Protocol for the Evaluation of Data-Related Processes and Challenges at Level 1-4 Trauma Centers in Arkansas: A Mixed-Methods Case Study.

Background: In the following manuscript, we describe the detailed protocol for a mixed-methods, observational case study conducted to identify and evaluate existing data-related processes and challenges currently faced by trauma centers in a rural state. The data will be utilized to assess the impact of these challenges on registry data collection. Methods: The study relies on a series of interviews and observations to collect data from trauma registry staff at level 1-4 trauma centers across the state of Arkansas. A think-aloud protocol will be used to facilitate observations to gather keystroke-level modeling data and insight into site processes and workflows for collecting and submitting data to the Arkansas Trauma Registry. Informal, semi-structured interviews will follow the observation period to assess the participant's perspective on current processes, potential barriers to data collection or submission to the registry, and recommendations for improvement. Each session will be recorded, and de-identified transcripts and session notes will be used for analysis. Keystroke level modeling data derived from observations will be extracted and analyzed quantitatively to determine time spent performing end-to-end registry-related activities. Qualitative data from interviews will be reviewed and coded by 2 independent reviewers following a thematic analysis methodology. Each set of codes will then be adjudicated by the reviewers using a consensus-driven approach to extrapolate the final set of themes. Discussion: We will utilize a mixed methods approach to understand existing processes and barriers to data collection for the Arkansas Trauma Registry. Anticipated results will provide a baseline measure of the data collection and submission processes at various trauma centers across the state. We aim to assess strengths and limitations of existing processes and identify existing barriers to interoperability. These results will provide first-hand knowledge on existing practices for the trauma registry use case and will provide quantifiable data that can be utilized in future research to measure outcomes of future process improvement efforts. The potential implications of this study can form the basis for identifying potential solutions for streamlining data collection, exchange, and utilization of trauma registry data for clinical practice, public health, and clinical and translational research.

Measuring the Coverage of the HL7® FHIR® Standard in Supporting Data Acquisition for 3 Public Health Registries.

With the increasing need for timely submission of data to state and national public health registries, current manual approaches to data acquisition and submission are insufficient. In clinical practice, federal regulations are now mandating the use of data messaging standards, i.e., the Health Level Seven (HL7®) Fast Healthcare Interoperability Resources (FHIR®) standard, to facilitate the electronic exchange of clinical (patient) data. In both research and public health practice, we can also leverage FHIR® ‒ and the infrastructure already in place for supporting exchange of clinical practice data ‒ to enable seamless exchange between the electronic medical record and public health registries. That said, in order to understand the current utility of FHIR® for supporting the public health use case, we must first measure the extent to which the standard resources map to the required registry data elements. Thus, using a systematic mapping approach, we evaluated the level of completeness of the FHIR® standard to support data collection for three public health registries (Trauma, Stroke, and National Surgical Quality Improvement Program). On average, approximately 80% of data elements were available in FHIR® (71%, 77%, and 92%, respectively; inter-annotator agreement rates: 82%, 78%, and 72%, respectively). This tells us that there is the potential for significant automation to support EHR-to-Registry data exchange, which will reduce the amount of manual, error-prone processes and ensure higher data quality. Further, identification of the remaining 20% of data elements that are "not mapped" will enable us to improve the standard and develop profiles that will better fit the registry data model.

Perceptions of tech-based mental health screening.

Background: An estimated one-third of patients experience post-traumatic stress disorder (PTSD) or depression in the year following a traumatic injury. The American College of Surgeons requires postinjury PTSD and depression screening in trauma centers, although implementation has been limited. Tech-based solutions have been proposed to improve uptake of postinjury mental health screening. The goals of this pilot study were to assess the usability and acceptability of Blueprint, a tech-based mental health screening platform, and explore attitudes toward tech-based screening and intervention. Methods: This pilot study included trauma patients (n=10) admitted to the trauma service. Participants completed the PTSD Checklist-5 and Patient Health Questionnaire-9 using Blueprint to test usability and acceptability of the platform. Participants completed the System Usability Scale (SUS) and a semi-structured interview to assess several domains including attitudes toward tech-based screening, potential barriers to implementation, and its usefulness in a postinjury context. Summative Template Analysis, a data abstraction procedure, was used to analyze qualitative data. Results: Blueprint received an average SUS score of 93.25/100 suggesting participants found the interface to be an 'excellent' means to assess postinjury mental health concerns. Participants were supportive of universal screening and identified several benefits to engaging in tech-based routine monitoring of postinjury PTSD and depressive symptoms including convenience, personalization, and trauma-informed care. Regarding intervention, patients valued web-based psychoeducation on topics related to their overall care and local resources. Conclusions: Tech-based mental health screening was highly usable and valuable to trauma patients at risk for postinjury PTSD and depression. Participants valued web-based psychoeducation and resources, but overall preferred Blueprint be used to facilitate access to in-person mental health services. Further evaluation of Blueprint as a means of assessment, intervention, and referral is needed.

Vitamin D Oral Replacement in Children With Obesity Related Asthma: VDORA1 Randomized Clinical Trial.

Children with asthma and obesity are more likely to have lower vitamin D levels, but the optimal replacement dose is unknown in this population. The objective of this study is identifying a vitamin D dose in children with obesity-related asthma that safely achieves serum vitamin D levels of ≥ 40 ng/mL. This prospective multisite randomized controlled trial recruited children/adolescents with asthma and body mass index ≥ 85% for age/sex. Part 1 (dose finding), evaluated 4 oral vitamin D regimens for 16 weeks to identify a replacement dose that achieved serum vitamin D levels ≥ 40 ng/mL. Part 2 compared the replacement dose calculated from part 1 (50,000 IU loading dose with 8,000 IU daily) to standard of care (SOC) for 16 weeks to identify the proportion of children achieving target serum 25(OH)D level. Part 1 included 48 randomized participants. Part 2 included 64 participants. In Part 1, no SOC participants achieved target serum level, but 50-72.7% of participants in cohorts A-C achieved the target serum level. In part 2, 78.6% of replacement dose participants achieved target serum level compared with none in the SOC arm. No related serious adverse events were reported. This trial confirmed a 50,000 IU loading dose plus 8,000 IU daily oral vitamin D as safe and effective in increasing serum 25(OH)D levels in children/adolescents with overweight/obesity to levels ≥ 40 ng/mL. Given the critical role of vitamin D in many conditions complicating childhood obesity, these data close a critical gap in our understanding of vitamin D dosing in children.

Protocol for the Evaluation of Data-Related Processes and Challenges at Level 1-4 Trauma Centers in Arkansas: A Mixed-Methods Case Study.

Background: In the following manuscript, we describe the detailed protocol for a mixed-methods, observational case study conducted to identify and evaluate existing data-related processes and challenges currently faced by trauma centers in a rural state. The data will be utilized to assess the impact of these challenges on registry data collection. Methods: The study relies on a series of interviews and observations to collect data from trauma registry staff at level 1-4 trauma centers across the state of Arkansas. A think-aloud protocol will be used to facilitate observations as a means to gather keystroke-level modeling data and insight into site processes and workflows for collecting and submitting data to the Arkansas Trauma Registry. Informal, semi-structured interviews will follow the observation period to assess the participant's perspective on current processes, potential barriers to data collection or submission to the registry, and recommendations for improvement. Each session will be recorded and de-identified transcripts and session notes will be used for analysis. Keystroke level modeling data derived from observations will be extracted and analyzed quantitatively to determine time spent performing end-to-end registry-related activities. Qualitative data from interviews will be reviewed and coded by 2 independent reviewers following a thematic analysis methodology. Each set of codes will then be adjudicated by the reviewers using a consensus-driven approach to extrapolate the final set of themes. Discussion: We will utilize a mixed methods approach to understand existing processes and barriers to data collection for the Arkansas Trauma Registry. Anticipated results will provide a baseline measure of the data collection and submission processes at various trauma centers across the state. We aim to assess strengths and limitations of existing processes and identify existing barriers to interoperability. These results will provide first-hand knowledge on existing practices for the trauma registry use case and will provide quantifiable data that can be utilized in future research to measure outcomes of future process improvement efforts. The potential implications of this study can form the basis for identifying potential solutions for streamlining data collection, exchange, and utilization of trauma registry data for clinical practice, public health, and clinical and translational research.

COVID-19 Vaccine Hesitancy Among Parents: A Qualitative Study.

OBJECTIVES: Addressing parental/caregivers' coronavirus disease 2019 (COVID-19) vaccine hesitancy is critical to improving vaccine uptake in children. Common concerns have been previously reported through online surveys, but qualitative data from KII and focus groups may add much-needed context. Our objective was to examine factors impacting pediatric COVID-19 vaccine decision-making in Black, Spanish-speaking, and rural white parents/caregivers to inform the content design of a mobile application to improve pediatric COVID-19 vaccine uptake. METHODS: Parents/caregivers of children aged 2 to 17 years from groups disproportionately affected by COVID-19-related vaccine hesitancy (rural-dwelling persons of any race/ethnicity, urban Black persons, and Spanish-speaking persons) were included on the basis of their self-reported vaccine hesitancy and stratified by race/ethnicity. Those expressing vaccine acceptance or refusal participated in KII, and those expressing hesitancy in focus groups. Deidentified transcripts underwent discourse analysis and thematic analysis, both individually and as a collection. Themes were revised until coders reached consensus. RESULTS: Overall, 36 participants completed the study: 4 vaccine acceptors and 4 refusers via KIIs, and the remaining 28 participated in focus groups. Participants from all focus groups expressed that they would listen to their doctor for information about COVID-19 vaccines. Infertility was a common concern, along with general concerns about vaccines. Vaccine decision-making was informed by the amount of information available to parents/caregivers, including scientific research; possible positive and negative long-term effects; and potential impacts of vaccination on preexisting medical conditions. CONCLUSIONS: Parents/caregivers report numerous addressable vaccine concerns. Our results will inform specific, targeted interventions for improving COVID-19 vaccine confidence.

Social Determinants of Health Data Quality at Different Levels of Geographic Detail.

Social determinants of health (SDOH) impact 80% of health outcomes from acute to chronic disorders, and attempts are underway to provide these data elements to clinicians. It is, however, difficult to collect SDOH data through (1) surveys, which provide inconsistent and incomplete data, or (2) aggregates at the neighborhood level. Data from these sources is not sufficiently accurate, complete, and up-to-date. To demonstrate this, we have compared the Area Deprivation Index (ADI) to purchased commercial consumer data at the individual-household level. The ADI is composed of income, education, employment, and housing quality information. Although this index does a good job of representing populations, it is not adequate to describe individuals, especially in a healthcare context. Aggregate measures are, by definition, not sufficiently granular to describe each individual within the population they represent and may result in biased or imprecise data when simply assigned to the individual. Moreover, this problem is generalizable to any community-level element, not just ADI, in so far as they are an aggregate of the individual community members.

Informatics infrastructure in a rural pediatric clinical trials network: Matching specific clinical research needs with best practices and industry guidelines.

Children have historically been underrepresented in randomized controlled trials and multi-center studies. This is particularly true for children who reside in rural and underserved areas. Conducting multi-center trials in rural areas presents unique informatics challenges. These challenges call for increased attention towards informatics infrastructure and the need for development and application of sound informatics approaches to the collection, processing, and management of data for clinical studies. By modifying existing local infrastructure and utilizing open source tools, we have been able to successfully deploy a multi-site data coordinating and operations center. We report our implementation decisions for data collection and management for the IDeA States Pediatric Clinical Trial Network (ISPCTN) based on the functionality needed for the ISPCTN, our synthesis of the extant literature in data collection and management methodology, and Good Clinical Data Management Practices.

Repeatable enhancement of healthcare data with social determinants of health.

Background: Social and behavioral aspects of our lives significantly impact our health, yet minimal social determinants of health (SDOH) data elements are collected in the healthcare system. Methods: In this proof-of-concept study we developed a repeatable SDOH enrichment and integration process to incorporate dynamically evolving SDOH domain concepts from consumers into clinical data. This process included SDOH mapping, linking compiled consumer data to patient records in Electronic Health Records, data quality analysis and preprocessing, and storage. Results: Consumer compilers data coverage ranged from ~90 to ~54% and the percentage match rate between compilers was between ~21 and 64%. Our preliminary analysis showed that apart from demographic factors, several SDOH factors like home-ownership, marital-status, presence of children, number of members per household, economic stability and education were significantly different between the COVID-19 positive and negative patient groups while estimated family-income and home market-value were not. Conclusion: Our preliminary analysis shows commercial consumer data can be a viable source of SDOH factor at an individual-level for clinical data thus providing a path for clinicians to improve patient treatment and care.

Preprocessing to Address Bias in Healthcare Data.

Multimorbidity, having a diagnosis of two or more chronic conditions, increases as people age. It is a predictor used in clinical decision-making, but underdiagnosis in underserved populations produces bias in the data that support algorithms used in the healthcare processes. Artificial intelligence (AI) systems could produce inaccurate predictions if patients have multiple unknown conditions. Rural patients are more likely to be underserved and also more likely to have multiple chronic conditions. In this study, data collected during the course of care in a centrally located academic hospital, multimorbidity decreased with rurality. This decrease suggests a bias against rural patients for algorithms that rely on diagnosis information to calculate risk. To test preprocessing to address bias in healthcare data, we measured the amount of discrimination in favor of metropolitan patients in the classification of multimorbidity. We built a model using the biased data to test optimum classification performance. A new unbiased training data set and model were created and tested against unaltered validation data. The new model's classification performance on unaltered data did not diverge significantly from the performance of the initial optimal model trained on the biased data suggesting that bias can be removed with preprocessing.

Covid-19 Positivity Differences Among Patients of a Rural, Southern US State Hospital System Based on Population Density, Rural-Urban Classification, and Area Deprivation Index.

In this study we examined the correlation of COVID-19 positivity with area deprivation index (ADI), social determinants of health (SDOH) factors based on a consumer and electronic medical record (EMR) data and population density in a patient population from a tertiary healthcare system in Arkansas. COVID-19 positivity was significantly associated with population density, age, race, and household size. Understanding health disparities and SDOH data can add value to health and the creation of trustable AI.

The h-ANN Model: Comprehensive Colonoscopy Concept Compilation Using Combined Contextual Embeddings.

Colonoscopy is a screening and diagnostic procedure for detection of colorectal carcinomas with specific quality metrics that monitor and improve adenoma detection rates. These quality metrics are stored in disparate documents i.e., colonoscopy, pathology, and radiology reports. The lack of integrated standardized documentation is impeding colorectal cancer research. Clinical concept extraction using Natural Language Processing (NLP) and Machine Learning (ML) techniques is an alternative to manual data abstraction. Contextual word embedding models such as BERT (Bidirectional Encoder Representations from Transformers) and FLAIR have enhanced performance of NLP tasks. Combining multiple clinically-trained embeddings can improve word representations and boost the performance of the clinical NLP systems. The objective of this study is to extract comprehensive clinical concepts from the consolidated colonoscopy documents using concatenated clinical embeddings. We built high-quality annotated corpora for three report types. BERT and FLAIR embeddings were trained on unlabeled colonoscopy related documents. We built a hybrid Artificial Neural Network (h-ANN) to concatenate and fine-tune BERT and FLAIR embeddings. To extract concepts of interest from three report types, 3 models were initialized from the h-ANN and fine-tuned using the annotated corpora. The models achieved best F1-scores of 91.76%, 92.25%, and 88.55% for colonoscopy, pathology, and radiology reports respectively.

DeIDNER Model: A Neural Network Named Entity Recognition Model for Use in the De-identification of Clinical Notes.

Clinical named entity recognition (NER) is an essential building block for many downstream natural language processing (NLP) applications such as information extraction and de-identification. Recently, deep learning (DL) methods that utilize word embeddings have become popular in clinical NLP tasks. However, there has been little work on evaluating and combining the word embeddings trained from different domains. The goal of this study is to improve the performance of NER in clinical discharge summaries by developing a DL model that combines different embeddings and investigate the combination of standard and contextual embeddings from the general and clinical domains. We developed: 1) A human-annotated high-quality internal corpus with discharge summaries and 2) A NER model with an input embedding layer that combines different embeddings: standard word embeddings, context-based word embeddings, a character-level word embedding using a convolutional neural network (CNN), and an external knowledge sources along with word features as one-hot vectors. Embedding was followed by bidirectional long short-term memory (Bi-LSTM) and conditional random field (CRF) layers. The proposed model reaches or overcomes state-of-the-art performance on two publicly available data sets and an F1 score of 94.31% on an internal corpus. After incorporating mixed-domain clinically pre-trained contextual embeddings, the F1 score further improved to 95.36% on the internal corpus. This study demonstrated an efficient way of combining different embeddings that will improve the recognition performance aiding the downstream de-identification of clinical notes.

TAX-Corpus: Taxonomy based Annotations for Colonoscopy Evaluation.

Colonoscopy plays a critical role in screening of colorectal carcinomas (CC). Unfortunately, the data related to this procedure are stored in disparate documents, colonoscopy, pathology, and radiology reports respectively. The lack of integrated standardized documentation is impeding accurate reporting of quality metrics and clinical and translational research. Natural language processing (NLP) has been used as an alternative to manual data abstraction. Performance of Machine Learning (ML) based NLP solutions is heavily dependent on the accuracy of annotated corpora. Availability of large volume annotated corpora is limited due to data privacy laws and the cost and effort required. In addition, the manual annotation process is error-prone, making the lack of quality annotated corpora the largest bottleneck in deploying ML solutions. The objective of this study is to identify clinical entities critical to colonoscopy quality, and build a high-quality annotated corpus using domain specific taxonomies following standardized annotation guidelines. The annotated corpus can be used to train ML models for a variety of downstream tasks.

Machine learning can identify patients at risk of hyperparathyroidism without known calcium and intact parathyroid hormone.

BACKGROUND: To prove the concept of diagnosing primary hyperparathyroidism (pHPT) without calcium and parathyroid hormone (PTH) values and identifying potential risk factors for pHPT. METHODS: Data were extracted from the clinical data warehouse (CDW) at the University of Arkansas for Medical Sciences (UAMS) Epic EHR (2014-2019). RESULTS: 1737 patients with over 185 000 rows of clinical data were provided in a relational structure and processed/flattened to facilitate modeling. Phenotype elements were identified for pHPT without advance knowledge of calcium and PTH levels. The area under the curve (AUC) for the prediction of pHPT using our model was 0.86 with sensitivity and specificity of 0.8953 and 0.6686, respectively, using a 0.45 probability threshold. CONCLUSION: Primary hyperparathyroidism was predicted from a dataset excluding calcium and PTH data with 86% accuracy. This approach needs to be validated/refined on larger samples of data and plans are in place to do this with other regional/national datasets.

Deep Learning Methods to Predict Mortality in COVID-19 Patients: A Rapid Scoping Review.

The ongoing COVID-19 pandemic has become the most impactful pandemic of the past century. The SARS-CoV-2 virus has spread rapidly across the globe affecting and straining global health systems. More than 2 million people have died from COVID-19 (as of 30 January 2021). To lessen the pandemic's impact, advanced methods such as Artificial Intelligence models are proposed to predict mortality, morbidity, disease severity, and other outcomes and sequelae. We performed a rapid scoping literature review to identify the deep learning techniques that have been applied to predict hospital mortality in COVID-19 patients. Our review findings provide insights on the important deep learning models, data types, and features that have been reported in the literature. These summary findings will help scientists build reliable and accurate models for better intervention strategies for predicting mortality in current and future pandemic situations.

COVID-19 Is Connected with Lower Health Literacy in Rural Areas.

The relationship between social determinants of health (SDoH) and health outcomes is established and extends to a higher risk of contracting COVID-19. Given the factors included in SDoH, such as education level, race, rurality, and socioeconomic status are interconnected, it is unclear how individual SDoH factors may uniquely impact risk. Lower socioeconomic status often occurs in concert with lower educational attainment, for example. Because literacy provides access to information needed to avoid infection and content can be made more accessible, it is essential to determine to what extent health literacy contributes to successful containment of a pandemic. By incorporating this information into clinical data, we have isolated literacy and geographic location as SDoH factors uniquely related to the risk of COVID-19 infection. For patients with comorbidities linked to higher illness severity, residents of rural areas associated with lower health literacy at the zip code level had a greater likelihood of positive COVID-19 results unrelated to their economic status.

Machine Learning Approach to Optimize Sedation Use in Endoscopic Procedures.

Endoscopy procedures are often performed with either moderate or deep sedation. While deep sedation is costly, procedures with moderate sedation are not always well tolerated resulting in patient discomfort, and are often aborted. Due to lack of clear guidelines, the decision to utilize moderate sedation or anesthesia for a procedure is made by the providers, leading to high variability in clinical practice. The objective of this study was to build a Machine Learning (ML) model that predicts if a colonoscopy can be successfully completed with moderate sedation based on patients' demographics, comorbidities, and prescribed medications. XGBoost model was trained and tested on 10,025 colonoscopies (70% - 30%) performed at University of Arkansas for Medical Sciences (UAMS). XGBoost achieved average area under receiver operating characteristic curve (AUC) of 0.762, F1-score to predict procedures that need moderate sedation was 0.85, and precision and recall were 0.81 and 0.89 respectively. The proposed model can be employed as a decision support tool for physicians to bolster their confidence while choosing between moderate sedation and anesthesia for a colonoscopy procedure.

Consolidated EHR Workflow for Endoscopy Quality Reporting.

Although colonoscopy is the most frequently performed endoscopic procedure, the lack of standardized reporting is impeding clinical and translational research. Inadequacies in data extraction from the raw, unstructured text in electronic health records (EHR) pose an additional challenge to procedure quality metric reporting, as vital details related to the procedure are stored in disparate documents. Currently, there is no EHR workflow that links these documents to the specific colonoscopy procedure, making the process of data extraction error prone. We hypothesize that extracting comprehensive colonoscopy quality metrics from consolidated procedure documents using computational linguistic techniques, and integrating it with discrete EHR data can improve quality of screening and cancer detection rate. As a first step, we developed an algorithm that links colonoscopy, pathology and imaging documents by analyzing the chronology of various orders placed relative to the colonoscopy procedure. The algorithm was installed and validated at the University of Arkansas for Medical Sciences (UAMS). The proposed algorithm in conjunction with Natural Language Processing (NLP) techniques can overcome current limitations of manual data abstraction.

DeIDNER Corpus: Annotation of Clinical Discharge Summary Notes for Named Entity Recognition Using BRAT Tool.

Named Entity Recognition (NER) aims to identify and classify entities into predefined categories is a critical pre-processing task in Natural Language Processing (NLP) pipeline. Readily available off-the-shelf NER algorithms or programs are trained on a general corpus and often need to be retrained when applied on a different domain. The end model's performance depends on the quality of named entities generated by these NER models used in the NLP task. To improve NER model accuracy, researchers build domain-specific corpora for both model training and evaluation. However, in the clinical domain, there is a dearth of training data because of privacy reasons, forcing many studies to use NER models that are trained in the non-clinical domain to generate NER feature-set. Thus, influencing the performance of the downstream NLP tasks like information extraction and de-identification. In this paper, our objective is to create a high quality annotated clinical corpus for training NER models that can be easily generalizable and can be used in a downstream de-identification task to generate named entities feature-set.