Project description:ObjectiveTo identify stigmatizing language in obstetric clinical notes using natural language processing (NLP).Materials and methodsWe analyzed electronic health records from birth admissions in the Northeast United States in 2017. We annotated 1771 clinical notes to generate the initial gold standard dataset. Annotators labeled for exemplars of 5 stigmatizing and 1 positive/preferred language categories. We used a semantic similarity-based search approach to expand the initial dataset by adding additional exemplars, composing an enhanced dataset. We employed traditional classifiers (Support Vector Machine, Decision Trees, and Random Forest) and a transformer-based model, ClinicalBERT (Bidirectional Encoder Representations from Transformers) and BERT base. Models were trained and validated on initial and enhanced datasets and were tested on enhanced testing dataset.ResultsIn the initial dataset, we annotated 963 exemplars as stigmatizing or positive/preferred. The most frequently identified category was marginalized language/identities (n = 397, 41%), and the least frequent was questioning patient credibility (n = 51, 5%). After employing a semantic similarity-based search approach, 502 additional exemplars were added, increasing the number of low-frequency categories. All NLP models also showed improved performance, with Decision Trees demonstrating the greatest improvement (21%). ClinicalBERT outperformed other models, with the highest average F1-score of 0.78.DiscussionClinical BERT seems to most effectively capture the nuanced and context-dependent stigmatizing language found in obstetric clinical notes, demonstrating its potential clinical applications for real-time monitoring and alerts to prevent usages of stigmatizing language use and reduce healthcare bias. Future research should explore stigmatizing language in diverse geographic locations and clinical settings to further contribute to high-quality and equitable perinatal care.ConclusionClinicalBERT effectively captures the nuanced stigmatizing language in obstetric clinical notes. Our semantic similarity-based search approach to rapidly extract additional exemplars enhanced the performances while reducing the need for labor-intensive annotation.

Project description:BackgroundClinical management guidelines (CMGs) are decision support tools for patient care used by professionals, patients, and family caregivers. Since clinical experts develop numerous CMGs, their technical language hinders comprehension and access by nonmedical stakeholders. Additionally, the views of affected individuals and their families are often not incorporated into treatment guidelines. We developed an adequate methodology for addressing the needs and preferences of family and professional stakeholders regarding CMGs, a recently developed protocol for managing congenital disorders of glycosylation (CDG), a family of rare metabolic diseases. We used the CDG community and phosphomannomutase 2 (PMM2)-CDG CMGs as a pilot to test and implement our methodology.ResultsWe listened to 89 PMM2-CDG families and 35 professional stakeholders and quantified their CMG-related needs and preferences through an electronic questionnaire. Most families and professionals rated CMGs as relevant (86.5% and 94.3%, respectively), and valuable (84.3% and 94.3%, respectively) in CDG management. The most identified challenges were the lack of CMG awareness (50.6% of families) and the lack of plain language CMG (39.3% of professionals). Concordantly, among families, the most suggested solution was involving them in CMG development (55.1%), while professionals proposed adapting CMGs to include plain language (71.4%). Based on these results, a participatory framework built upon health literacy principles was created to improve CMG comprehension and accessibility. The outputs are six complementary CMG-related resources differentially adapted to the CDG community's needs and preferences, with a plain language PMM2-CDG CMG as the primary outcome. Additionally, the participants established a distribution plan to ensure wider access to all resources.ConclusionsThis empowering, people-centric methodology accelerates CMG development and accessibility to all stakeholders, ultimately improving the quality of life of individuals living with a specific condition and raising the possibility of application to other clinical guidelines.

Project description:BACKGROUND: Significant health disparities exist between limited English proficient and English-proficient patients. Little is known about the impact of language services on chronic disease outcomes such as for diabetes. METHODS/PRINCIPAL FINDINGS: To determine whether the amount and type of language services received during primary care visits had an impact on diabetes-related outcomes (hospitalization, emergency room utilization, glycemic control) in limited English proficient patients, a retrospective cohort design was utilized. Hospital and medical record data was examined for 1425 limited English proficient patients in the Cambridge Health Alliance diabetes registry. We categorized patients receiving usual care into 7 groups based on the amount and combination of language services (language concordant providers, formal interpretation and nothing) received at primary care visits during a 9 month period. Bivariate analyses and multiple logistic regression were used to determine relationships between language service categories and outcomes in the subsequent 6 months. Thirty-one percent of patients (445) had no documentation of interpreter use or seeing a language concordant provider in any visits. Patients who received 100% of their primary care visits with language concordant providers were least likely to have diabetes-related emergency department visits compared to other groups (p<0001) in the following 6 months. Patients with higher numbers of co-morbidities were more likely to receive formal interpretation. CONCLUSIONS/SIGNIFICANCE: Language concordant providers may help reduce health care utilization for limited English proficient patients with diabetes. However, given the lack of such providers in sufficient numbers to meet patients' communication needs, strategies are needed to both increase their numbers and ensure that the highest risk patients receive the most appropriate language services. In addition, systems serving diverse populations must clarify why some limited English proficient patients do not receive language services at some or all of their visits and whether this has an impact on quality of care.

Project description:PurposeBreast masses in pediatric patients are often managed similarly to adult breast masses despite significant differences in pathology and natural history. Emerging evidence suggests that clinical observation is safe. The purpose of this study was to quantify the clinical appropriateness of the management of benign breast disease in pediatric patients.MethodsA multi-institutional retrospective cohort study was completed between 1995 and 2017. Patients were included if they had benign breast disease and were 19 years old or younger. A timeline of all interventions (ultrasound, biopsy, or excision) was generated to quantify the number of patients who were observed for at least 90 days, deemed appropriate care. To quantify inappropriate care, the number of interventions performed within 90 days, and the pathologic concordance to clinical decisions was determined by reviewing the radiology reports of all ultrasounds and pathology reports of all biopsies and excisions.ResultsA total of 1,909 patients were analyzed. Mean age was 16.4 years old (± 2.1). The majority of masses were fibroadenomas (60.4%). Only half of patients (54.3%) were observed for 90 or more days. 81.1% of interventions were unnecessary, with pathology revealing masses that would be safe to observe. The positive predictive value (PPV) of clinical decisions made based on suspicious ultrasound findings was 16.2%, not different than a PPV of 21.9% (p < 0.25) for decisions made on clinical suspicion alone.ConclusionDespite literature supporting an observation period for pediatric breast masses, half of patients had an intervention within three months with one out of ten patients undergoing an invasive procedure within this time frame. Furthermore, 81.1% of invasive interventions were unnecessary based on final pathologic findings. A formal consensus clinical guideline for the management of pediatric benign breast disease including a standardized clinical observation period is needed to decrease unnecessary procedures in pediatric patients with breast masses.

Project description:BACKGROUND:Bronchiolitis is the commonest respiratory infection in children less than 12 months and cause of hospitalisation in infants under 6 months of age in Australasia. Unfortunately there is substantial variation in management, despite high levels of supporting evidence. This paper reports on the process, strengths and challenges of the hybrid approach used to develop the first Australasian management guideline relevant to the local population. METHOD:An adaption of the nine steps recommended by the National Health and Medical Research Council (NHMRC) and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology were utilised. Following establishment of the Guideline Development Committee (GDC), we identified the population, intervention, comparator, outcomes and time of interest (PICOt) questions, undertook a systematic literature search and graded the evidence and recommendations using the NHMRC and GRADE processes. Using Nominal Group Techniques (NGT), consensus was sought in formulating the clinical practice recommendations and practice points. Key health professional bodies were consulted to ensure relevance in the Australasian emergency and ward settings. RESULTS:From 33 PICOT questions, clinical recommendations for practice that were deemed relevant to the Australasian population were identified. Specific considerations for the management of Australian and New Zealand indigenous infants in relation to the use of azithromycin and risk factors for more serious illness are included. Using NGT, consensus demonstrated by a median Likert score > 8 for all recommendations was achieved. The guideline presents clinical guidance, followed by the key recommendations and evidence review behind each recommendation. CONCLUSION:Developing evidence-based clinical guidelines is a complex process with considerable challenges. Challenges included having committee members located over two countries and five time zones, large volume of literature and variation of member's knowledge of grading of evidence and recommendations. The GRADE and NHMRC processes provided a systematic and transparent approach ensuring a final structure including bedside interface, and a descriptive summary of the evidence base and tables for each key statement. Involvement of stakeholders who will ultimately be end-users as members of the GDC provided valuable knowledge. Lessons learnt during this guideline development process provide valuable insight for those planning development of evidence-based guidelines.

Dataset Information

Impact of a Bronchiolitis Clinical Pathway on Management Decisions by Preferred Language

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets