OBJECTIVE: Language used by providers in medical documentation may reveal evidence of race-related implicit bias. We aimed to use natural language processing (NLP) to examine if prevalence of stigmatizing language in emergency medicine (EM) encounter notes differs across patient race/ethnicity.
METHODS: In a retrospective cohort of EM encounters, NLP techniques identified stigmatizing and positive themes. Logistic regression models analyzed the association of race/ethnicity and themes within notes. Outcomes were the presence (or absence) of 7 different themes: 5 stigmatizing (difficult, non-compliant, skepticism, substance abuse/seeking, and financial difficulty) and 2 positive (compliment and compliant).
RESULTS: The sample included notes from 26,363 unique patients. NH Black patient notes were less likely to contain difficult (odds ratio (OR) 0.80, 95% confidence interval (CI), 0.73-0.88), skepticism (OR 0.87, 95% CI, 0.79-0.96), and substance abuse/seeking (OR 0.62, 95% CI, 0.56-0.70) compared to NH White patient notes but more likely to contain non-compliant (OR 1.26, 95% CI, 1.17-1.36) and financial difficulty (OR 1.14, 95% CI, 1.04-1.25). Hispanic patient notes were less likely to contain difficult (OR 0.68, 95% CI, 0.58-0.80) and substance abuse/seeking (OR 0.78, 95% CI, 0.66-0.93). NH NA/AI patient notes had twice the odds as NH White patient notes to contain a stigmatizing theme (OR 2.02, 95% CI, 1.64-2.49).
CONCLUSIONS: Using an NLP model to analyze themes in EM notes across racial groups, we identified several inequities in the usage of positive and stigmatizing language. Interventions to minimize race-related implicit bias should be undertaken.

BACKGROUND: Medical documentation plays a crucial role in clinical practice, facilitating accurate patient management and communication among health care professionals. However, inaccuracies in medical notes can lead to miscommunication and diagnostic errors. Additionally, the demands of documentation contribute to physician burnout. Although intermediaries like medical scribes and speech recognition software have been used to ease this burden, they have limitations in terms of accuracy and addressing provider-specific metrics. The integration of ambient artificial intelligence (AI)-powered solutions offers a promising way to improve documentation while fitting seamlessly into existing workflows.
OBJECTIVE: This study aims to assess the accuracy and quality of Subjective, Objective, Assessment, and Plan (SOAP) notes generated by ChatGPT-4, an AI model, using established transcripts of History and Physical Examination as the gold standard. We seek to identify potential errors and evaluate the model\'s performance across different categories.
METHODS: We conducted simulated patient-provider encounters representing various ambulatory specialties and transcribed the audio files. Key reportable elements were identified, and ChatGPT-4 was used to generate SOAP notes based on these transcripts. Three versions of each note were created and compared to the gold standard via chart review; errors generated from the comparison were categorized as omissions, incorrect information, or additions. We compared the accuracy of data elements across versions, transcript length, and data categories. Additionally, we assessed note quality using the Physician Documentation Quality Instrument (PDQI) scoring system.
RESULTS: Although ChatGPT-4 consistently generated SOAP-style notes, there were, on average, 23.6 errors per clinical case, with errors of omission (86%) being the most common, followed by addition errors (10.5%) and inclusion of incorrect facts (3.2%). There was significant variance between replicates of the same case, with only 52.9% of data elements reported correctly across all 3 replicates. The accuracy of data elements varied across cases, with the highest accuracy observed in the \"Objective\" section. Consequently, the measure of note quality, assessed by PDQI, demonstrated intra- and intercase variance. Finally, the accuracy of ChatGPT-4 was inversely correlated to both the transcript length (P=.05) and the number of scorable data elements (P=.05).
CONCLUSIONS: Our study reveals substantial variability in errors, accuracy, and note quality generated by ChatGPT-4. Errors were not limited to specific sections, and the inconsistency in error types across replicates complicated predictability. Transcript length and data complexity were inversely correlated with note accuracy, raising concerns about the model\'s effectiveness in handling complex medical cases. The quality and reliability of clinical notes produced by ChatGPT-4 do not meet the standards required for clinical use. Although AI holds promise in health care, caution should be exercised before widespread adoption. Further research is needed to address accuracy, variability, and potential errors. ChatGPT-4, while valuable in various applications, should not be considered a safe alternative to human-generated clinical documentation at this time.

To evaluate the quality of Electronic Health Record (EHR) documentation practices of Female Genital Cutting (FGC) by medical providers. A retrospective chart review study of 99 patient encounter notes within the University of Minnesota health system (inclusive of 40 hospitals and clinics) was conducted. Extracted data included but was not limited to patient demographics, reason for patient visit, ICD code used in note, and provider description of FGC anatomy. Data was entered into REDCAP and categorized according to descriptive statistics. Out of 99 encounters, 45% used the unspecified code for FGC. The most common reason for patient visits was sexual pain, though many notes contained several reasons for the visit regarding reproductive, urological, or sexual concerns. 56% of visits discussed deinfibulation. 11 different terms for FGC were used, with \"female circumcision\" being the most common. 14 different terms for deinfibulation were found within 64 notes. 42% of encounters included a description of introitus size in the anatomical description, and only 38% of these provided a metric measurement. This study found significant variation in the quality of FGC documentation practices. Medical providers often used the unspecified FGC code, subjective and/or seemingly inaccurate descriptions of FGC/anatomy, and several different terms for both FGC and deinfibulation. Clearly, more education is needed in clinical training programs to (1) identify FGC type, (2) use the corresponding ICD code, and (3) use specific, objective descriptions (including presence/absence of structures and infibulation status).

OBJECTIVE: Chat generative pre-trained transformer (GPT) is a novel large pre-trained natural language processing software that can enable scientific writing amongst a litany of other features. Given this, there is a growing interest in exploring the use of ChatGPT models as a modality to facilitate/assist in the provision of clinical care.
METHODS: We investigated the time taken for the composition of neurosurgical discharge summaries and operative reports at a major University hospital. In so doing, we compared currently employed speech recognition software (i.e., SpeaKING) vs novel ChatGPT for three distinct neurosurgical diseases: chronic subdural hematoma, spinal decompression, and craniotomy. Furthermore, factual correctness was analyzed for the abovementioned diseases.
RESULTS: The composition of neurosurgical discharge summaries and operative reports with the assistance of ChatGPT leads to a statistically significant time reduction across all three diseases/report types: p < 0.001 for chronic subdural hematoma, p < 0.001 for decompression of spinal stenosis, and p < 0.001 for craniotomy and tumor resection. However, despite a high degree of factual correctness, the preparation of a surgical report for craniotomy proved to be significantly lower (p = 0.002).
CONCLUSIONS: ChatGPT assisted in the writing of discharge summaries and operative reports as evidenced by an impressive reduction in time spent as compared to standard speech recognition software. While promising, the optimal use cases and ethics of AI-generated medical writing remain to be fully elucidated and must be further explored in future studies.

Introduction To comply with the Information Blocking Rule in the 21st Century Cures Act, many hospitals began to release inpatient electronic health information such as clinical notes and results to patients immediately, starting in April 2021. We sought to understand the perceptions of hospital-based clinicians regarding the impact of these changes in information sharing on clinicians and patients. Materials and methods We developed and distributed an electronic survey to 122 inpatient attending physicians, resident physicians, and physician assistants within the internal medicine and family medicine departments at an academic medical center. The survey asked clinicians to rate their comfort with information-sharing protocols and describe their perceptions of the impact of immediate information sharing on their documentation habits and patient interactions following the implementation of the Cures Act. Results The survey response rate was 37.7% (46/122). Of the respondents, 56.5% felt comfortable with the note-sharing process, 84.8% reported omitting specific information from their notes to prevent patients from reading it, and 39.1% of clinicians agreed that patients have found clinical notes \"more confusing than helpful.\" Conclusions Immediate sharing of electronic health information has the potential to be a powerful tool for communicating with hospitalized patients. However, our results show many hospital-based clinicians report limited comfort with the note-sharing process and perceive it to be confusing to patients. Efforts are needed to educate clinicians regarding information sharing, understand patient and family perspectives, and develop best practices to enhance communication through electronic notes.

UNASSIGNED: Outpatient diabetes mellitus (DM) care over video telehealth (TH) requires modifications to how endocrinologists complete physical examinations (PEs). But there is little guidance on what PE components to include, which may incur wide variation in practice. We compared endocrinologists\' documentation of DM PE components for in-person (IP) vs TH visits.
UNASSIGNED: Retrospective chart review of 200 notes for new patients with DM from 10 endocrinologists (10 IP and 10 TH visits each) in the Veterans Health Administration between April 1, 2020, and April 1, 2022. Notes were scored from 0 to 10 based on documentation of 10 standard PE components. We compared mean PE scores for IP vs TH across all clinicians using mixed effects models. Independent samples t-tests were used to compare both mean PE scores within clinician and mean scores for each PE component across clinicians for IP vs TH. We described virtual care-specific and foot assessment techniques.
UNASSIGNED: The overall mean (SE) PE score was higher for IP vs TH (8.3 [0.5] vs 2.2 [0.5]; P < .001). Every endocrinologist had higher PE scores for IP vs TH. Every PE component was more commonly documented for IP vs TH. Virtual care-specific techniques and foot assessment were rare.
UNASSIGNED: Our study quantifies the degree to which Pes for TH were attenuated among a sample of endocrinologists, raising a flag that process improvements and research are needed for virtual Pes. Organizational support and training could help increase PE completion via TH. Research should examine reliability and accuracy of virtual PE, its value to clinical decision-making, and its impact on clinical outcomes.

Background and aim Ward-round documentation is important for clinical communication and patient safety. Standardized checklists have improved ward-round documentation in surgical and medical settings. This quality improvement project aimed to introduce a standardized ward round proforma to improve documentation in a UK specialist stroke unit. Methods Ward round entries were assessed against internally agreed standardized criteria. A stroke-specific ward round proforma was designed and introduced with input from the multidisciplinary team. A repeat audit was performed, including assessment of the use of different proforma sections. Multidisciplinary team members were invited to provide feedback via an anonymous online survey. Results A total of 111 ward round entries were reviewed before the proforma was introduced. Ninety-five ward round entries were reviewed following introduction of the proforma, and 84.2% of these used the proforma for documentation. Overall documentation of standardized criteria improved from 48.7% to 62.1% with substantial improvement seen in documentation of neurological examination, presence/absence of mechanical venous thromboembolism prophylaxis, and blood test results. Multidisciplinary team feedback was positive. Conclusions The stroke-specific ward round proforma improved the quality and consistency of documentation in the unit. An updated proforma was designed using these results and multidisciplinary team feedback.

BACKGROUND: Medical documentation is an important part of the medical process as it is an essential way of communication within the health care system. However, medical documentation practice in the private sector is not well studied in Ethiopian context. The aim of this study was to assess the practice of medical documentation and its associated factors among health workers at private hospitals in the Amhara region, Ethiopia.
METHODS: An institution-based cross-sectional quantitative study supplemented with a qualitative design was conducted among 419 health workers at the private hospitals in the Amhara Region, Ethiopia from March 29 to April 29 /2021. Data were collected using both a self-administered questionnaire and interview guide for quantitative and qualitative respectively. Data were entered using Epi data version 3.1 and analyzed using SPSS version 20. Descriptive statistics, Bi-variable, and multivariable logistic regression analysis were performed. In-depth interviews were conducted using semi-structured questionnaires with eight respondents to explore the challenges related to the practice of medical documentation. Respondent\'s response were analyzed using OpenCode version 4.03 thematically.
RESULTS: Four hundred seven study participants returned the questionnaire. Nearly 50 % (47.2%) health workers had of good medical documentation practice. Health workers who received in-service training on medical documentation AOR = 2.77(95% CI: [1.49,5.14]), good knowledge AOR = 2.28 (95% CI: [1.34,3.89]), favorable attitude AOR = 1.78 (95%CI: [1.06,2.97]), strong motivation AOR = 3.49 (95% CI: [2.10,5.80]), available guide line formats AOR = 3.12 (95% CI: [1.41,6.84]), eHealth literacy AOR = 1.73(95% CI: [1.02,2.96]), younger age AOR = 2.64 (95% CI:[1.27,5.46]) were statistically associated with medical documentation.
CONCLUSIONS: More than half of the medical services provided were not registered. Therefore, it is important to put extra efforts to improve documentation practice by providing planed trainings on standards of documentation to all health workers, creating positive attitudes and enhancing their knowledge by motivating them to develop a culture of information.

To create and validate a checklist for high-quality documentation and pilot a multi-modal, immersive educational module across multiple institutions. We hypothesized that this module would improve knowledge, skills, and attitudes in medical documentation.
Module design was grounded in an established curriculum design framework. We conducted the study across 12 pediatric critical care fellowship programs between September 2017 and January 2018. Workshops were allotted 90 minutes for completion. We utilized a pre-/post- study design to determine the workshop\'s impact. Changes in knowledge were assessed through pre and post testing. Changes in skills were evaluated with a validated checklist for inclusion of key documentation elements. Changes in attitudes were determined through learner self-assessment RESULTS: 83 of 138 eligible fellows (60%) started the module and 62 of 83 (75%) completed data sets for analysis. Immediate post-testing demonstrated modest statistically significant improvement in knowledge, skills, and attitudes. The workshop was easily disseminated and deployed CONCLUSIONS: This study demonstrates that a multi-modal educational intervention can lead to improvement in medical documentation knowledge, skills, and attitudes in a cohort of PCCM fellows and be easily disseminated for use by other specialties and types of clinicians.

We assessed the accuracy of the ICD-10 code for delirium (F05) and its relationship with delirium discharge summary documentation.
We performed a retrospective chart review at three academic hospitals. The Chart-based Delirium Identification Instrument (CHART-DEL) was used to identify 108 hospitalized patients aged ≥65 years with delirium, and 758 patients without delirium as controls. We assessed the proportion of patients who received the F05 code and calculated the sensitivity and specificity. We compared the rates of F05 code received between patients with and without \"delirium\" documented in the discharge summary.
Among delirious patients, 46.3% received a F05 code, which has a sensitivity of 46.3% and specificity of 99.6% for delirium. Of charts with \"delirium\" in the discharge summary (n = 67), 67.2% were appropriately coded.
Current ICD-10 data inadequately capture delirium. Delirium documentation in the discharge summary is associated with improved delirium coding.