Reading Level Research Articles

Are large language models a useful resource to address common patient concerns on hallux valgus? A readability analysis

BackgroundThis study evaluates the accuracy and readability of Google, ChatGPT-3.5, and 4.0 (two versions of an artificial intelligence model) responses to common questions regarding bunion surgery. MethodsA Google search of “bunionectomy” was performed, and the first ten questions under “People Also Ask” were recorded. ChatGPT-3.5 and 4.0 were asked these ten questions individually, and their answers were analyzed using the Flesch-Kincaid Reading Ease and Gunning-Fog Level algorithms. ResultsWhen compared to Google, ChatGPT-3.5 and 4.0 had a larger word count with 315 ± 39 words (p < .0001) and 294 ± 39 words (p < .0001), respectively. A significant difference was found between ChatGPT-3.5 and 4.0 compared to Google using Flesch-Kincaid Reading Ease (p < .0001). ConclusionsOur findings demonstrate that ChatGPT provided significantly lengthier responses than Google and there was a significant difference in reading ease. Both platforms exceeded the seventh to eighth-grade reading level of the U.S. population. Level of evidenceN/A

Bridging the Gap: Can Large Language Models Match Human Expertise in Writing Neurosurgical Operative Notes?

Proper documentation is essential for patient care. The popularity of artificial intelligence (AI) offers the potential for improvements in neurosurgical note-writing. The study aimed to assess how AI can optimize documentation in neurosurgical procedures. Thirty-six notes were included. All identifiable data were removed. Essential information, such as perioperative data and diagnosis, was sourced from these notes. ChatGPT 4.0 was trained to draft notes from surgical vignettes using each surgeon's note template. One hundred forty-four surveys, with a surgeon or AI note, were shared with three surgeons to evaluate accuracy, content, and organization using a five-point scale. Accuracy was the factual correctness. Content was the comprehensiveness. Organization was the arrangement of the note. Flesch-Kincaid Grade Level (FKGL) and Flesch Reading Ease (FRE) scores quantified each note's readability. The mean AI accuracy (4.44) was not different from the mean surgeon accuracy (4.33, p = 0.512). The mean AI content (3.73) was lower than the mean surgeon content (4.42, p < 0.001). The mean AI organization (4.54) was greater than the mean surgeon organization (4.24, p = 0.064). The mean AI note's FKGL (13.13) was greater than the mean surgeon FKGL (9.99, p <0.001). The mean AI FRE (21.42) was lower than the mean surgeon FRE (41.70, p <0.001). AI notes were on par with surgeon notes in accuracy and organization, but lacked in content. Additionally, AI notes utilized language at an advanced reading level. These findings underscore the potential for ChatGPT to enhance the efficiency of neurosurgery documentation.

Readability and Accessibility of Patient-Education Materials for Heart Failure in the United States

Online education materials are widely used by patients and caregivers to understand the management of complex chronic diseases such as heart failure (HF). Organizations such as the American Medical Association and National Institutes of Health recommend that materials be written at a 6th grade reading level. The current study examined the readability and accessibility of online education materials for patients with HF. Whole page text from each included website was entered into an online readability calculator. Five validated readability indices (Flesch-Kincaid Grade Level, Flesch Reading Ease Scale, Gunning Fog Index, Coleman-Liau Index, and Simple Measure of Gobbledygook (SMOG Index)) were used to evaluate each source. Websites were categorized by source (government, public, and private). The availability of audiovisual accessibility features and content in non-English languages were assessed for each website. Of the 36 online resources analyzed, the median readability level was 9-10th grade by the Flesch-Kincaid Grade Level and college level using the Flesch Reading Ease Scale. The Gunning Fog Index and Coleman-Liau Index both showed median readability scores corresponding to a 12th grade reading level, while the SMOG Index showed a median score corresponding to that of the 9th grade. Only 10 websites (28%) offered information in languages other than English, and none provided comprehensive accessibility features for users with disabilities. Common online educational materials for patients with HF are characterized by a higher readability level than that recommended by the National Institutes of Health and American Medical Association with limited multilingual and accessibility options, potentially limiting the accessibility of resources to patients and caregivers.

KETERAMPILAN MEMBACA PEMAHAMAN MATERI RAGAM TEKS BERBASIS HOTS DI SD: STUDI LITERATUR

This research aims to examine the very low reading comprehension skills in Indonesia. It can be seen that the position of students' reading literacy skills in Indonesia from the results of several PIRLS surveys in 2006, IEA 2011, and PISA in 2018 can be categorized as very worrying. This international research assessment standard refers to advanced reading (comprehension) or what is often called HOTS (High Order Thinking Skills). The purpose of this article is to provide information regarding Indonesian students' ability to read comprehension with HOTS-based text variation material in Indonesian language learning, both regarding various problems that result in low reading comprehension skills and how to improve reading comprehension skills in elementary schools. This research uses qualitative research methods by studying literature or literature studies using several sources of written literature, data, and book references and other scientific journal articles. The results of the research show that reading comprehension problems are caused by incorrect comprehension orientation regarding advanced reading levels and lack of variety in learning to read questions in elementary schools so that reading comprehension skills are very low.

Analyzing Lexical Density and Readability in IELTS Band 4 Students’ Writing Test

This study explores the lexical density and readability levels of writing produced by IELTS Band 4 students. The main objectives are to assess the lexical density and readability of their writing across ten writing tests. The study employs a descriptive qualitative research method, using data collected from 10 writing tests conducted by IELTS Band 4 students. Lexical density is determined using the "Analyze My Writing" website, while readability is assessed using the Flesch Reading Ease formula and scale. The findings reveal a wide range of lexical density, ranging from 46.98% to 63.93%, indicating varying levels of content richness and complexity in the students' writing. Similarly, readability scores vary significantly, with scores ranging from 34.21 to 66.32, reflecting differences in text difficulty and ease of comprehension. These findings underscore the diverse linguistic proficiency and accessibility of writing among IELTS Band 4 students.

Leveraging Generative Artificial Intelligence Models in Patient Education on Inferior Vena Cava Filters.

Background: Inferior Vena Cava (IVC) filters have become an advantageous treatment modality for patients with venous thromboembolism. As the use of these filters continues to grow, it is imperative for providers to appropriately educate patients in a comprehensive yet understandable manner. Likewise, generative artificial intelligence models are a growing tool in patient education, but there is little understanding of the readability of these tools on IVC filters. Methods: This study aimed to determine the Flesch Reading Ease (FRE), Flesch-Kincaid, and Gunning Fog readability of IVC Filter patient educational materials generated by these artificial intelligence models. Results: The ChatGPT cohort had the highest mean Gunning Fog score at 17.76 ± 1.62 and the lowest at 11.58 ± 1.55 among the Copilot cohort. The difference between groups for Flesch Reading Ease scores (p = 8.70408 × 10-8) was found to be statistically significant albeit with priori power found to be low at 0.392. Conclusions: The results of this study indicate that the answers generated by the Microsoft Copilot cohort offers a greater degree of readability compared to ChatGPT cohort regarding IVC filters. Nevertheless, the mean Flesch-Kincaid readability for both cohorts does not meet the recommended U.S. grade reading levels.

Quality of Information About Kidney Stones from Artificial Intelligence Chatbots.

Introduction: Kidney stones are common and morbid conditions in the general population with a rising incidence globally. Previous studies show substantial limitations of online sources of information regarding prevention and treatment. The objective of this study was to examine the quality of information on kidney stones from artificial intelligence (AI) chatbots. Methods: The most common online searches about kidney stones from Google Trends and headers from the National Institute of Diabetes and Digestive and Kidney Diseases website were used as inputs to four AI chatbots (ChatGPT version 3.5, Perplexity, Chat Sonic, and Bing AI). Validated instruments were used to assess the quality (DISCERN instrument from 1 low to 5 high), understandability, and actionability (PEMAT, from 0% to 100%) of the chatbot outputs. In addition, we examined the reading level of the information and whether there was misinformation compared with guidelines (5 point Likert scale). Results: AI chatbots generally provided high-quality consumer health information (median DISCERN 4 out of 5) and did not include misinformation (median 1 out of 5). The median understandability was moderate (median 69.6%), and actionability was moderate to poor (median 40%). Responses were presented at an advanced reading level (11th grade; median Flesch-Kincaid score 11.3). Conclusions: AI chatbots provide generally accurate information on kidney stones and lack misinformation; however, it is not easily actionable and is presented above the recommended reading level for consumer health information.

Using artificial intelligence to generate medical literature for urology patients: a comparison of three different large language models

PurposeLarge language models (LLMs) are a form of artificial intelligence (AI) that uses deep learning techniques to understand, summarize and generate content. The potential benefits of LLMs in healthcare is predicted to be immense. The objective of this study was to examine the quality of patient information leaflets (PILs) produced by 3 LLMs on urological topics.MethodsPrompts were created to generate PILs from 3 LLMs: ChatGPT-4, PaLM 2 (Google Bard) and Llama 2 (Meta) across four urology topics (circumcision, nephrectomy, overactive bladder syndrome, and transurethral resection of the prostate). PILs were evaluated using a quality assessment checklist. PIL readability was assessed by the Average Reading Level Consensus Calculator.ResultsPILs generated by PaLM 2 had the highest overall average quality score (3.58), followed by Llama 2 (3.34) and ChatGPT-4 (3.08). PaLM 2 generated PILs were of the highest quality in all topics except TURP and was the only LLM to include images. Medical inaccuracies were present in all generated content including instances of significant error. Readability analysis identified PaLM 2 generated PILs as the simplest (age 14–15 average reading level). Llama 2 PILs were the most difficult (age 16–17 average).ConclusionWhile LLMs can generate PILs that may help reduce healthcare professional workload, generated content requires clinician input for accuracy and inclusion of health literacy aids, such as images. LLM-generated PILs were above the average reading level for adults, necessitating improvement in LLM algorithms and/or prompt design. How satisfied patients are to LLM-generated PILs remains to be evaluated.

Both English- and Spanish-Language Anterior Cruciate Ligament Reconstruction Online Patient Education Materials Are Written at Higher-Than-Recommended Reading Levels

PurposeThe purpose of this study is to examine the overall reading levels of anterior cruciate ligament reconstruction online patient education materials (OPEMs) written in English and Spanish. MethodsInvestigators conducted Google searches for OPEMs utilizing “ACL surgery” and “cirugía LCA” as English and Spanish search terms. Several measures of readability were used to analyze 25 English (Flesch Reading Ease, Flesch Reading Ease Grade Level, Flesch-Kincaid Grade Level, Coleman-Liau Index, Gunning-Fox Index, and the Simple Measure of Gobbledygook) and 25 Spanish (Fernández-Huerta Index, Fernández-Huerta Grade Level, and INFLESZ) OPEMs. English and Spanish OPEMs were compared based on mean overall grade level and number of OPEMs written below a seventh or ninth-grade reading level. ResultsEnglish OPEMs demonstrated a higher mean overall grade level compared to Spanish OPEMs (10.48 ± 1.86 vs 8.64 ± 1.22, p<.001). No significant differences were noted in the number of OPEMs written below a seventh-grade reading level. However, significantly more Spanish OPEMs were written below a ninth-grade reading level, compared to English OPEMs (56% vs 16%, p=.003). ConclusionAlthough Spanish OPEMs were written at a lower reading level, average readability for both English and Spanish OPEMs was significantly higher than the recommended level. Across both languages, only a single English webpage met the AMA-recommended sixth-grade reading level. More Spanish articles were written at or below the average adult reading level in the United States.

Understanding the cancer health literacy gap: NCCN member institutions’ website readability across the high mortality cancers in the US

BackgroundThe readability of cancer educational content often exceeds the average reading level of the U.S. population. This study evaluated the readability of online educational material for the five deadliest cancers in the US. MethodsWe assessed the readability of diagnosis and treatment information across 34 NCCN member institution websites. Readability was analyzed using four validated formulas: Fry Readability Score, Flesch-Kincaid Grade Level, Gunning-Fog Index, and Simple Measure of Gobbledygook. The primary outcome was the mean organizational readability level for each resource, and secondary outcomes included differences in readability between diagnosis and treatment materials and an analysis of sentence and word complexity. ResultsThe mean readability level across 34 websites was 13 ± 1.96, equivalent to a university freshman level. Treatment materials were significantly harder to read than diagnosis materials (14 ± 1.8 vs. 12 ± 1.8, p < 0.01) across all cancers. Readability varied significantly depending on the type of cancer, with material relating to the diagnosis of pancreatic cancer requiring the highest educational level (13 ± 1.9) and material relating to treating breast cancer requiring the lowest reading proficiency (12 ± 1.7). Furthermore, the analysis demonstrated that cancer diagnosis materials contained a lower median percentage of complex and long words than treatment materials but had a higher percentage of long sentences. ConclusionsThese findings underscore the need for organizations to prioritize matching the nation’s literacy levels. Despite the growing reliance on online resources for patient education, the current readability levels hinder patient comprehension

Readability analysis of ChatGPT's responses on lung cancer

For common diseases such as lung cancer, patients often use the internet to obtain medical information. As a result of advances in artificial intelligence and large language models such as ChatGPT, patients and health professionals use these tools to obtain medical information. The aim of this study was to evaluate the readability of ChatGPT-generated responses with different readability scales in the context of lung cancer. The most common questions in the lung cancer section of Medscape® were reviewed, and questions on the definition, etiology, risk factors, diagnosis, treatment, and prognosis of lung cancer (both NSCLC and SCLC) were selected. A set of 80 questions were asked 10 times to ChatGPT via the OpenAI API. ChatGPT's responses were tested using various readability formulas. The mean Flesch Reading Ease, Flesch-Kincaid Grade Level, Gunning FOG Scale, SMOG Index, Automated Readability Index, Coleman-Liau Index, Linsear Write Formula, Dale-Chall Readability Score, and Spache Readability Formula scores are at a moderate level (mean and standard deviation: 40.52 ± 9.81, 12.56 ± 1.66, 13.63 ± 1.54, 14.61 ± 1.45, 15.04 ± 1.97, 14.24 ± 1.90, 11.96 ± 2.55, 10.03 ± 0.63 and 5.93 ± 0.50, respectively). The readability levels of the answers generated by ChatGPT are "collage" and above and are difficult to read. Perhaps in the near future, the ChatGPT can be programmed to produce responses that are appropriate for people of different educational and age groups.

The Readability of Non-Financial Information. The Role of Stakeholders' Pressure in the European Setting

This study explores the factors that may explain the readability of non-financial information (NFI) disclosure through the interpretive lens of legitimacy theory. In particular, it investigates whether the readability of the NFI disclosure, prepared under Directive 2014/95/EU, is affected by the pressure of certain stakeholders and by the presence of a corporate social responsibility committee (CSRc). The sample is based on 9,218 European listed firms, covering the period 2017-2020. The results show that pressure from environmentally sensitive industries, consumers, and some investors plays a crucial role in improving the readability level of NFI, while employees and the State owner do not exercise an influence on the level of readability. In addition, the presence of a CSRc is not only relevant per se, but also acts as a moderator in enhancing the association with these stakeholders on the readability of NFI disclosure.

Caution Regarding ChatGPT's Appropriateness and Reliability Regarding Surgery for Wrist Arthritis.

Chat Generative Pre-Trained Transformer (ChatGPT), an artificial intelligence (AI) program, is widely used for information compilation. This study sought to analyze the quality and consistency of the information generated by ChatGPT regarding common procedures for wrist arthritis. 32 standardized questions regarding wrist osteoarthritis and related procedures (4-corner-fusion [4CF], proximal row carpectomy [PRC], resurfacing capitate pyrocarbon implant, wrist denervation, and total wrist arthrodesis and arthroplasty) were presented to the ChatGPT-3.5 interface 3 separate times, without feedback. ChatGPT's answers were evaluated for medical accuracy by 3 reviewers and rated as "appropriate," "appropriate but incomplete," or "inappropriate." Ratings were then converted to numerical values to calculate an intraclass correlation coefficient (ICC). A DISCERN score was used to assess quality, and Flesch-Kincade Grade Level and Flesch Reading Ease Score for readability. 75% of the responses were deemed "appropriate," with 23 questions receiving unanimous appropriate ratings across all responses. The ICC was 0.97 (95% CI [0.46, 0.98]), indicating excellent reliability. DISCERN score was 60 (good). The Flesch-Kincaid Grade Level was 14.6 ± 1.9, and the Flesch Reading Ease Score was 25.3 ± 6.7, implying a college reading level. The information that ChatGPT provided for PRC and total wrist arthrodesis and arthroplasty, appeared to be more reliable than for 4CF and denervation. ChatGPT's reliability and accuracy of information varied across procedures, possibly due to unknown and diverse sources. Furthermore, while some answers were factually correct, many provided generic information across differing questions, limiting usefulness. ChatGPT must be used cautiously and the limitations understood.

The diagnostic and triage accuracy of the GPT-3 artificial intelligence model: an observational study

Artificial intelligence (AI) applications in health care have been effective in many areas of medicine, but they are often trained for a single task using labelled data, making deployment and generalisability challenging. How well a general-purpose AI language model performs diagnosis and triage relative to physicians and laypeople is not well understood. We compared the predictive accuracy of Generative Pre-trained Transformer 3 (GPT-3)'s diagnostic and triage ability for 48 validated synthetic case vignettes (<50 words; sixth-grade reading level or below) of both common (eg, viral illness) and severe (eg, heart attack) conditions to a nationally representative sample of 5000 lay people from the USA who could use the internet to find the correct options and 21 practising physicians at Harvard Medical School. There were 12 vignettes for each of four triage categories: emergent, within one day, within 1 week, and self-care. The correct diagnosis and triage category (ie, ground truth) for each vignette was determined by two general internists at Harvard Medical School. For each vignette, human respondents and GPT-3 were prompted to list diagnoses in order of likelihood, and the vignette was marked as correct if the ground-truth diagnosis was in the top three of the listed diagnoses. For triage accuracy, we examined whether the human respondents' and GPT-3's selected triage was exactly correct according to the four triage categories, or matched a dichotomised triage variable (emergent or within 1 day vs within 1 week or self-care). We estimated GPT-3's diagnostic and triage confidence on a given vignette using a modified bootstrap resampling procedure, and examined how well calibrated GPT-3's confidence was by computing calibration curves and Brier scores. We also performed subgroup analysis by case acuity, and an error analysis for triage advice to characterise how its advice might affect patients using this tool to decide if they should seek medical care immediately. Among all cases, GPT-3 replied with the correct diagnosis in its top three for 88% (42/48, 95% CI 75-94) of cases, compared with 54% (2700/5000, 53-55) for lay individuals (p<0.0001) and 96% (637/666, 94-97) for physicians (p=0·012). GPT-3 triaged 70% correct (34/48, 57-82) versus 74% (3706/5000, 73-75; p=0.60) for lay individuals and 91% (608/666, 89-93%; p<0.0001) for physicians. As measured by the Brier score, GPT-3 confidence in its top prediction was reasonably well calibrated for diagnosis (Brier score=0·18) and triage (Brier score=0·22). We observed an inverse relationship between case acuity and GPT-3 accuracy (p<0·0001) with a fitted trend line of -8·33% decrease in accuracy for every level of increase in case acuity. For triage error analysis, GPT-3 deprioritised truly emergent cases in seven instances. A general-purpose AI language model without any content-specific training could perform diagnosis at levels close to, but below, physicians and better than lay individuals. We found that GPT-3's performance was inferior to physicians for triage, sometimes by a large margin, and its performance was closer to that of lay individuals. Although the diagnostic performance of GPT-3 was comparable to physicians, it was significantly better than a typical person using a search engine. The National Heart, Lung, and Blood Institute.

Evaluating Health Literacy Resources for Secondary Hyperparathyroidism in End-Stage Kidney Disease

IntroductionParathyroidectomy is recommended for severe secondary hyperparathyroidism (SHPT) due to end-stage kidney disease (ESKD), but surgery is underutilized. High quality and accessible online health information, recommended to be at a 6th-grade reading level, is vital to improve patient health literacy. This study evaluated available online resources for SHPT from ESKD based on information quality and readability. MethodsThree search engines were queried using the terms “parathyroidectomy for secondary hyperparathyroidism,” “parathyroidectomy kidney/renal failure,” “parathyroidectomy dialysis patients,” “should I have surgery for hyperparathyroidism due to kidney failure?,” and “do I need surgery for hyperparathyroidism due to kidney failure if I do not have symptoms?” Websites were categorized by source and origin. Two independent reviewers determined information quality using JAMA (0-4) and DISCERN (1-5) frameworks, and scores were averaged. Cohen's kappa evaluated inter-rater reliability. Readability was determined using the Flesch Kincaid Reading Ease, Flesch Kincaid Grade Level, and Simple Measure of Gobbledygook tools. Median readability scores were calculated, and corresponding grade level determined. Websites with reading difficulties >6th grade level were calculated. ResultsThirty one (86.1%) websites originated from the U.S., with most from hospital-associated (63.9%) and foundation/advocacy sources (30.6%). The mean JAMA and DISCERN scores for all websites were 1.3 ± 1.4 and 2.6 ± 0.7, respectively. Readability scores ranged from grade level 5–college level, and most websites scored above the recommended 6th grade level. ConclusionsPatient-oriented websites tailoring SHPT from ESKD are at a reading level higher than recommended, and the quality of information is low. Efforts must be made to improve the accessibility and quality of information for all patients.

Word recognition thresholds in novice readers: exploring when reading and listening comprehension are comparable

AbstractThis study proposes and identifies two levels of word reading at which good comprehenders can obtain a similar understanding of written and spoken narrative texts: The oral-written matching functional and efficient thresholds. These thresholds are compared with the level of word reading over which a relationship between listening comprehension and reading comprehension is evident (a basic word recognition threshold) and with the level after which better word recognition does not result in improvements in reading comprehension (an upper word recognition threshold). The three first thresholds were calculated in a sample of 344 first to third-grade Spanish students who read and listened to two narrative texts of different difficulty and who were assessed on measures of comprehension and word recognition. Listening and reading comprehension were related when participants accurately read more than 73% of pseudowords in a list (basic word recognition threshold). For participants with good listening comprehension, reading comprehension performance matched listening comprehension once a moderate reading speed of ~ 66 decontextualized words per minute was achieved (oral-written matching functional threshold) or when the rate of reading the text was ~ 140 words per minute (oral-written matching efficient threshold). The value of the oral-written matching efficient threshold was, in this study, similar to the upper word recognition threshold identified by prior research. The thresholds calculated in this study delineate a sequence of increasingly challenges to the developing reader, allowing a nuanced description of the initial stages of reading development.

Handwriting Legibility is Related to Spelling Level, but Not to Reading and Spelling Progress

ABSTRACT In three studies the relationship between handwriting difficulties and dyslexia was investigated. Study 1 examined differences in handwriting difficulties between children with (n = 331) and without dyslexia (n = 305). Children with dyslexia experienced more difficulties with legible handwriting than those without. Study 2 explored the relationship between handwriting difficulties and reading/spelling proficiency among children with dyslexia (n = 311). Reading level and legibility were not related, whereas spelling level and legibility were. Children whose handwriting was extremely poor were more likely to have spelling problems. Study 3 investigated whether handwriting difficulties affected progress in reading and spelling during remediation (n = 249). While children with limited handwriting legibility generally exhibited poorer spelling skills, it did not impede their progress in reading and spelling during the intervention. These findings suggest that handwriting proficiency does not necessarily affect remediation efforts for reading and spelling.

Assessing the readability and quality of online patient information for laser tattoo removal

Just as tattoos continue to increase in popularity, many people with tattoos also seek removal, often due to career concerns. Prospective clients interested in laser tattoo removal may do research about the procedure online, as the internet increasingly becomes a resource to get preliminary health information. However, it is important that the online health information on the topic be of high quality and be accessible to all patients. We analyzed 77 websites from a Google search query using the terms “Laser tattoo removal patient Information” and “Laser tattoo removal patient Instructions” to assess this. The websites were evaluated for their readability using multiple validated indices and comprehensiveness. We found that websites had a broad readability range, from elementary to college, though most were above the recommended eighth-grade reading level. Less than half of the websites adequately discussed the increased risk of pigmentary complications in the skin of color clients or emphasized the importance of consulting with a board-certified dermatologist/plastic surgeon before the procedure. Over 90% of the websites noted that multiple laser treatments are likely needed for complete clearance of tattoos. The findings from our study underscore a significant gap in the accessibility and quality of online information for patients considering laser tattoo removal, particularly in addressing specific risks for patients with darker skin tones and emphasizing the need for consulting a board-certified physician before undergoing the procedure. It is important that online resources for laser tattoo removal be appropriately written to allow better decision-making, expectations, and future satisfaction for potential clients interested in the procedure.

Application of English Multi-Modal Reading Teaching Mode in the Context of “Internet Plus”

Under the background of “Internet plus”, the application of English multi-modal reading teaching mode is of great significance. With the continuous development of Internet technology, the field of education has also been deeply influenced, especially in English reading teaching. Therefore, this study aims to explore how to integrate multi-modal resources into English reading teaching under the background of “Internet plus”. Through literature review and case analysis, it is found that multi-modal reading teaching mode can improve students' reading interest and effectiveness. The results show that the use of images, audio, video and other forms of media resources in the teaching process can help students better understand and absorb English reading materials. Therefore, this paper proposes that in the era of “Internet plus”, teachers should actively design multi-modal English reading teaching activities with the help of various digital media resources, so as to improve students' reading level and ability and promote the development and innovation of English teaching.

TrachGPT: Appraisal of tracheostomy care recommendations from an artificial intelligent Chatbot.

Safe home tracheostomy care requires engagement and troubleshooting by patients, who may turn to online, AI-generated information sources. This study assessed the quality of ChatGPT responses to such queries. In this cross-sectional study, ChatGPT was prompted with 10 hypothetical tracheostomy care questions in three domains (complication management, self-care advice, and lifestyle adjustment). Responses were graded by four otolaryngologists for appropriateness, accuracy, and overall score. The readability of responses was evaluated using the Flesch Reading Ease (FRE) and Flesch-Kincaid Reading Grade Level (FKRGL). Descriptive statistics and ANOVA testing were performed with statistical significance set to p < .05. On a scale of 1-5, with 5 representing the greatest appropriateness or overall score and a 4-point scale with 4 representing the highest accuracy, the responses exhibited moderately high appropriateness (mean = 4.10, SD = 0.90), high accuracy (mean = 3.55, SD = 0.50), and moderately high overall scores (mean = 4.02, SD = 0.86). Scoring between response categories (self-care recommendations, complication recommendations, lifestyle adjustments, and special device considerations) revealed no significant scoring differences. Suboptimal responses lacked nuance and contained incorrect information and recommendations. Readability indicated college and advanced levels for FRE (Mean = 39.5, SD = 7.17) and FKRGL (Mean = 13.1, SD = 1.47), higher than the sixth-grade level recommended for patient-targeted resources by the NIH. While ChatGPT-generated tracheostomy care responses may exhibit acceptable appropriateness, incomplete or misleading information may have dire clinical consequences. Further, inappropriately high reading levels may limit patient comprehension and accessibility. At this point in its technological infancy, AI-generated information should not be solely relied upon as a direct patient care resource.