and poor growth). Conclusion Infants with SBstoma and poor growth have a unique bacteriome community and those with poor enteral tolerance have metabolic differences compared to infants with improved absorption.
Infants with intestinal failure have an increased risk of intestinal failure-associated liver disease (IFALD). Composite intravenous lipid emulsion (ILE) may reduce the risk of cholestasis. The primary outcome was to compare IFALD rates in infants with intestinal failure, between those receiving a composite ILE versus soybean oil ILE. The secondary outcome compared growth between these 2 groups.At our 2 tertiary neonatal/pediatric hospitals, we identified all patients (≤1 year old) who received ≥6 weeks parenteral nutrition (PN) from 2010 to 2018. Data included liver and growth parameters. IFALD was defined as serum conjugated bilirubin (CB) >33 µmol/L (≥2 mg/dL). Nonparametric tests were used for all comparisons.Fifty infants (35 composite ILE, 15 soybean oil ILE) were included. Those on composite ILE received PN for longer (10.1 vs 7.6 weeks; P = 0.001) and had higher baseline CB (29 vs 6.5 μmol/L; P = 0.001). No differences were found by 6 weeks (14.5 vs 5 μmol/L; P = 0.54) and by PN cessation (4 vs 4 μmol/L; P = 0.33). The proportion of patients with IFALD decreased from 54% to 20% for composite ILE, while stable given soybean oil ILE (7%). There were no differences in weight, length, or head circumference z scores ( P > 0.05).In our institutions, over 8 years, chronic intestinal failure was rare. Composite ILE was the predominant lipid choice for infants who needed longer courses of PN or had developed cholestasis. Despite longer PN duration, and higher baseline CB, overall rates of IFALD decreased with composite ILE. Regardless of parenteral lipid used, there were no differences in growth.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
