Free-text Narratives Research Articles

Understanding the risk factors for student attrition across pre-registration nursing and midwifery programmes in a United Kingdom university: A sequential explanatory mixed methods study

BackgroundThe World Health Organization estimates a global shortage of 4.5 million nurses and 0.31 million midwives by 2030, which presents a significant problem for workforce planning. Despite aspirational United Kingdom (UK) government policy targets and ongoing efforts, the number of nursing and midwifery vacancies continues to increase. One contributing factor is student nurse and midwife attrition.AimTo identify risk factors that contribute to nursing and midwifery students interrupting or withdrawing from pre-registration programmes. DesignSequential explanatory mixed-methods design. SettingA University School of Nursing and Midwifery in the Southwest of England. Participants534 pre-registration nursing and midwifery student survey respondents and 8 focus group participants. MethodsStudents' perspectives of attrition risk factors were evaluated using a Likert scale with free text questions used to capture additional factors affecting interruption or withdrawal. Students were also invited to take part in focus group discussions. Survey data were coded and analysed utilising SPSS (Version 25) software. Survey free text narrative and focus group data were analysed using reflexive thematic analysis. ResultsSurvey results revealed 55 students had interrupted and returned and 281 students had considered leaving or interrupting their programme. Survey descriptive statistics data were grouped into top attrition risk factors. Free text narratives and focus group data identified four themes: multiple and overlapping challenges; the threats and opportunities of clinical placements; the availability of the right support at the right time and investing for a better future for myself and others. ConclusionNursing and Midwifery students reported challenges across their pre-registration programmes resulting in risk factors that might lead to students interrupting or withdrawing. Optimal cooperation between higher education institutions (HEI) and clinical stakeholders is of paramount importance to address these challenges as is government and Nursing and Midwifery Council (NMC) acknowledgement and consideration of the identified risks linked to attrition; specifically relating to national funding arrangements and the clinical practice hours required for registration within the UK. This may also resonate with professional programmes internationally, faced with similar staffing shortages.

Automated Transformation of Unstructured Cardiovascular Diagnostic Reports into Structured Datasets Using Sequentially Deployed Large Language Models.

Rich data in cardiovascular diagnostic testing are often sequestered in unstructured reports, with the necessity of manual abstraction limiting their use in real-time applications in patient care and research. We developed a two-step process that sequentially deploys generative and interpretative large language models (LLMs; Llama2 70b and Llama2 13b). Using a Llama2 70b model, we generated varying formats of transthoracic echocardiogram (TTE) reports from 3,000 real-world echo reports with paired structured elements, leveraging temporal changes in reporting formats to define the variations. Subsequently, we fine-tuned Llama2 13b using sequentially larger batches of generated echo reports as inputs, to extract data from free-text narratives across 18 clinically relevant echocardiographic fields. This was set up as a prompt-based supervised training task. We evaluated the fine-tuned Llama2 13b model, HeartDx-LM, on several distinct echocardiographic datasets: (i) reports across the different time periods and formats at Yale New Haven Health System (YNHHS), (ii) the Medical Information Mart for Intensive Care (MIMIC) III dataset, and (iii) the MIMIC IV dataset. We used the accuracy of extracted fields and Cohen's Kappa as the metrics and have publicly released the HeartDX-LM model. The HeartDX-LM model was trained on randomly selected 2,000 synthetic echo reports with varying formats and paired structured labels, with a wide range of clinical findings. We identified a lower threshold of 500 annotated reports required for fine-tuning Llama2 13b to achieve stable and consistent performance. At YNHHS, the HeartDx-LM model accurately extracted 69,144 out of 70,032 values (98.7%) across 18 clinical fields from unstructured reports in the test set from contemporary records where paired structured data were also available. In older echo reports where only unstructured reports were available, the model achieved 87.1% accuracy against expert annotations for the same 18 fields for a random sample of 100 reports. Similarly, in expert-annotated external validation sets from MIMIC-IV and MIMIC-III, HeartDx-LM correctly extracted 201 out of 220 available values (91.3%) and 615 out of 707 available values (87.9%), respectively, from 100 randomly chosen and expert annotated echo reports from each set. We developed a novel method using paired large and moderate-sized LLMs to automate the extraction of unstructured echocardiographic reports into tabular datasets. Our approach represents a scalable strategy that transforms unstructured reports into computable elements that can be leveraged to improve cardiovascular care quality and enable research.

Visualizing Nursing Narratives: An Evaluation of Latent Dirichlet Allocation Topic Modeling for Care Reports.

The increasing volume of unstructured textual data in healthcare, particularly in nursing care reports, presents both challenges and opportunities for enhancing patient care and operational efficiency. This study explores the application of Latent Dirichlet Allocation (LDA) topic modeling to analyze free-text nursing narratives from inpatient stays in three different clinics, aiming to uncover the latent thematic structures within. Utilizing the R programming environment and the visualization tool LDAvis, we identified three main themes: "Patient Well-being," "Patient Mobility and Care Activities," and "Treatment and Pain Management," the latter combining two closely related but initially distinct topics due to their overlapping content. Our findings demonstrate the potential of LDA topic modeling in extracting meaningful insights from nursing narratives, which could inform patient care strategies and healthcare practices. However, the study also highlights significant challenges associated with the method, including the sensitivity to parameter settings, the lack of updates for key software packages, and concerns about reproducibility. These issues highlight the need for meticulous parameter validation and the exploration of alternative text analysis methodologies for future research. By addressing these methodological challenges and emphasizing the importance of comparative method analysis, this study contributes to the advancement of text analytics in healthcare. It opens avenues for further research aimed at developing more robust, efficient, and accessible tools for analyzing free-text data, thereby enhancing the ability of healthcare professionals to use unstructured data to improve decision making and patient outcomes.

Incorporating Patient Values in Large Language Model Recommendations for Surrogate and Proxy Decisions.

Surrogates, proxies, and clinicians making shared treatment decisions for patients who have lost decision-making capacity often fail to honor patients' wishes, due to stress, time pressures, misunderstanding patient values, and projecting personal biases. Advance directives intend to align care with patient values but are limited by low completion rates and application to only a subset of medical decisions. Here, we investigate the potential of large language models (LLMs) to incorporate patient values in supporting critical care clinical decision-making for incapacitated patients in a proof-of-concept study. We simulated text-based scenarios for 50 decisionally incapacitated patients for whom a medical condition required imminent clinical decisions regarding specific interventions. For each patient, we also simulated five unique value profiles captured using alternative formats: numeric ranking questionnaires, text-based questionnaires, and free-text narratives. We used pre-trained generative LLMs for two tasks: 1) text extraction of the treatments under consideration and 2) prompt-based question-answering to generate a recommendation in response to the scenario information, extracted treatment, and patient value profiles. Model outputs were compared with adjudications by three domain experts who independently evaluated each scenario and decision. Automated extractions of the treatment in question were accurate for 88% (n = 44/50) of scenarios. LLM treatment recommendations received an average Likert score by the adjudicators of 3.92 of 5.00 (five being best) across all patients for being medically plausible and reasonable treatment recommendations, and 3.58 of 5.00 for reflecting the documented values of the patient. Scores were highest when patient values were captured as short, unstructured, and free-text narratives based on simulated patient profiles. This proof-of-concept study demonstrates the potential for LLMs to function as support tools for surrogates, proxies, and clinicians aiming to honor the wishes and values of decisionally incapacitated patients.

Drug Misuse Among Pediatric Patients: Encounters in the Prehospital Field.

Few data describing prehospital pediatric substance misuse exist. The objective of this study was to characterize substance misuse in North Carolina pediatric patients receiving emergency medical services (EMS) care. We conducted a retrospective cross-sectional study of patients aged < 16 years transported by EMS for substance misuse. Data were obtained from a statewide repository including patients treated for confirmed or suspected misuse of marijuana, alcohol, benzodiazepines, opioids, stimulants, acetaminophen, antidepressants, or other drugs. Of 45,855 EMS encounters, 448 patients misused drugs. Most patients were female (56.7%), White (50.9%), adolescent (73.8%), and resided in Central North Carolina (54.0%). A plurality of cases was rural (46.2%), followed by urban (27.9%) and regional city/suburban counties (25.7%). Calls most often originated in residential locations (68.5%). Drugs identified during EMS calls included alcohol (24.2%), marijuana (24.2%), benzodiazepines (8.8%), antidepressants (8.8%), stimulants (6.8%), opioids (5.9%), and other medications (17.6%). Motivations for drug use were recreational (58.4%), self-harm (24.6%), and accidental (17.0%). Due to insufficient data documentation, one EMS system was removed from analysis. Additionally, misclassification of type of drug or intention of drug use is possible due to the use of free-text patient narratives. EMS responded to a vast variety of drug misuse among pediatric patients including prescription medications, alcohol, marijuana, and illicit drugs. Accidental ingestions occurred exclusively in infant/preschool ages and intent for recreation or self-harm primarily occurred in adolescents. By increasing awareness of the more common pediatric patient characteristics associated with the type and reason for drug use, EMS agencies can improve pediatric readiness among prehospital clinicians.

Explainable text-tabular models for predicting mortality risk in companion animals

As interest in using machine learning models to support clinical decision-making increases, explainability is an unequivocal priority for clinicians, researchers and regulators to comprehend and trust their results. With many clinical datasets containing a range of modalities, from the free-text of clinician notes to structured tabular data entries, there is a need for frameworks capable of providing comprehensive explanation values across diverse modalities. Here, we present a multimodal masking framework to extend the reach of SHapley Additive exPlanations (SHAP) to text and tabular datasets to identify risk factors for companion animal mortality in first-opinion veterinary electronic health records (EHRs) from across the United Kingdom. The framework is designed to treat each modality consistently, ensuring uniform and consistent treatment of features and thereby fostering predictability in unimodal and multimodal contexts. We present five multimodality approaches, with the best-performing method utilising PetBERT, a language model pre-trained on a veterinary dataset. Utilising our framework, we shed light for the first time on the reasons each model makes its decision and identify the inclination of PetBERT towards a more pronounced engagement with free-text narratives compared to BERT-base’s predominant emphasis on tabular data. The investigation also explores the important features on a more granular level, identifying distinct words and phrases that substantially influenced an animal’s life status prediction. PetBERT showcased a heightened ability to grasp phrases associated with veterinary clinical nomenclature, signalling the productivity of additional pre-training of language models.

Recording of suspected adverse drug reaction reporting in veterinary free-text clinical narratives.

To use text mining approaches to identify instances of suspected adverse drug reactions recorded in first opinion veterinary free-text clinical narratives, and to evaluate whether these were also reported to either the Veterinary Medicines Directorate or the relevant Marketing Authorisation holder in order to derive an estimate of the suspected adverse drug reaction (sADR) minimum under-reporting rate. To characterise sADR reports and explore whether particular features are associated with report submission. Two regular expressions were developed to identify mentions of "adverse drug reactions" and "side effects" in the free-text clinical narratives of electronic health records contained within the Small Animal Veterinary Surveillance Network database. Consultations containing a match for the developed regular expressions were manually reviewed for inclusion and further classified to determine the suspected product, seriousness and expectedness of the event, and an indication of whether the event had been reported. The associations between event characteristics and reporting were explored using Fisher's exact tests. A total of 10,565 records were manually reviewed from which 827 sADRs were identified. Approximately 90% of these sADRs were not recorded as reported. Suspected adverse drug reactions that were not considered "expected" were recorded as reported more frequently than "expected" sADRs. However, clinical severity did not appear to impact on whether there was a record of reporting. This is the first estimate of under reporting sADRs based on real world evidence from veterinary clinical records. The under-reporting rate implied by this study highlights that further interventions are required to improve reporting rate within the veterinary profession in order to support pharmacovigilance activities and improve drug safety.

A Natural Language Processing Framework for Detecting Adverse Drug Reactions in Clinical Structured Drug Reviews

Background: Adverse drug reactions (ADRs) pose significant concerns in healthcare, yet their underreporting remains a challenge. Extracting spontaneous and non-automatic reports from free-text narratives contributes to this low rate of reporting. An automatic ADR detection system can mitigate these issues by identifying, summarizing, and reporting ADRs in a document. This study presents an adverse drug reaction detector (ADRD), a natural language processing (NLP) framework applied to the psychiatric treatment adverse reactions (PsyTAR) dataset. Aiming to automate ADR analysis, the framework explores the relationship between ADRs and patient satisfaction. Methods: A comprehensive eight-phase approach was employed in the ADRD framework, utilizing Python programming language libraries and NLP tools. The dataset underwent meticulous preprocessing, and the subsequent phases involved data summarization, pattern identification, data cleaning, sentiment calculation, assessment of drug effectiveness and usefulness, analysis of medical conditions, and identification of the most effective and ineffective drugs for each condition. Results: Analyzing 891 comments related to four unique drugs (i.e., Zoloft, Lexapro, Cymbalta, and Effexor XR) from patients with 285 distinct conditions, the framework offered insights into the dataset structure, statistical indicators, distribution of ratings and ADR counts, the impact of ratings on ADR counts, and length of comments’ influence on ratings. Conclusion: The challenges of extracting ADR reports from free-text narratives have led to their underreporting. ADRD offers an automated and insightful approach for enhancing ADR analysis and reporting processes, making strides toward bridging the gap in ADR reporting.

“I knew it was a scam”: Understanding the triggers for recognizing romance fraud

AbstractResearch SummaryThis article investigates the contributing factors (or triggers) to a realization of romance fraud victimization, based on 1015 reports lodged with Scamwatch (Australian online reporting portal for fraud) between July 2018 and July 2019 (inclusive). The article examines the free text narrative of each report to propose five discernible trigger categories: further requests for money; characteristics of communications; verification checks; an offender's action(s); and being told by a third party.Policy ImplicationsBased on a comprehensive understanding of these five categories, the article advocates for broader messaging approaches to encompass financial literacy and well‐being, cyberliteracy and critical thinking skills, cybersecurity practices, and respectful and healthy relationships. Leveraging the use of these wider education and awareness campaigns could improve recognition of romance fraud. Importantly, it is suggested that banks and other financial institutions are in a strong position to focus their efforts on some of these broader messages to effect positive change.

Social Determinants of Health in EMS Records: A Mixed-methods Analysis Using Natural Language Processing and Qualitative Content Analysis.

Introduction: Social determinants of health (SDoH) are known to impact the health and well-being of patients. However, information regarding them is not always collected in healthcare interactions, and healthcare professionals are not always well-trained or equipped to address them. Emergency medical services (EMS) professionals are uniquely positioned to observe and attend to SDoH because of their presence in patients' environments; however, the transmission of that information may be lost during transitions of care. Documentation of SDoH in EMS records may be helpful in identifying and addressing patients' insecurities and improving their health outcomes. Our objective in this study was to determine the presence of SDoH information in adult EMS records and understand how such information is referenced, appraised, and linked to other determinants by EMS personnel. Methods: Using EMS records for adult patients in the 2019 ESO Data Collaborative public-use research dataset using a natural language processing (NLP) algorithm, we identified free-text narratives containing documentation of at least one SDoH from categories associated with food, housing, employment, insurance, financial, and social support insecurities. From the NLP corpus, we randomly selected 100 records from each of the SDoH categories for qualitative content analysis using grounded theory. Results: Of the 5,665,229 records analyzed by the NLP algorithm, 175,378 (3.1%) were identified as containing at least one reference to SDoH. References to those SDoH were centered around the social topics of accessibility, mental health, physical health, and substance use. There were infrequent explicit references to other SDoH in the EMS records, but some relationships between categories could be inferred from contexts. Appraisals of patients' employment, food, and housing insecurities were mostly negative. Narratives including social support and financial insecurities were less negatively appraised, while those regarding insurance insecurities were mostly neutral and related to EMS operations and procedures. Conclusion: The social determinants of health are infrequently documented in EMS records. When they are included, they are infrequently explicitly linked to other SDoH categories and are often negatively appraised by EMS professionals. Given their unique position to observe and share patients' SDoH information, EMS professionals should be trained to understand, document, and address SDoH in their practice.

Using motor vehicle crash records for injury surveillance and research in agriculture and forestry

Problem: Fatal injuries in the agriculture, forestry, and fishing sector (AgFF) outweigh those across all sectors in the United States. Transportation-related injuries are among the top contributors to these fatal events. However, traditional occupational injury surveillance systems may not completely capture crashes involving farm vehicles and logging trucks, specifically nonfatal events. Methods: The study aimed to develop an integrated database of AgFF-related motor-vehicle crashes for the southwest (Arkansas, Louisiana, New Mexico, Oklahoma, and Texas) and to use these data to conduct surveillance and research. Lessons learned during the pursuit of these aims were cataloged. Activities centered around the conduct of traditional statistical and geospatial analyses of structured data fields and natural language processing of free-text crash narratives. Results: The structured crash data in each state include fields that allowed farm vehicles or equipment and logging trucks to be identified. The variable definitions and coding were not consistent across states but could be harmonized. All states recorded data fields pertaining to person, vehicle, and crash/environmental factors. Structured data supported the construction of crash severity models and geospatial analyses. Law enforcement provided additional details on crash causation in free-text narratives. Crash narratives contained sufficient text to support viable machine learning models for farm vehicle or equipment crashes, but not for logging truck narratives. Discussion: Crash records can help to fill research and surveillance gaps in AgFF in the southwest region. This supports traffic safety’s evolution to the current Safe System paradigm. There is a conceptual linkage between the Safe System and Total Worker Health approaches, providing a bridge between traffic safety and occupational health. Practical Applications: Despite limitations, crash records can be an important component of injury surveillance for events involving AgFF vehicles. They also can be used to inform the selection and evaluation of traffic countermeasures and behavioral interventions.

Analyzing free-text clinical narratives for veterans with lymphoid malignancies using natural language processing (NLP).

e13576 Background: Free-text clinical narratives contain rich patient information, which is labor-intensive to extract through chart review. We developed an NLP pipeline to enable automatic extraction of performance status (PS), staging, and diagnosis from clinical narratives from Veterans Affairs (VA) patients with lymphoid malignancies (LM). Methods: The rule-based NLP algorithm was developed and iteratively refined using a development corpus of 287 notes independently annotated by two clinicians. The F1-score for PS was 95.8 (precision 98.6, recall 93.2), 92.7 for staging (precision 94.0, recall 81.6), and 67 (precision 80.2, recall 57.9) for diagnosis. The NLP pipeline was then externally validated using an evaluation corpus of 97 notes from another group of 100 veterans with T-cell LM. Results: The results are reported. In the 97 notes, primary diagnosis was most routinely documented, with 2.76 mentions per note. In comparison, staging was most sparsely documented with only 34 mentions (note that 11 patients with large granular lymphocytic leukemia were not staged). The NLP pipeline performed relatively well in extracting PS and staging (F1-scores were 0.74 and 0.72, respectively). It also achieved high precision in extracting diagnosis information (precision 0.93). However, recall (0.44) for diagnosis was poor, likely due to the complexities and inconsistencies of how diagnoses are documented for LM. Frequency of documentation and performance in the external validation set. Conclusions: The pipeline shows promising performance on the external validation set, demonstrating the feasibility of using NLP to extract information from notes of patients with LM for clinical research. The NLP pipeline generally has lower recall than precision, indicating that the pipeline may miss clinical information that should be captured. FPs incorrectly capture entities that are easily confused with the clinical entities of interest, such as nutritional status versus performance status. Future work includes capturing more lexical variations and indicators of documentation, as well as contextual information, such as in which sections of notes elements are likely documented. In addition, we describe how diagnosis may exist as primary, secondary, and in a differential, and we are building an NLP-based classifier to distinguish between these types of diagnoses. We will use results from the rule-based NLP pipeline as labels to fine-tune transformer-based, weak-supervised models to further enhance the performance. [Table: see text]

PetSavers research

Heather Davies from the University of Liverpool received PetSavers funding for the student research project Exploring recording of adverse drug reaction reporting in veterinary free-text clinical narratives. Student Emily Blackwell, whose presentation of her work at BSAVA Congress 2022 won her the PetSavers Clinical Abstract Award, describes the study.

Assessing longitudinal housing status using Electronic Health Record data: a comparison of natural language processing, structured data, and patient-reported history.

Measuring long-term housing outcomes is important for evaluating the impacts of services for individuals with homeless experience. However, assessing long-term housing status using traditional methods is challenging. The Veterans Affairs (VA) Electronic Health Record (EHR) provides detailed data for a large population of patients with homeless experiences and contains several indicators of housing instability, including structured data elements (e.g., diagnosis codes) and free-text clinical narratives. However, the validity of each of these data elements for measuring housing stability over time is not well-studied. We compared VA EHR indicators of housing instability, including information extracted from clinical notes using natural language processing (NLP), with patient-reported housing outcomes in a cohort of homeless-experienced Veterans. NLP achieved higher sensitivity and specificity than standard diagnosis codes for detecting episodes of unstable housing. Other structured data elements in the VA EHR showed promising performance, particularly when combined with NLP. Evaluation efforts and research studies assessing longitudinal housing outcomes should incorporate multiple data sources of documentation to achieve optimal performance.

A natural language processing approach to categorise contributing factors from patient safety event reports

ObjectivesThe objective of this study was to explore the use of natural language processing (NLP) algorithm to categorise contributing factors from patient safety event (PSE). Contributing factors are elements in...

The apparent burden of unexplained sudden infant deaths in Lusaka, Zambia: Findings from analysis of verbal autopsies

Background: The contribution of sudden unexpected infant death (SUID) has received little attention in global health. The objective of this study was to estimate the burden of SUID in Lusaka, Zambia. Methods: Verbal autopsies were conducted on infants who died in Lusaka, between 2017 and 2020. From these, we performed a qualitative analysis of the free text narratives of the final series of events leading to each infant’s death and classified these as symptomatic deaths or SUID. Any narrative that described an infant who was otherwise healthy with no antecedent illness prior to death and found dead in bed after a sleep episode was classified as SUID. We used logistic regression to test for statistical differences between asymptomatic deaths and SUIDs on key infant, maternal and other risk factors of SUIDs. Results: Eight hundred and nine verbal autopsies were conducted with families of decedent infants younger than six months of age. A total of 92.6% (749/809) had presented with symptoms prior to death, whereas 7.4% (60/809) died without preceding symptoms or obvious cause of death. Of these, 16/60 were compatible with accidental suffocation deaths, and 54/60 appeared to be sudden infant death syndrome. SUID deaths were concentrated in infants younger than two months of age with peak age of one to two months. Age at death was the only significant factor in multivariate analysis. Infants aged between one and two months had 2.84 increased odds of suspected SUIDs compared to infants in the first month of life (aOR = 2.84, 95% CI: 1.31, 6.16). Conclusions: Our findings suggest SUID could be accounting for a significant proportion of infant deaths in Zambia, but this cause of infant mortality is going unrecognized. Public health interventions in Zambia, and Africa more broadly, are likely overlooking SUIDs as an important cause of infant mortality.

The Use and Structure of Emergency Nurses’ Triage Narrative Data: Scoping Review

Emergency departments use triage to ensure that patients with the highest level of acuity receive care quickly and safely. Triage is typically a nursing process that is documented as structured and unstructured (free text) data. Free-text triage narratives have been studied for specific conditions but never reviewed in a comprehensive manner. The objective of this paper was to identify and map the academic literature that examines triage narratives. The paper described the types of research conducted, identified gaps in the research, and determined where additional review may be warranted. We conducted a scoping review of unstructured triage narratives. We mapped the literature, described the use of triage narrative data, examined the information available on the form and structure of narratives, highlighted similarities among publications, and identified opportunities for future research. We screened 18,074 studies published between 1990 and 2022 in CINAHL, MEDLINE, Embase, Cochrane, and ProQuest Central. We identified 0.53% (96/18,074) of studies that directly examined the use of triage nurses' narratives. More than 12 million visits were made to 2438 emergency departments included in the review. In total, 82% (79/96) of these studies were conducted in the United States (43/96, 45%), Australia (31/96, 32%), or Canada (5/96, 5%). Triage narratives were used for research and case identification, as input variables for predictive modeling, and for quality improvement. Overall, 31% (30/96) of the studies offered a description of the triage narrative, including a list of the keywords used (27/96, 28%) or more fulsome descriptions (such as word counts, character counts, abbreviation, etc; 7/96, 7%). We found limited use of reporting guidelines (8/96, 8%). The breadth of the identified studies suggests that there is widespread routine collection and research use of triage narrative data. Despite the use of triage narratives as a source of data in studies, the narratives and nurses who generate them are poorly described in the literature, and data reporting is inconsistent. Additional research is needed to describe the structure of triage narratives, determine the best use of triage narratives, and improve the consistent use of triage-specific data reporting guidelines. RR2-10.1136/bmjopen-2021-055132.

Assessing the healthcare quality issues for digital incident reporting in Sweden: Incident reports analysis.

This study explored healthcare quality issues affecting the reporting and investigation levels of digital incident reporting systems. A total of 38 health information technology-related incident reports (free-text narratives) were collected from one of Sweden's national incident reporting repositories. The incidents were analysed using an existing framework, i.e., the Health Information Technology Classification System, to identify the types of issues and consequences. The framework was applied in two fields, 'event description' by the reporters and 'manufacturer's measures', to assess the quality of reporting incidents by the reporters. Additionally, the contributing factors, i.e., either human or technical factors for both fields, were identified to evaluate the quality of the reported incidents. Five types of issues were identified and changes made between before-and-after investigations: Machine to software-related issues (n = 8), machine to use-related issues (n = 5), software to software-related issues (n = 5), use to software-related issues (n = 4) and use to use-related issues (n = 1). Over two-thirds (n = 15) of the incidents demonstrated a change in the contributing factors after the investigation. Only four incidents were identified as altering the consequences after the investigation. This study shed some light on the issues of incident reporting and the gap between the reporting and investigation levels. Facilitating sufficient staff training sessions, agreeing on common terms for health information technology systems, refining the existing classifications systems, enforcing mini-root cause analysis, and ensuring unit-based local reporting and standard national reporting may help bridge the gap between reporting and investigation levels in digital incident reporting.

Transforming clinical trials: the emerging roles of large language models.

Clinical trials are essential for medical research, but they often face challenges in matching patients to trials and planning. Large language models (LLMs) offer a promising solution, signaling a transformative shift in the field of clinical trials. This review explores the multifaceted applications of LLMs within clinical trials, focusing on five main areas expected to be implemented in the near future: enhancing patient-trial matching, streamlining clinical trial planning, analyzing free text narratives for coding and classification, assisting in technical writing tasks, and providing cognizant consent via LLM-powered chatbots. While the application of LLMs is promising, it poses challenges such as accuracy validation and legal concerns. The convergence of LLMs with clinical trials has the potential to revolutionize the efficiency of clinical trials, paving the way for innovative methodologies and enhancing patient engagement. However, this development requires careful consideration and investment to overcome potential hurdles.

Barriers and Facilitators of Obtaining Social Determinants of Health of Patients With Cancer Through the Electronic Health Record Using Natural Language Processing Technology: Qualitative Feasibility Study With Stakeholder Interviews.

Social determinants of health (SDoH), such as geographic neighborhoods, access to health care, education, and social structure, are important factors affecting people's health and health outcomes. The SDoH of patients are scarcely documented in a discrete format in electronic health records (EHRs) but are often available in free-text clinical narratives such as physician notes. Innovative methods like natural language processing (NLP) are being developed to identify and extract SDoH from EHRs, but it is imperative that the input of key stakeholders is included as NLP systems are designed. This study aims to understand the feasibility, challenges, and benefits of developing an NLP system to uncover SDoH from clinical narratives by conducting interviews with key stakeholders: (1) oncologists, (2) data analysts, (3) citizen scientists, and (4) patient navigators. Individuals who frequently work with SDoH data were invited to participate in semistructured interviews. All interviews were recorded and subsequently transcribed. After coding transcripts and developing a codebook, the constant comparative method was used to generate themes. A total of 16 participants were interviewed (5 data analysts, 4 patient navigators, 4 physicians, and 3 citizen scientists). Three main themes emerged, accompanied by subthemes. The first theme, importance and approaches to obtaining SDoH, describes how every participant (n=16, 100%) regarded SDoH as important. In particular, proximity to the hospital and income levels were frequently relied upon. Communication about SDoH typically occurs during the initial conversation with the oncologist, but more personal information is often acquired by patient navigators. The second theme, SDoH exists in numerous forms, exemplified how SDoH arises during informal communication and can be difficult to enter into the EHR. The final theme, incorporating SDoH into health services research, addresses how more informed SDoH can be collected. One strategy is to empower patients so they are aware about the importance of SDoH, as well as employing NLP techniques to make narrative data available in a discrete format, which can provide oncologists with actionable data summaries. Extracting SDoH from EHRs was considered valuable and necessary, but obstacles such as narrative data format can make the process difficult. NLP can be a potential solution, but as the technology is developed, it is important to consider how key stakeholders document SDoH, apply the NLP systems, and use the extracted SDoH in health outcome studies.