Comparative Performance of ChatGPT 3.5 and GPT4 on Rhinology Standardized Board Examination Questions.

Objective: Advances in deep learning and artificial intelligence (AI) have led to the emergence of large language models (LLM) like ChatGPT from OpenAI. The study aimed to evaluate the performance of ChatGPT 3.5 and GPT4 on Otolaryngology (Rhinology) Standardized Board Examination questions in comparison to Otolaryngology residents. Methods: This study selected all 127 rhinology standardized questions from www.boardvitals.com, a commonly used study tool by otolaryngology residents preparing for board exams. Ninety-three text-based questions were administered to ChatGPT 3.5 and GPT4, and their answers were compared with the average results of the question bank (used primarily by otolaryngology residents). Thirty-four image-based questions were provided to GPT4 and underwent the same analysis. Based on the findings of an earlier study, a pass-fail cutoff was set at the 10th percentile. Results: On text-based questions, ChatGPT 3.5 answered correctly 45.2% of the time (8th percentile) (P = .0001), while GPT4 achieved 86.0% (66th percentile) (P = .001). GPT4 answered image-based questions correctly 64.7% of the time. Projections suggest that ChatGPT 3.5 might not pass the American Board of Otolaryngology Written Question Exam (ABOto WQE), whereas GPT4 stands a strong chance of passing. Discussion: The older LLM, ChatGPT 3.5, is unlikely to pass the ABOto WQE. However, the advanced GPT4 model exhibits a much higher likelihood of success. This rapid progression in AI indicates its potential future role in otolaryngology education. Implications for Practice: As AI technology rapidly advances, it may be that AI-assisted medical education, diagnosis, and treatment planning become commonplace in the medical and surgical landscape. Level of Evidence: Level 5.

The Use of Artificial Intelligence to Improve Readability of Otolaryngology Patient Education Materials.

OBJECTIVE: The recommended readability of health education materials is at the sixth-grade level. Artificial intelligence (AI) large language models such as the newly released ChatGPT4 might facilitate the conversion of patient-education materials at scale. We sought to ascertain whether online otolaryngology education materials meet recommended reading levels and whether ChatGPT4 could rewrite these materials to the sixth-grade level. We also wished to ensure that converted materials were accurate and retained sufficient content. METHODS: Seventy-one articles from patient educational materials published online by the American Academy of Otolaryngology-Head and Neck Surgery were selected. Articles were entered into ChatGPT4 with the prompt "translate this text to a sixth-grade reading level." Flesch Reading Ease Score (FRES) and Flesch-Kincaid Grade Level (FKGL) were determined for each article before and after AI conversion. Each article and conversion were reviewed for factual inaccuracies, and each conversion was reviewed for content retention. RESULTS: The 71 articles had an initial average FKGL of 11.03 and FRES of 46.79. After conversion by ChatGPT4, the average FKGL across all articles was 5.80 and FRES was 77.27. Converted materials provided enough detail for patient education with no factual errors. DISCUSSION: We found that ChatGPT4 improved the reading accessibility of otolaryngology online patient education materials to recommended levels quickly and effectively. IMPLICATIONS FOR PRACTICE: Physicians can determine whether their patient education materials exceed current recommended reading levels by using widely available measurement tools, and then apply AI dialogue platforms to modify materials to more accessible levels as needed. LEVEL OF EVIDENCE: Level 5.

Otolaryngology Care Disparities in American Indian Populations.

Our objectives were to quantify geographical disparities in otolaryngology care access with respect to American Indian (AI) populations and to identify gaps in care. Although increased incidence and mortality rates of ear, nose, and throat (ENT) conditions in AI populations are well documented, few studies address factors contributing to these differential outcomes. We conducted a cross-sectional study of US states with AI areas that either met the population threshold for the American Community Survey annual estimate or annual supplemental estimate. A 2-tailed t test was used to compare the geographic distribution of ENT providers practicing within AI areas against non-AI areas, showing a statistically significant difference (P < .001) in the concentration of providers (0.409 vs 2.233 providers per 100,000 patients). To our knowledge, this is the first study to explore geographic barriers contributing to AI disparities within otolaryngology.

An Integrative Model of ENT Healthcare for the Homeless Population.

OBJECTIVE: Our aim was to determine the need for otolaryngology care within the homeless population, identify barriers to access that the homeless population may face, and develop a model system which would address these needs with respect to the barriers. METHODS: A retrospective chart review of 812 patients seen between 1/25/16-3/21/2020 was performed. Charts were obtained from homeless patients seen at free general clinics held shelters in Chicago, IL (781 charts) and in Champaign, IL (31 charts). Records reporting at least one otolaryngology disease in a patient experiencing homelessness were included in this study. Patients were considered homeless if they resided at the shelter at the time of their appointment. To determine common barriers to care, a simple yes/no questionnaire was administered to residents at west-side Chicago homeless shelters. Questions addressed barriers to health care access that had been mentioned by patients seen at free clinics. RESULTS: Chart review findings demonstrated that 14.3% (n = 142) of all homeless patients seen at free in-shelter clinics were seen for ENT-related disorders. Survey results revealed that 76.3% (n = 71) of respondents believed that telemedicine services would be useful in shelters. 74.2% (n = 69) stated they were unable to afford prescribed medications. 93.5% (n = 87) stated that better access to transportation would increase their likelihood of seeking care. CONCLUSIONS: In our attempt to bridge this inequity, we have launched a hybrid in-person/virtual care program to improve access to otolaryngology care for the homeless community. LEVEL OF EVIDENCE: N/A Laryngoscope, 134:2705-2709, 2024.

Racial Disparities in Head and Neck Cancer: It's Not Just About Access.

OBJECTIVE: Medical literature identifies stark racial disparities in head and neck cancer (HNC) in the United States, primarily between non-Hispanic white (NHW) and non-Hispanic black (NHB) populations. The etiology of this disparity is often attributed to inequitable access to health care and socioeconomic status (SES). However, other contributors have been reported. We performed a systematic review to better understand the multifactorial landscape driving racial disparities in HNC. DATA SOURCES: A systematic review was conducted in Covidence following Preferred Reporting Items for Systematic Reviews and Meta-analyses Guidelines. A search of PubMed, SCOPUS, and CINAHL for literature published through November 2022 evaluating racial disparities in HNC identified 2309 publications. REVIEW METHODS: Full texts were screened by 2 authors independently, and inconsistencies were resolved by consensus. Three hundred forty publications were ultimately selected and categorized into themes including disparities in access/SES, treatment, lifestyle, and biology. Racial groups examined included NHB and NHW patients but also included Hispanic, Native American, and Asian/Pacific Islander patients to a lesser extent. RESULTS: Of the 340 articles, 192 focused on themes of access/SES, including access to high-quality hospitals, insurance coverage, and transportation contributing to disparate HNC outcomes. Additional themes discussed in 148 articles included incongruities in surgical recommendations, tobacco/alcohol use, human papillomavirus-associated malignancies, and race-informed silencing of tumor suppressor genes. CONCLUSION: Differential access to care plays a significant role in racial disparities in HNC, disproportionately affecting NHB populations. However, there are other significant themes driving racial disparities. Future studies should focus on providing equitable access to care while also addressing these additional sources of disparities in HNC.

Multiple skull base defects in the setting of spontaneous cerebrospinal fluid rhinorrhea; a dual institution view.

PURPOSE: Idiopathic intracranial hypertension (IIH) is a condition of high cerebrospinal fluid (CSF) pressure that presents with CSF leak. The implications of multiple skull base defects (SBD) and associated synchronous CSF leaks have not been previously explored. MATERIALS AND METHODS: A dual institutional case-control study examined multiple SBD's and encephaloceles on the risk of CSF leak and postoperative failures post-repair. IIH patients with CSF leaks and IIH controls without leaks were selected retrospectively. Chi square analysis evaluated for statistically significant alterations in probability with CSF leak development. RESULTS: 192 patients were selected with 108 IIH controls and 84 spontaneous CSF leak cases. Signs and symptoms for IIH controls and CSF leak cases respectively were pulsatile tinnitus (60.2 % and 29.8 %), headaches (96.3 % and 63.1 %), papilledema (74.1 % and 12.5 %), visual field defects (60.8 % and 13 %) (p < 0.001). Encephalocele formation in controls was 3.7 % compared to cases at 91.6 % (p < 0.001). Multiple SBD's in controls compared cases was 0.9 % and 46.4 % respectively (p < 0.001). Subgroup analysis of CSF leak cases showed 15 patients with two CSF leak repairs due to a recurrence. 27 (39.1 %) single leak cases had multiple SDB's while 12 (80 %) recurrent leaks had multiple SDB's (p = 0.004). CONCLUSIONS: Patients with radiographic evidence of multiple SBD's and encephaloceles represent a high-risk population with a propensity for CSF leaks. Secondary SBD's are common in patients with spontaneous CSF rhinorrhea and higher in patients with a recurrence.

Evaluation of Wait Times for Otolaryngology Appointments in Illinois.

Objective: The objective of this study was to quantify the wait times that patients may encounter for common clinical diagnoses when seeking otolaryngology care, while determining whether a wait time disparity exists based on geographic location within Illinois. Methods: In November 2022, a list of Illinois otolaryngologists was obtained from www.entnet.org. Using a uniform script, each of the 291 otolaryngologists were contacted. The caller posed as a new patient with either sudden sensorineural hearing loss, a neck mass, or chronic sinusitis. Each clinic was called 3 times and wait times were recorded. One hundred fifty-eight otolaryngologists were included in the analysis. Results: The average statewide wait time for a new patient presenting with sudden unilateral hearing loss, a neck mass, and chronic sinusitis was 18.0, 22.6, and 25.5 days, respectively. There was no statistically significant difference between urban and rural wait times. Discussion: Although wait time differences were noted, the lack of urban versus rural p value significance may be attributed to the small sample size (n = 11) of rural otolaryngologists in Illinois. However, the overall wait times in this study were longer compared to those reported in other studies, suggesting that the current number of otolaryngologists in Illinois is inadequate to meet the public need. Implications for Practice: We have demonstrated that the current demand for otolaryngology care is outstripping the existing supply in Illinois. This suggests that an emphasis should be placed on training more otolaryngologists, or increasing the use of physician extenders, while incentivizing otolaryngologists to practice in rural areas. Level of Evidence: 5.