Clinical Documentation Research Articles

Safe implementation of minimally invasive surgery in a specialized colorectal cancer unit.

In the past 30 years, minimally invasive surgery (MIS) has made remarkable progress and has become the standard of care in colorectal cancer treatment. The implementation of new techniques or platforms is, therefore, a challenge for surgical teams. This study aims to analyze the experience in the implementation of minimally invasive surgery in the colorectal unit in a specialized colorectal cancer center. We will report and compare the clinical outcomes of the patients submitted to the different surgical approaches, reflecting the importance of surgical training in the laparoscopic and robotic field for the reduction of surgical complications and improve short-term outcomes. This study involved a retrospective analysis of data collected from a prospectively maintained database at the colorectal unit of Champalimaud Foundation between 2012 and 2023. Data were collected as part of routine clinical documentation and included variables on patient's demographics, staging, short-term outcomes, and follow-up. A total of 661 patients treated at the Champalimaud Foundation between 2012 and 2023 were included, of which 389 (59%) had colon and 272 (41%) rectal cancer. Most of the patients underwent elective surgery, with a minimally invasive approach performed in 91% of cases. A complete resection (R0) was achieved in 95.1% (619) of the procedures with a pathology report staging 64.5% (409) of tumors as pT3-4. Eleven percent (70) of patients had complications classified as Clavien-Dindo (CD) ≥ 3. This study supports the safety of the implementation of minimally invasive surgery in colorectal cancer care, with improvement in postoperative outcomes and surgical quality, supporting the importance of surgical training and specialized teams.

A Qualitative Study of School Mental Health Providers’ Experiences with Chart Notes: Perceived Utility, Burden, and Areas for Growth

ABSTRACT Background Despite the ubiquity of chart notes and their well-known impact on medical providers’ job stress and satisfaction, little is known about how mental health providers interact with and are impacted by documentation requirements, particularly in school environments. Objective This qualitative study sought to investigate school mental health providers’ perspectives on and experiences with documentation. Method Four focus groups with eight school mental health providers were conducted. Providers were asked about their current documentation practices and desired changes to their electronic health record system. Recordings of focus groups were transcribed and analyzed for themes using a rapid qualitative analytic approach. Results Key themes revolved around the purpose, features, time demands, training needs, and writing processes surrounding chart notes. Providers reported spending substantial time and energy on documentation and viewed chart notes primarily as a billing requirement, but also valued their utility for treatment planning, outcome monitoring, and clinical accountability. Providers expressed frustration with redundant data entry, lack of protected time for documentation, insufficient training for and communication around updates to documentation requirements, and strained interactions with supervisors during editing. Providers highlighted scheduling protected work time for documentation, enhancing documentation training with a focus on habit formation, standardizing supervisory expectations and providing documentation specific training for supervisors, and redesigning electronic health record interfaces with automation and checklists as strategies that could improve their experiences with chart notes. Conclusions Providers highlighted the ways current training and supervision practices, time constraints, and electronic health record functionality contribute to job stress and burnout while also noting the value of documentation for treatment planning and processing difficult emotions. Roadmaps outlining further research that is needed to understand these suggestions from providers in the larger context of clinical documentation, as well their potential benefits and tradeoffs for providers and other stakeholders, are discussed.

Ocular outcomes of toxoplasmosis in the Amazon-Cerrado transition zone, Brazil: a population-based study.

Eye damage represents one of the gravest outcomes of toxoplasmosis, potentially leading to varying degrees of vision impairment, up to and including blindness. Monitoring ophthalmological indicators in individuals affected by toxoplasmosis, alongside understanding its impact on visual acuity, can motivate the scientific community to advocate for strategies to prevent vision loss associated with toxoplasmosis. This is a retrospective study analysing medical records and clinical documentation from patients at an ophthalmological clinic in a city located at the intersection of the Cerrado and Amazon Forest biomes. It was noted that >20% of patients exhibited visual acuity ≤20/200, a condition medically recognized as blindness for the affected eye. Our study highlights a significant portion of individuals experiencing lasting visual impairments due to toxoplasmosis. These findings remain obscured by the existing national guidelines, which only mandate reporting for congenital and gestational toxoplasmosis cases, thereby overlooking sporadic manifestations and their consequential effects. The effort to identify ophthalmologic consequences is crucial both for personalized clinical care and for shaping national guidelines for managing toxoplasmosis.

Abstract B045: The impact of smoking status on the genomic landscape of lung squamous cell carcinoma

Abstract Introduction: Molecular diagnostics have become a standard of care in determining the appropriate treatment paradigm for patients with advanced non-small cell lung cancer (NSCLC). Current guidelines indicate that histology is the most important criteria in deciding on conducting molecular testing. In lung adenocarcinoma, guidelines recommend that all patients with advanced disease undergo comprehensive molecular testing. However, the current National Comprehensive Cancer Guidelines (NCCN) suggest only consideration of molecular testing in lung squamous cell carcinoma (SCC). Despite this recommendation, numerous genomic alterations (GAs) have been observed in lung SCC including amplifications, fusions, and mutations. The list of druggable genetic alterations has been growing, now including ROS1 rearrangements, RET rearrangements, and MET alterations with associated pharmacologic therapies. This suggests a role for broader comprehensive molecular profiling in NSCLC. We sought to assess the impact of smoking status on the proportion of actionable GAs in a cohort of lung SCC patients. Methods: We retrospectively reviewed liquid biopsy reports from Guardant 360 between October 2020 and July 2023 that were obtained in the context of routine clinical care at a single institution, the City of Hope Comprehensive Cancer Center. We then obtained social history from clinical documentation. Patients were categorized based on smoking status with the following categories: non-smoker, remote smoking history (quit >20 years prior), or smoking history (ongoing or quit within the last 20 year). Results: 56 patients were included in the final analysis. 98% (n = 55) patients had a documented smoking history, with 24% (n=13) classified as non-smokers. In total, 96% (n = 54) patients had detectable alterations of variable significance Of the 13 patients classified as non-smokers, 62% (n=8) patients had an actionable GA detected by liquid biopsy which has a current FDA approved agent. Of targetable GAs in the non-smoking cohort, the following alterations were observed: EGFR exon 19 deletion (44% , n=4), MET exon 14 skipping mutation (22%, n=2), EGFR G719S (11%, n=1), EGFR E114K (11% n=1). Of the 8 patients with targetable alterations detected, responses by RECIST included 4 partial response, 2 stable disease, and 2 progressive disease. The objective response rate (ORR) for this group was 55.6% and the clinical benefit rate (CBR) was 77.8%. Discussion: These data help illustrate the importance of genomic testing in lung SCC, particularly in patients with a non- smoking status. Given that 62% of non-smokers harbored a mutation treated with an FDA- approved agent, up-front genomic testing is of paramount importance. For smokers, numerous alterations were also detected which have been shown to result in decreased efficacy of immunotherapy regimens. This analysis helps highlight the potential value of liquid biopsy testing in lung SCC patients with both non-smoking and smoking histories. Citation Format: Adam Rock, Sahil Garg, Ravi Salgia, Victoria Villaflor, Ramya Muddasani, Erminia Massarelli, Colby Jenkins. The impact of smoking status on the genomic landscape of lung squamous cell carcinoma [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr B045.

Exploring Clinical Documentation Practices among Healthcare Staff: A Clinical Audit of National Hospital Galle, Sri Lanka

Introduction: Valid and trustworthy clinical documentation improves patient outcomes, boosts hospital data accuracy, and enhances reflective action to improve healthcare service quality and safety. In Sri Lanka, the bedhead ticket (BHT) is the legal documentation record for inpatients. In 2022, the Ministry of Health (MoH) issued new rules for the correct maintenance of BHTs.Objective: This clinical audit's objective was to assess the in-patient clinical documenting practices at National Hospital Galle (NHG), Sri Lanka.Methodology: A total of 384 randomly selected BHTs from the record room, which belonged to patients who received inward care from April 2023 to September 2023, were included in this clinical audit. A checklist and a marking scheme were prepared according to MoH guidelines and filled out after going through each BHT.Results: The level of completeness of the first page was good in 96% of BHTs. Maintenance of continuation sheets was average in 82%of BHTs and good in only 2%. Around 67% of BHTs obtained good marks for the level of completeness of entries. Only 1% of BHTs had entries with a signature placed at the end. The date and time of the entry were recorded in 73.5% and 64.3% consecutively.Conclusion: Adherence to the MoH guidelines for clinical documentation practices was satisfactory at NHG, and regular audits and staff awareness and training are required to maintain the standards.

Abstract Su601: Missed Hypoxia in the Prehospital Care of Major Traumatic Brain Injury: Discrepancies Between Continuous Monitor Data and Clinical Documentation

Background: It is well established that prehospital hypoxia dramatically increases mortality in Traumatic Brain Injury (TBI). Thus, in EMS TBI research, case ascertainment and risk adjustment are highly dependent upon documentation of in-field O 2 saturation (SpO 2 ). Objective: To compare the rate of hypoxia identified by EMS personnel and documented in EMS patient care records (PCR) versus the actual rate of hypoxia recorded by continuous, non-invasive monitoring in TBI. Methods: A subset of major TBI cases (moderate/severe) from 5 EMS agencies reporting monitor data (Philips MRx™) in the EPIC EMS TBI Study (NIH 1R01NS071049) were evaluated (02/13-02/18). All monitor data available for post-hoc review were displayed and accessible to the providers during EMS care. We compared PCR documentation of hypoxia (SpO 2 <90%) to continuous monitor data in TBI patients (nasal canula O 2 , mask, basic or advanced airway). Statistic: exact McNemar’s test, α=0.05. Results: 120 cases were included [med. age 50; IQR (23.8, 68); 65% male. Monitors recorded 44 hypoxic cases (36.7%; 95%CI 28.1, 45.9%). Only 8 (6.7%; 2.9, 12.7%) were documented in the PCR (p<0.0001; Figure). Three machine-detected hypoxia cases had no PCR-documented SpO 2 at all. Conclusions: The difference in PCR-documented and monitor-identified hypoxia was dramatic. The monitor data revealed a five-fold greater incidence of hypoxia than that documented by EMS providers. This may be partially explained by the continuous nature of pulse oximetry in the face of ongoing care responsibilities and scene distractions. The field environment may cause providers to miss low readings as they fluctuate moment-by-moment. These findings have major clinical implications. While some of the discrepancy between PCR reporting and monitor data might simply be failure to document hypoxia, the much more concerning implication is that many cases might be unrecognized and untreated. Furthermore, these findings have important implications for case ascertainment, confounding, and risk-adjustment in EMS TBI research. Whenever possible, quality improvement and research projects should utilize continuous non-invasive monitor data to identify and evaluate hypoxic patients in the setting of TBI. These findings may also have implications for identifying hypoxia in EMS patients with other critical conditions.

BIOM-46. REAL-WORLD EXPERIENCE WITH CIRCULATING TUMOR DNA IN CEREBROSPINAL FLUID IN PATIENTS WITH CENTRAL NERVOUS SYSTEM TUMORS

Abstract Next generation sequencing (NGS) of circulating tumor DNA (ctDNA) from cerebrospinal fluid (CSF) offers a minimally invasive approach to identify somatic alterations for individuals with central nervous system (CNS) cancers. Here, we review our ‘real world’ experience using a clinically validated, FDA-authorized platform, MSK-IMPACTTM, to comprehensively profile ctDNA from CSF samples with and without clinically documented CNS involvement by cancer (1007 samples, 711 patients). The patient cohort included patients with over 90 distinct tumor types of primary CNS and metastatic origin, with lung cancer (n = 188), breast cancer (n = 150), and glioma (n = 148) representing the most common broad categories. We identified genetic alterations (ctDNA-positive) in 489/922 (53.0%) CSF samples from patients with clinically documented CNS cancer. Over half of these ctDNA-positive samples (248, 50.7%) possessed a biomarker that was predictive of response to an FDA-approved drug (OncoKB level 1). Furthermore, the distribution of clinically actionable somatic alterations was consistent with tumor-type specific alterations across the AACR GENIE cohort. Of patients without clinical documentation of CNS cancer, none of these 85 CSF samples were ctDNA-positive (specificity = 100%). Successive CSF sequencing over time identified clonal evolution and the emergence of resistance mechanisms to therapies. ctDNA positivity corresponded with shortened overall survival after CSF collection (HR: 3.23, 95% CI: 2.58-4.05, P < 0.001). NGS of CSF cfDNA can detect clinically actionable somatic alterations in a large subset of CNS cancer patients and allows tracking of tumor evolution. Detection of somatic alterations in CSF may be a useful prognostic biomarker for survival in cancer patients.

Symptomatic Acetabular Erosion After Hip Hemiarthroplasty: Is It a Major Concern? A Retrospective Analysis of 2477 Hemiarthroplasty Cases

Background/Objectives: Acetabular erosion is a postoperative condition that can occur after hip hemiarthroplasty (HA), potentially leading to pain and requiring conversion to total hip arthroplasty (THA). Given the discrepancy in its incidence and impact in the literature, this study aims to report the incidence of symptomatic acetabular erosion and the subsequent conversion to THA in all HA cases performed in a single health system. Methods: A total of 2477 HA cases had their clinical notes and serial radiographs examined for this retrospective study. Outcome measures included any records of hip or groin pain and conversion to THA that were attributed to acetabular erosion as documented in the clinical notes. Results: Two thousand four hundred and seventy-seven HA cases were reviewed in this study. The mean age for all patients in the study was 81.5 years and the mean follow up was 3.7 years. Out of the 2477 HA cases, only 12 HA cases (0.48%) in 12 patients had data records of chronic hip pain, attributable to acetabular wear in the clinical notes, of variable severity and presentations. The mean duration until the clinical documentation of acetabular wear-induced pain was 25.6 months (range, 1.4–146.4 months), with most symptomatic hip erosion cases presented within the first year (50% presented within the first 6 months) after the index HA surgery. Despite that, only five cases underwent conversion to THA (0.2%) while seven patients received conservative management. Conclusions: In patients older than 65 years of age who are candidates for HA, the incidence of symptomatic hip erosion and the subsequent conversion to THA is low and hence HA remains a viable treatment option. Based on the duration until clinical presentation of acetabular erosion, this study suggests that the state of acetabular cartilage during surgery may influence the development of early acetabular wear in most symptomatic patients. On the other hand, hip hemiarthroplasty is a rare treatment option for displaced femoral neck fractures in patients younger than 65 years of age, and its use in this patient subset depends on conditional and patient-related factors such as the activity level, cognitive function, and medical comorbidities.

Gout Flare Burden in the United States: A Multiyear Cross-Sectional Survey Study.

The objective of this study was to assess the burden of gout flares and examine associated patient characteristics and outcomes in a sample of US adults. Data were collected via an online survey of US adults ≥18 years using a random stratified sampling framework. Participants with gout completed questions about treatments, serum urate (SU) levels, severity, satisfaction with control, and gout flares. All participants completed the Veterans RAND 12-Item Health Survey, the Generalized Anxiety Disorder 7-Item Scale, and the Patient Health Questionnaire 9-Item Scale. Data were summarized using descriptive statistics. Multivariable-adjusted logistic regression analyses examined factors predictive of reporting gout flares to a physician. A total of 933 participants met the study criteria for having gout. Those with gout tended to be older (58.3 [SD 13.3] years vs 45.4 [SD 16.1] years; P < 0.001), male (76.3% vs 46.9%; P < 0.001), White (80.5% vs 76.8%; P = 0.01), and married or living with their partner (58.9% vs 52.8%; P < 0.001) compared with those without gout (n = 30,146). The total gout flare burden for those with gout was 6.6 gout flares per year. Nearly 72% of gout flares were either not reported to physicians or pretreated or prevented. Characteristics of those who were less likely to report gout flares included being younger, being less educated, having a lower Charlson Comorbidity Index score, not being diagnosed with gout by their doctor, and not taking a urate-lowering therapy. This study confirmed that gout flares are common in US adults with gout and found that gout flares are underreported. Reliance on clinical documentation of physician-reported gout flares is insufficient to assess the true patient burden of gout.

Natural Language Processing of Clinical Documentation to Assess Functional Status in Patients With Heart Failure

Serial functional status assessments are critical to heart failure (HF) management but are often described narratively in documentation, limiting their use in quality improvement or patient selection for clinical trials. To develop and validate a deep learning natural language processing (NLP) strategy for extracting functional status assessments from unstructured clinical documentation. This diagnostic study used electronic health record data collected from January 1, 2013, through June 30, 2022, from patients diagnosed with HF seeking outpatient care within 3 large practice networks in Connecticut (Yale New Haven Hospital [YNHH], Northeast Medical Group [NMG], and Greenwich Hospital [GH]). Expert-annotated notes were used for NLP model development and validation. Data were analyzed from February to April 2024. Development and validation of NLP models to detect explicit New York Heart Association (NYHA) classification, HF symptoms during activity or rest, and frequency of functional status assessments. Outcomes of interest were model performance metrics, including area under the receiver operating characteristic curve (AUROC), and frequency of NYHA class documentation and HF symptom descriptions in unannotated notes. This study included 34 070 patients with HF (mean [SD] age 76.1 [12.6] years; 17 728 [52.0]% female). Among 3000 expert-annotated notes (2000 from YNHH and 500 each from NMG and GH), 374 notes (12.4%) mentioned NYHA class and 1190 notes (39.7%) described HF symptoms. The NYHA class detection model achieved a class-weighted AUROC of 0.99 (95% CI, 0.98-1.00) at YNHH, the development site. At the 2 validation sites, NMG and GH, the model achieved class-weighted AUROCs of 0.98 (95% CI, 0.96-1.00) and 0.98 (95% CI, 0.92-1.00), respectively. The model for detecting activity- or rest-related symptoms achieved an AUROC of 0.94 (95% CI, 0.89-0.98) at YNHH, 0.94 (95% CI, 0.91-0.97) at NMG, and 0.95 (95% CI, 0.92-0.99) at GH. Deploying the NYHA model among 182 308 unannotated notes from the 3 sites identified 23 830 (13.1%) notes with NYHA mentions, specifically 10 913 notes (6.0%) with class I, 12 034 notes (6.6%) with classes II or III, and 883 notes (0.5%) with class IV. An additional 19 730 encounters (10.8%) could be classified into functional status groups based on activity- or rest-related symptoms, resulting in a total of 43 560 medical notes (23.9%) categorized by NYHA, an 83% increase compared with explicit mentions alone. In this diagnostic study of 34 070 patients with HF, the NLP approach accurately extracted a patient's NYHA symptom class and activity- or rest-related HF symptoms from clinical notes, enhancing the ability to track optimal care delivery and identify patients eligible for clinical trial participation from unstructured documentation.

Quantifying Plastic and Reconstructive Surgery Engagement in the Evolution of ICD-10 Codes.

International Classification of Diseases (ICD)-10 code specificity is important for detailed diagnostic documentation. Insufficient coding granularity in plastic surgery may hinder accurate clinical documentation, impacting education, research, and patient care. This study examines ICD-10 code modification efforts by plastic surgeons and plastic surgery organizations within the last decade, subsequently detailing the formal proposal process and highlighting potential areas where coding modifications may be explored. A retrospective review of all modification proposals presented to the ICD Coordination and Maintenance Committee from March 2013 to March 2023 was conducted. Plastic surgery-related proposals were identified and evaluated for successful implementation. These results were compared to those of other specialties using Fisher exact tests and 2-tailed t tests. A total of 472 unique proposals were identified and reviewed. Surgeons and/or surgical organizations contributed to 111 (23.5%) novel proposals, 41 (36.9%) of which were initiated by surgeons. Nine surgical specialties and 14 supporting organizations were represented. The success rate for all surgical-related proposals was 85.6%, with an average time to implementation of 24.7 months. Only 2 (1.8%) of all surgical-related proposals pertained to plastic surgery, both of which were successfully implemented. Though a few plastic surgeons have taken advantage of the ICD code revision process within the last decade, the high overall success rate for surgical-related proposals suggests that this seems to be a feasible method by which plastic surgeons and plastic surgery organizations can bring about useful coding changes that meaningfully impact clinical practice.

Disparities in time to diagnosis of Radiographic Axial Spondyloarthritis.

Radiographic axial spondyloarthritis (r-axSpA) has a 7-year average diagnostic delay. While the impact of sex or gender on time to diagnosis has been evaluated, the role of social determinants of health remains understudied. We assessed whether time from initial clinical documentation of r-axSpA symptoms to r-axSpA diagnosis (diagnostic delay) varies based on sex, race, ethnicity, or the presence of social needs. We studied patients with r-axSpA from a tertiary center from 2000 to 2022. The cohort was built with the Observational Health Data Sciences and Informatics (OHDSI) network. For the primary analysis, we assessed the time from back pain/spinal pain to r-axSpA diagnosis and secondarily, the time to r-axSpA from any other r-axSpA -related condition. To estimate differences in diagnostic delay, we employed a parametric survival model, accelerated failure time. We included 404 patients (mean age 49, 37% female), with 25.5% identifying as Black, 31.1% as other or unknown race and 14.1% as Hispanic. Patients with a documented social need had a 21% increase in time from back pain to r-axSpA diagnosis (95% CI 0.93, 1.56). In patients with any r-axSpA related condition, time to diagnosis increased similarly by 21% (95% CI 0.92, 1.57). Considering an average time to diagnosis of 34 months, a social need increased time to diagnosis by 7 months. This study reveals a trend toward diagnostic delay in r-axSpA related to social need, sex, race, and ethnicity. Future studies should focus on referral strategies to enable prompt diagnosis and optimize care.

Evaluation and Use of Natural Language Processing (NLP) Reasoning and Classification Models to Support Clinical Trial Patient Identification and Enrollment in the Community Oncology Setting

Clinical trial research in oncology relies heavily on clinical documentation within the electronic medical record (EMR) to ascertain patient eligibility in clinical trials based on inclusion and exclusion criteria. The structured data elements within the EMR serve as the primary information source for defining patient cohorts, with clinical cancer stage and performance status being two pivotal criteria determining trial eligibility. The challenge arises from the inconsistent availability of clinical stage and performance status data within the structured fields of the EMR despite their consistent presence in clinical notes. Additionally, there is a deficiency of standardization of this data that exists in the unstructured field. Hence, due to lack of structured data and standardization of said data, there are limitations in developing artificial intelligence (AI) models. To increase the comprehensiveness of clinical records, a clinical research team at a community oncology practice was consulted to identify requirements and extract essential clinical features from de-identified data. The methods outlined in this paper focused on eliminating false positives to allow future development of Large Language Models (LLM) using the outputted structured fields which resulted in an increase in patient record completeness with high accuracy. The accuracy ranged from 97.5-97.75% for the models that were developed. Out of the 60,000+ patients, the numerical staging, TNM (tumor, node, metastasis) staging, and Karnofsky performance score models added a structured field for 29.62%, 21.01%, and 40.64% patients respectively. Additionally, a semi-supervised NLP algorithm was applied on the performance status algorithm which achieved a mean absolute error (MAE) of 1.57. This work demonstrates the use case of natural language processing (NLP) in optimizing the clinical research enrollment process by providing an efficient and accurate method to detect key clinical values in unstructured patient data. Similar methodology with more advanced algorithms such as LLM can be employed to detect additional patient elements such as molecular biomarkers, imaging reports, postoperative surgical outcomes (i.e., clear margins etc.) and patient treatment outcomes using the extracted structured fields.

A novel digital health approach to improving global pediatric sepsis care in Bangladesh using wearable technology and machine learning.

Sepsis is the leading cause of child death globally with low- and middle-income countries (LMICs) bearing a disproportionate burden of pediatric sepsis deaths. Limited diagnostic and critical care capacity and health worker shortages contribute to delayed recognition of advanced sepsis (severe sepsis, septic shock, and/or multiple organ dysfunction) in LMICs. The aims of this study were to 1) assess the feasibility of a wearable device for physiologic monitoring of septic children in a LMIC setting and 2) develop machine learning models that utilize readily available wearable and clinical data to predict advanced sepsis in children. This was a prospective observational study of children with sepsis admitted to an intensive care unit in Dhaka, Bangladesh. A wireless, wearable device linked to a smartphone was used to collect continuous recordings of physiologic data for the duration of each patient's admission. The correlation between wearable device-collected vital signs (heart rate [HR], respiratory rate [RR], temperature [T]) and manually collected vital signs was assessed using Pearson's correlation coefficients and agreement was assessed using Bland-Altman plots. Clinical and laboratory data were used to calculate twice daily pediatric Sequential Organ Failure Assessment (pSOFA) scores. Ridge regression was used to develop three candidate models for advanced sepsis (pSOFA > 8) using combinations of clinical and wearable device data. In addition, the lead time between the models' detection of advanced sepsis and physicians' documentation was compared. 100 children were enrolled of whom 41% were female with a mean age of 15.4 (SD 29.6) months. In-hospital mortality rate was 24%. Patients were monitored for an average of 2.2 days, with > 99% data capture from the wearable device during this period. Pearson's r was 0.93 and 0.94 for HR and RR, respectively) with r = 0.72 for core T). Mean difference (limits of agreement) was 0.04 (-14.26, 14.34) for HR, 0.29 (-5.91, 6.48) for RR, and -0.0004 (-1.48, 1.47) for core T. Model B, which included two manually measured variables (mean arterial pressure and SpO2:FiO2) and wearable device data had excellent discrimination, with an area under the Receiver-Operating Curve (AUC) of 0.86. Model C, which consisted of only wearable device features, also performed well, with an AUC of 0.78. Model B was able to predict the development of advanced sepsis more than 2.5 hours earlier compared to clinical documentation. A wireless, wearable device was feasible for continuous, remote physiologic monitoring among children with sepsis in a LMIC setting. Additionally, machine-learning models using wearable device data could discriminate cases of advanced sepsis without any laboratory tests and minimal or no clinician inputs. Future research will develop this technology into a smartphone-based system which can serve as both a low-cost telemetry monitor and an early warning clinical alert system, providing the potential for high-quality critical care capacity for pediatric sepsis in resource-limited settings.

Fifty shades of green and blue: autopsy findings after administration of xenobiotics.

Unusual findings during an autopsy may come from peculiarities in the position, shape, size, weight, consistency, smell or color of organs. The following study was triggered by an autopsy case in which an unusual blue-green discoloration of organs, which changed during the autopsy, was noticed. A review of the local autopsy database, selected cases including the antemortem clinical documentation and the literature has been performed to clarify the etiology of these conspicuous discolorations and to evaluate their diagnostic value. The study showed that certain xenobiotics may lead to such discoloration. After systemic administration of methylene blue, darkening blue-green discoloration of organs, especially the brain and heart, can be observed. In addition, the systemic administration of toluidine blue also appears to be capable of causing such discoloration. Beyond that, drugs (like Rohypnol®) or other foreign substances (like detergents) containing warning colors, i.e. indigocarmin (E132) or Brilliant Blue FCF (E133) may cause discolorations of the upper gastrointestinal tract or the urinary bladder respectively. A blue-green, possibly darkening discoloration of organs during autopsy may point towards an antemortem administration of certain xenobiotics. The affected organs give an indication of the possible route of application and the type of substance. A differentiated interpretation of the etiology of such conspicuous discolorations at autopsy should only be made considering the (medical) history and, if necessary, complementary (toxicological) examinations.

Employment, work ability and sick leave in melanoma patients within the first year of diagnosis.

Malignant melanoma affects younger working individuals. This study investigated work ability and sick leave within one year after diagnosis, as well as the impact of rehabilitation and psycho-oncological measures on employment outcomes. In this monocentric, prospective, observational cohort study, 221 patients (62.4% females), aged 19-65, participated. 78.5% had tumor stageIB or lower at baseline. Routine clinical documentation, occupational history, work ability, subjective prognosis of employment, need for and satisfaction with rehabilitation measures were repeatedly assessed. 181 patients (82%) were employed at first visit, 172 (78%) at last visit. Approximately 75% of patients initially rated work ability as "moderate", with up to 9 sick leave days. In the following year, sick days increased significantly in patients with stageIB and IIA (p=0.044) and highly significantly in patients with stageII B and above (p<0.001). Psycho-oncologic consultation (24%) and rehabilitation (18%) did not change the significantly worse self-rated work ability of these patients. Melanoma affects work ability, even in tumor stages without lymph node involvement or distant metastasis. Controlled clinical trials would be useful to evaluate the success of rehabilitation and psycho-oncological interventions for patients with melanoma.

Automated Extraction of Stroke Severity From Unstructured Electronic Health Records Using Natural Language Processing.

Multicenter electronic health records can support quality improvement and comparative effectiveness research in stroke. However, limitations of electronic health record-based research include challenges in abstracting key clinical variables, including stroke severity, along with missing data. We developed a natural language processing model that reads electronic health record notes to directly extract the National Institutes of Health Stroke Scale score when documented and predict the score from clinical documentation when missing. The study included notes from patients with acute stroke (aged ≥18 years) admitted to Massachusetts General Hospital (2015-2022). The Massachusetts General Hospital data were divided into training/holdout test (70%/30%) sets. We developed a 2-stage model to predict the admission National Institutes of Health Stroke Scale, obtained from the GWTG (Get With The Guidelines) stroke registry. We trained a model with the least absolute shrinkage and selection operator. For test notes with documented National Institutes of Health Stroke Scale, scores were extracted using regular expressions (stage 1); when not documented, least absolute shrinkage and selection operator was used for prediction (stage 2). The 2-stage model was tested on the holdout test set and validated in the Medical Information Mart for Intensive Care (2001-2012) version 1.4, using root mean squared error and Spearman correlation. We included 4163 patients (Massachusetts General Hospital, 3876; Medical Information Mart for Intensive Care, 287); average age, 69 (SD, 15) years; 53% men, and 72% White individuals. The model achieved a root mean squared error of 2.89 (95% CI, 2.62-3.19) and Spearman correlation of 0.92 (95% CI, 0.91-0.93) in the Massachusetts General Hospital test set, and 2.20 (95% CI, 1.69-2.66) and 0.96 (95% CI, 0.94-0.97) in the MIMIC validation set, respectively. The automatic natural language processing-based model can enable large-scale stroke severity phenotyping from the electronic health record and support real-world quality improvement and comparative effectiveness studies in stroke.

Anxious/depressed individuals exhibit disrupted frontotemporal synchrony during cognitive conflict encoding.

Anxiety and depressive disorders are associated with cognitive control deficits, yet their underlying neural mechanisms remain poorly understood. Here, we used high-resolution stereotactic EEG (sEEG) to determine how anxiety and/or depression modulates neural and behavioral responses when cognitive control is engaged in individuals with medically refractory epilepsy undergoing sEEG monitoring for surgical evaluation. We analyzed sEEG data recorded from frontotemporal regions of 29 participants (age range: 19-55, mean age: 35.5, female: 16/29) while they performed a Multi-Source Interference Task (MSIT) designed to elicit cognitive conflict. Neurobehavioral interviews, symptom rating scales, and clinical documentation were used to categorize participants as demonstrating anxiety and/or depression symptoms (A/D, n=13) or as epilepsy controls (EC, n=16). Generalized linear mixed-effects (GLME) models were used to analyze behavioral and neural data. Models of oscillatory power were used to identify brain regions within conflict-encoding networks in which coherence and phase locking values (PLV) were examined in A/D and EC. A/D participants demonstrated a greater conflict effect (response time slowing with higher cognitive load), without impairment in response time (RT) or accuracy compared to EC. A/D participants also showed significantly enhanced conflict-evoked theta (4-8Hz) and alpha (8-15Hz) power in the dorsolateral prefrontal cortex (dlPFC) and amygdala as well as widespread broadband activity in the lateral temporal lobe (LTL) compared to EC. Additionally, theta coherence and PLV between dlPFC-LTL and dlPFC-amygdala were reduced by conflict in A/D. Our findings suggest individuals with anxiety/depression symptoms exhibit heightened frontotemporal oscillatory activity and disrupted frontotemporal synchrony during cognitive conflict encoding, which may indicate a greater need for cognitive resources due to ineffective cognitive processing. These results highlight a potential role of frontotemporal circuits in conflict encoding that are altered in anxiety/depression, and may further inform future therapeutic interventions aimed at enhancing cognitive control in these populations.

Occurrence of Stigmatizing Documentation Among Hospital Medicine Encounters With Opioid-Related Diagnosis Codes: Cohort Study.

Physician use of stigmatizing language in the clinical documentation of hospitalized adults with opioid use is common. However, patient factors associated with stigmatizing language in this setting remain poorly characterized. This study aimed to determine whether specific demographic factors and clinical outcomes are associated with the presence of stigmatizing language by physicians in the clinical documentation of encounters with opioid-related ICD-10 (International Statistical Classification of Diseases, Tenth Revision) codes. Hospital encounters with one or more associated opioid-related ICD-10 admission diagnoses on the hospital medicine service during the 2020 calendar year were analyzed for the presence of stigmatizing language in history and physical and discharge summaries. Multivariable adjusted logistic regression models were used to determine associations of age, race, gender, medication for addiction treatment use, against medical advice discharge, homelessness, comorbid polysubstance use, comorbid psychiatric disorder, comorbid chronic pain, cost, and 30-day readmission with the presence of stigmatizing language. A total of 221 encounters were identified, of which 64 (29%) encounters had stigmatizing language present in physician documentation. Most stigmatizing language was due to use of "substance abuse" rather than the preferred term "substance use" (63/66 instances). Polysubstance use and homelessness were independently associated with the presence of stigmatizing language (adjusted odds ratio [aOR] 7.83; 95% CI 3.42-19.24 and aOR 2.44; 95% CI 1.03-5.90) when controlling for chronic pain and other covariates. Among hospital medicine encounters with an opioid-related diagnosis, stigmatizing language by physicians in clinical documentation was common and independently associated with comorbid polysubstance use and homelessness.

Assessing Retrieval-Augmented Large Language Model Performance in Emergency Department ICD-10-CM Coding Compared to Human Coders.

Accurate medical coding is essential for clinical and administrative purposes but complicated, time-consuming, and biased. This study compares Retrieval-Augmented Generation (RAG)-enhanced LLMs to provider-assigned codes in producing ICD-10-CM codes from emergency department (ED) clinical records. Retrospective cohort study using 500 ED visits randomly selected from the Mount Sinai Health System between January and April 2024. The RAG system integrated past 1,038,066 ED visits data (2021-2023) into the LLMs' predictions to improve coding accuracy. Nine commercial and open-source LLMs were evaluated. The primary outcome was a head-to-head comparison of the ICD-10-CM codes generated by the RAG-enhanced LLMs and those assigned by the original providers. A panel of four physicians and two LLMs blindly reviewed the codes, comparing the RAG-enhanced LLM and provider-assigned codes on accuracy and specificity. RAG-enhanced LLMs demonstrated superior performance to provider coders in both the accuracy and specificity of code assignments. In a targeted evaluation of 200 cases where discrepancies existed between GPT-4 and provider-assigned codes, human reviewers favored GPT-4 for accuracy in 447 instances, compared to 277 instances where providers' codes were preferred (p<0.001). Similarly, GPT-4 was selected for its superior specificity in 509 cases, whereas human coders were preferred in only 181 cases (p<0.001). Smaller open-access models, such as Llama-3.1-70B, also demonstrated substantial scalability when enhanced with RAG, with 218 instances of accuracy preference compared to 90 for providers' codes. Furthermore, across all models, the exact match rate between LLM-generated and provider-assigned codes significantly improved following RAG integration, with Qwen-2-7B increasing from 0.8% to 17.6% and Gemma-2-9b-it improving from 7.2% to 26.4%. RAG-enhanced LLMs improve medical coding accuracy in EDs, suggesting clinical workflow applications. These findings show that generative AI can improve clinical outcomes and reduce administrative burdens. This work was supported in part through the computational and data resources and staff expertise provided by Scientific Computing and Data at the Icahn School of Medicine at Mount Sinai and supported by the Clinical and Translational Science Awards (CTSA) grant UL1TR004419 from the National Center for Advancing Translational Sciences. Research reported in this publication was also supported by the Office of Research Infrastructure of the National Institutes of Health under award number S10OD026880 and S10OD030463. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The funders played no role in study design, data collection, analysis and interpretation of data, or the writing of this manuscript. A study showed AI models with retrieval-augmented generation outperformed human doctors in ED diagnostic coding accuracy and specificity. Even smaller AI models perform favorably when using RAG. This suggests potential for reducing administrative burden in healthcare, improving coding efficiency, and enhancing clinical documentation.