Search

Introduction: Emergency departments (ED) play a central role in defining the effectiveness and quality of the overall hospital’s mass casualty incident (MCI) response. The use of electronic health records (EHR) in hospital settings has been rapidly growing globally. There is, however, a paucity of literature on the use and performance of EHR during MCIs.Methods: In this study we aimed to describe EHR use, as well as the challenges and lessons learnt in response to the 2020 explosion in the Port of Beirut, Lebanon, during which the hospital received over 360 casualties.Results: Information technology support, reducing EHR system restrictions, cross-function training, focus on registration and patient identification, patient flow and tracking, mobility and bedside access, and alternate sites of care are all important areas to focus on during emergency/disaster response planning.Conclusion: Innovative solutions that help address logistical challenges for different aspects of the disaster response are needed.

Neighborhood socioeconomic context where Latinx children live may influence body weight status. Los Angeles County and Orange County of Southern California both are on the list of the top ten counties with the largest Latinx population in the USA. This heterogeneity allowed us to estimate differential impacts of neighborhood environment on childrens body mass index z-scores by race/ethnicity using novel methods and a rich data source. We geocoded pediatric electronic medical record data from a predominantly Latinx sample and characterized neighborhoods into unique residential contexts using latent profile modeling techniques. We estimated multilevel linear regression models that adjust for comorbid conditions and found that a childs place of residence independently associates with higher body mass index z-scores. Interactions further reveal that Latinx children living in Middle-Class neighborhoods have higher BMI z-scores than Asian and Other Race children residing in the most disadvantaged communities. Our findings underscore the complex relationship between community racial/ethnic composition and neighborhood socioeconomic context on body weight status during childhood.

Background: Despite substantial progress in maternal and neonatal health, Rwanda's mortality rates remain high, necessitating innovative approaches to meet health related Sustainable Development Goals (SDGs). By leveraging data collected from Electronic Medical Records, this study explores the application of machine learning models to predict adverse pregnancy outcomes, thereby improving risk assessment and enhancing care delivery. Methods: This study utilized retrospective cohort data from the electronic medical record (EMR) system of 25 hospitals in Rwanda from 2020 to 2023. The independent variables included socioeconomic status, health status, reproductive health, and pregnancy-related factors. The outcome variable was a binary composite feature that combined adverse pregnancy outcomes in both the mother and the newborn. Extensive data cleaning was performed, with missing values addressed through various strategies, including the exclusion of variables and instances, imputation techniques using K-Nearest Neighbors and Multiple Imputation by Chained Equations. Data imbalance was managed using a synthetic minority oversampling technique. Six machine learning models—Logistic Regression, Decision Trees, Support Vector Machine, Gradient Boosting, Random Forest, and Multilayer Perceptron—were trained using 10-fold cross-validation and evaluated on an unseen dataset with–70 − 30 training and evaluation splits. Results: Data from 117,069 women across 25 hospitals in Rwanda were analyzed, leading to a final dataset of 32,783 women after removing entries with significant missing values. Among these women, 5,424 (16.5%) experienced adverse pregnancy outcomes. Random Forest and Gradient Boosting Classifiers demonstrated high accuracy and precision. After hyperparameter tuning, the Random Forest model achieved an accuracy of 90.6% and an ROC-AUC score of 0.85, underscoring its effectiveness in predicting adverse outcomes. However, a recall rate of 46.5% suggests challenges in detecting all the adverse cases. Key predictors of adverse outcomes identified in this study included gestational age, number of pregnancies, antenatal care visits, maternal age, vital signs, and delivery methods. Conclusions: This study recommends enhancing EMR data quality, integrating machine learning into routine practice, and conducting further research to refine predictive models and address evolving pregnancy outcomes. In addition, this study recommends the design of AI-based interventions for high-risk pregnancies. Clinical trial number: Not applicable. [ABSTRACT FROM AUTHOR]

Speech and language disorders are known to have a substantial genetic contribution. Although frequently examined as components of other conditions, research on the genetic basis of linguistic differences as separate phenotypic subgroups has been limited so far. Here, we performed an in-depth characterization of speech and language disorders in 52 143 individuals, reconstructing clinical histories using a large-scale data-mining approach of the electronic medical records from an entire large paediatric healthcare network. The reported frequency of these disorders was the highest between 2 and 5 years old and spanned a spectrum of 26 broad speech and language diagnoses. We used natural language processing to assess the degree to which clinical diagnoses in full-text notes were reflected in ICD-10 diagnosis codes. We found that aphasia and speech apraxia could be retrieved easily through ICD-10 diagnosis codes, whereas stuttering as a speech phenotype was coded in only 12% of individuals through appropriate ICD-10 codes. We found significant comorbidity of speech and language disorders in neurodevelopmental conditions (30.31%) and, to a lesser degree, with epilepsies (6.07%) and movement disorders (2.05%). The most common genetic disorders retrievable in our analysis of electronic medical records were STXBP1 (n = 21), PTEN (n = 20) and CACNA1A (n = 18). When assessing associations of genetic diagnoses with specific linguistic phenotypes, we observed associations of STXBP1 and aphasia (P = 8.57 × 10−7, 95% confidence interval = 18.62–130.39) and MYO7A with speech and language development delay attributable to hearing loss (P = 1.24 × 10−5, 95% confidence interval = 17.46–infinity). Finally, in a sub-cohort of 726 individuals with whole-exome sequencing data, we identified an enrichment of rare variants in neuronal receptor pathways, in addition to associations of UQCRC1 and KIF17 with expressive aphasia, MROH8 and BCHE with poor speech, and USP37 , SLC22A9 and UMODL1 with aphasia. In summary, our study outlines the landscape of paediatric speech and language disorders, confirming the phenotypic complexity of linguistic traits and novel genotype–phenotype associations. Subgroups of paediatric speech and language disorders differ significantly with respect to the composition of monogenic aetiologies. [ABSTRACT FROM AUTHOR]

Background: Self-management education is integral for proper asthma management. However, there is an accessibility gap to self-management education following asthma hospitalizations. Most pediatric patients and their families receive suboptimal or no education. Objective: To implement a comprehensive pediatric asthma education program and evaluate subsequent self-management knowledge in patients as well as behavior change outcomes reflected in the frequency of asthma related repeat emergency department visits and hospitalization. The program implementation was informed by the Knowledge to Translation Action Framework and the i-PARIHS model for quality improvement and involved several iterative stages. Methods: We implemented a comprehensive asthma education program for the families of all children 0-18 years old who had been admitted for an asthma exacerbation to the Children's Hospital of Eastern Ontario (CHEO), beginning on April 1, 2018. The program was adapted to the stages of the Knowledge Translation to Action Framework including undertaking an environmental scan, expert stakeholder feedback, reviews, addressing barriers, and tailoring the intervention, along with evaluating knowledge and health outcomes. Education was delivered over 1-2 h in personalized individual or small group settings, within 4 wk of hospital discharge. All education was provided by registered nurses or respiratory therapists who were also certified asthma educators. The EPIC electronic medical record was used to facilitate referral and scheduling of asthma education sessions, and to track subsequent acute asthma visits. We compared the frequency of a repeat asthma emergency department (ED) visit or hospitalization within 1-year following an initial asthma hospitalization for children who would have received comprehensive asthma education, to a historical cohort of children who were hospitalized between April 9, 2017 – Apr 8, 2018, and did not receive asthma education. Results: The program had a high enrollment, capturing nearly 75% of the target population. Most families found the program to be acceptable and reported increased knowledge of how to manage asthma. We identified a crude overall 54% reduction in repeat hospitalizations among children 1 year after implementation of the asthma education program (i.e. 10.2% (23/225) repeat hospitalization rate pre- implementation versus 4.8% (11/227) post-implementation). In adjusted time-to event analysis, this reduction was prominent at 3 months among those who received comprehensive asthma education, relative to those who did not, but this improvement was not sustained by 1 year (HR =1.1, 95% CI =0.55- 2.05; p-value = 0.6). Discussion: Although we did not find long-term improvements in ED visits, or hospitalizations, in children of caregivers who participated in comprehensive asthma education, the asthma education program holds potential given that most patients found it to be acceptable and that it increased asthma management knowledge. A future asthma education program should include multiple sessions to ensure that the knowledge and behavior change will be sustained, leading ultimately to long-term reductions in repeat ED visits and hospitalizations. [ABSTRACT FROM AUTHOR]

Objective Event capture in clinical trials is resource-intensive, and electronic medical records (EMRs) offer a potential solution. This study develops algorithms for EMR-based death and hospitalization capture and compares them with traditional event capture methods. Materials and Methods We compared the effectiveness of EMR-based event capture and site-captured events adjudicated by a clinical endpoint committee in the multi-center INfluenza Vaccine to Effectively Stop cardio Thoracic Events and Decompensated heart failure (INVESTED) trial for participants from the Veterans Affairs healthcare system. Varying time windows around event dates were used to optimize events matching. The algorithms were externally validated for heart failure hospitalizations in the Medical Information Mart for Intensive Care (MIMIC)-IV database. Results We observed 100% sensitivity for death events with a 1-day window. Sensitivity for cardiovascular, heart failure, pulmonary, and nonspecific cardiopulmonary hospitalizations using discharge diagnosis codes varied between 75% and 95%. Including Centers for Medicare & Medicaid Services data improved sensitivity with no meaningful decrease in specificity. The MIMIC-IV analysis showed 82% sensitivity and 99% specificity for heart failure hospitalizations. Discussion EMR-based method accurately identifies all-cause mortality and demonstrates high accuracy for cardiopulmonary hospitalizations. This study underscores the importance of optimal time windows, data completeness, and domain variability in EMR systems. Conclusion EMR-based methods are effective strategies for capturing death and hospitalizations in clinical trials; however, their effectiveness may be influenced by the complexity of events and domain variability across different EMR systems. Nonetheless, EMR-based methods can serve as a valuable complement to traditional methods. [ABSTRACT FROM AUTHOR]

Background: Effective interventions to reduce drowning incidents require accurate and reliable data for scientific analysis. However, the lack of high-quality evidence and the variability in drowning terminology, definitions, and outcomes present significant challenges in assessing studies to inform drowning guidelines. Many drowning reports use inappropriate classifications for drowning incidents, which significantly contributes to the underreporting of drowning. In particular, non-fatal drowning incidents are underreported because many countries do not routinely collect this data. The Danish Drowning Cohort: The Danish Drowning Cohort was established in 2016 to facilitate research to improve preventative, rescue, and treatment interventions to reduce the incidence, mortality, and morbidity of drowning. The Danish Drowning Cohort contains nationwide data on all fatal and non-fatal drowning incidents treated by the Danish Emergency Medical Services. Data are extracted from the Danish prehospital electronic medical record using a text-search algorithm (Danish Drowning Formula) and a manual validation process. The WHO definition of drowning, supported by the clarification statement for non-fatal drowning, is used as the case definition to identify drowning. All drowning patients are included, including unwitnessed incidents, non-conveyed patients, patients declared dead prehospital, or patients with obvious clinical signs of irreversible death. This method allows syndromic surveillance and monitors a nationwide cohort of fatal and non-fatal drowning incidents in near-real time to inform future prevention strategies. The Danish Drowning Cohort complies with the Utstein style for drowning reporting guidelines. The 30-day mortality is obtained through the Civil Personal Register to differentiate between fatal and non-fatal drowning incidents. In addition to prehospital data, new data linkages with other Danish registries via the patient's civil registration number will enable the examination of various additional factors associated with drowning risk. Conclusion: The Danish Drowning Cohort contains nationwide prehospital data on all fatal and non-fatal drowning incidents treated by the Danish Emergency Medical Service. It is a basis for all research on drowning in Denmark and may improve preventative, rescue, and treatment interventions to reduce the incidence, mortality, and morbidity of drowning. Plain Language Summary: The Danish Drowning Cohort includes data on fatal and non-fatal drowning incidents treated by the Emergency Medical Services from 2016 and onwards and serves as the foundation for drowning research in Denmark. Data are extracted from the Danish Prehospital Electronic Medical Record using the Danish Drowning Formula and manual validation. The research data can advance prevention, rescue, and treatment interventions, aiming to decrease drowning incidence, mortality, and morbidity. The research data follows the Utstein style for drowning reporting guidelines linked with 30-day survival. [ABSTRACT FROM AUTHOR]

Objective To construct a machine learning model to predict the risk of all-cause mortality in intensive care unit (ICU) patients. Methods Based on the intensive care medical information market Ⅲ (MIMIC-Ⅲ) database, the machine learning method was used to integrate the structured and unstructured data in the electronic medical record (EHR) to create a mortality risk prediction model for ICU patients. Results The machine learning model combined with structured and unstructured data improved the accuracy of clinical outcome prediction of ICU patients. The AUROC value of the optimized gradient enhancement model was 0. 88,indicating that the patient′s life state could be accurately predicted. Conclusion Using machine learning models, based on a small number of easily collected structured variables combined with unstructured data, can significantly improve the prediction performance of ICU patients′ mortality risk prediction models. [ABSTRACT FROM AUTHOR]

Background: Electronic medical records (EMR) can be utilized to understand the impact of the disruption in care provision caused by the pandemic. We aimed to develop and validate an algorithm to identify persons with epilepsy (PWE) from our EMR and to use it to explore the effect of the pandemic on outpatient service utilization. Methods: EMRs from the neurology specialty, covering the period from January 2018 to December 2023, were used. An algorithm was developed using an iterative approach to identify PWE with a critical lower bound of 0.91 for negative predictive value. Manual internal validation was performed. Outpatient visit data were extracted and modeled as a time series using the autoregressive integrated moving average model. All statistical analyses were performed using STATA version 14.2 (Statacorp, USA). Results: Four iterations resulted in an algorithm, with a negative predictive value 0.98 (95% CI: 0.95–0.99), positive predictive value of 0.98 (95% CI: 0.85–0.99), and an F-score accuracy of 0.96, which identified 4474 PWE. The outpatient service utilization was abruptly reduced by the pandemic, with a change of -902.1 (95%CI: -936.55 to -867.70), and the recovery has also been slow, with a decrease of -5.51(95%CI: -7.00 to -4.02). Model predictions aligned closely with actual visits with median error of -3.5%. Conclusions: We developed an algorithm for identifying people with epilepsy with good accuracy. Similar methods can be adapted for use in other resource-limited settings and for other diseases. The COVID pandemic appears to have caused a lasting reduction of service utilization among PWE. [ABSTRACT FROM AUTHOR]

The deployment of Health Information Systems (HIS) in Sri Lanka has been low in adoption compared to developed countries. There has been a dearth of studies to identify the factors that improve the adoption of HIS in developing countries. Thus, this study investigates the factors influencing the acceptance of HIS among public healthcare staff. A survey was administered among 170 medical professionals, including nurses and doctors. Partial Least Squares Structural Equation Modelling (PLS-SEM) was applied to the dataset with 5000 bootstrap subsamples. The research model was developed based on the prior literature and by extending the Technology Acceptance Model (TAM) to the context of public healthcare. A positive relationship was observed between the actual use of HIS and constructs such as perceived usefulness, perceived ease of use, attitude, behavioural intention, prior use of HIS by supervisors, computer anxiety and facilitating conditions. These findings confirm the applicability of the proposed extended TAM in the public healthcare system of a developing country. Furthermore, HIS practitioners and policymakers in the healthcare sector would find these results valuable. [ABSTRACT FROM AUTHOR]

Purpose: In the context of Chinese clinical texts, this paper aims to propose a deep learning algorithm based on Bidirectional Encoder Representation from Transformers (BERT) to identify privacy information and to verify the feasibility of our method for privacy protection in the Chinese clinical context. Methods: We collected and double-annotated 33,017 discharge summaries from 151 medical institutions on a municipal regional health information platform, developed a BERT-based Bidirectional Long Short-Term Memory Model (BiLSTM) and Conditional Random Field (CRF) model, and tested the performance of privacy identification on the dataset. To explore the performance of different substructures of the neural network, we created five additional baseline models and evaluated the impact of different models on performance. Results: Based on the annotated data, the BERT model pre-trained with the medical corpus showed a significant performance improvement to the BiLSTM-CRF model with a micro-recall of 0.979 and an F1 value of 0.976, which indicates that the model has promising performance in identifying private information in Chinese clinical texts. Conclusions: The BERT-based BiLSTM-CRF model excels in identifying privacy information in Chinese clinical texts, and the application of this model is very effective in protecting patient privacy and facilitating data sharing. [ABSTRACT FROM AUTHOR]

Background Antibiotic durations for uncomplicated skin/soft tissue infections (SSTI) often exceed the guideline-recommended 5–7 days. We assessed the effectiveness of a default duration order panel in the Electronic Health Record to reduce long prescriptions. Methods Cluster randomized trial of an SSTI order panel with default antibiotic durations (implemented 12/2021), compared to a control panel (no decision support) in 14 pediatric primary care clinics. We assessed long prescription rates from 23 months before to 12 months after order panel implementation (1/2020–12/2022). Antibiotic duration was considered long if >5 days for cellulitis or drained abscess, or >7 days for undrained abscess, impetigo, or other SSTI. Results We included 1123 and 511 encounters in intervention and control clinics, respectively. In a piecewise generalized linear model, the long prescription rate decreased from 63.8% to 54.6% (absolute difference, −9.2%) in the intervention group and from 70.0% to 54.9% (absolute difference, −15.1%) in the control group. The relative change in trajectories from pre-panel to post-panel periods did not differ significantly between intervention and control groups (P  = .488). Although used in only 29.4% of eligible encounters, intervention panel use had lower odds of long prescription compared to all other prescriptions (odds ratio 0.18). Conclusions We did not detect an overall impact of an order panel with default durations in reducing long antibiotic prescriptions for SSTIs. When ordered from the intervention panel, prescriptions were usually guideline-concordant. Effective strategies to make choosing a default duration more automatic are necessary to further reduce long prescriptions. [ABSTRACT FROM AUTHOR]

Introduction Traumatic brain injury (TBI) is among the most common conditions in the military. VA Caribbean Healthcare System (VACHS) patients with Traumatic Brain Injury (TBI) have a higher mortality rate than Veterans in other VA health care systems in the United States. The main goal of this study was to develop sociodemographic profiles and outline health characteristics of Hispanic patients with TBI treated at the VA Caribbean Healthcare System in a search for potential explanations to account for the higher mortality rate. This study advocates for equity in health services provided for minorities inside the militia. Materials and Methods Data collected from electronic medical records and VA databases were used to create sociodemographic and health characteristics profiles, in addition to survival models. The population of the study were post 911 Veteran soldiers who had been diagnosed with TBI. Adjusted models were created to provide hazard ratios (HR) for mortality risk. Results Out of the 16,549 files available from all 10 selected VA sites, 526 individuals were identified as treated at the VACHS. Of 526 subjects screened, 39 complied with the inclusion/exclusion criteria. Results include: 94.4% male, 48.7% between the ages of 21 and 41 years, 89.7% have depression, 66.7% have post-traumatic stress disorder (PTSD), 82.1% receive occupational therapy, 94.9% have severe headaches, 100% suffer from pain, 94.9% have memory problems, and 10.3% have had suicidal thoughts. Over 60% had a first-hand explosion experience, be it just the explosion or with another type of injury. Data showed that 33% of our patients had a Magnetic Resonance Imaging (MRI), 31% had a CT, 15.4% had a SPECT, and 2.6% had PET scan. Significant associations were found between MRIs and speech therapies, and MRIs and total comorbidities. The Cox proportional-hazards model for survival adjusted for age, gender, race/ethnicity, and comorbidities shows that VACHS Veterans diagnosed with a TBI had a higher mortality risk rate (HR 1.23 [95% CI 1.10, 1.37]) when compared to the other 9 health centers with the highest percentage of Hispanic Veterans. Conclusions Since explosions were the most common mechanism of injury, further research is needed into the experiences of Veterans in connection with this specific variable. A high percentage of the patients suffered from depression and PTSD. Additionally, over half of the patients had an unmeasured TBI severity. The effects these aspects have on symptomatology and how they hinder the recovery process in Hispanic patients should be examined in further detail. It is also important to highlight that family and friends' support could be key for injury treatment. This study highlights the use of the 4 types of scans (MRI, CT, PET/CT, and SPECT/CT) as ideal diagnosis tools. The alarming number of patients with suicidal thoughts should be a focus in upcoming studies. Future studies should aim to determine whether increased death rates in TBI Veterans can be linked to other United States islander territories. Concepts, such as language barriers, equal resource allocation, and the experiences of Veterans with TBIs should be further explored in this Veteran population. [ABSTRACT FROM AUTHOR]

Introduction Low back pain (LBP) is highly prevalent after lower limb amputation (LLA) and contributes to substantial reductions in quality of life and function. Towards understanding pathophysiological mechanisms underlying LBP after LLA, this article compares lumbar spine pathologies and muscle morphologies between individuals with LBP, with and without LLA. Materials and Methods We queried electronic medical records of Service members with and without LLA who sought care for LBP at military treatment facilities between January 2002 and May 2020. Two groups with cLBP, one with (n  = 15) and one without unilateral transtibial LLA (n  = 15), were identified and randomly chosen from a larger sample. Groups were matched by age, mass, and sex. Lumbar muscle morphology, Pfirrmann grades, Modic changes, facet arthrosis, Meyerding grades, and lordosis angle were determined from radiographs and magnetic resonance images available in the medical record. Independent t -tests compared variables between cohorts while multiple regression models determined if intramuscular fat influenced Pfirrmann grades. Chi-square determined differences in presence of spondylolysis and facet arthrosis. Results Lordosis angle was larger with LLA (P  = 0.01). Spondylolysis was more prevalent with LLA (P  = 0.008; 40%) whereas facet arthrosis was similar between cohorts (P  = 0.3). Muscle area was not different between cohorts, yet intramuscular fat was greater with LLA (P  ≤ 0.05). Intramuscular fat did not influence Pfirrmann grades (P  > 0.15). Conclusions Despite similar lumbar muscle size, those with unilateral LLA may be predisposed to progress to symptomatic spondylolisthesis and intramuscular fat. Surgical and/or rehabilitation interventions may mitigate long-term effects of diminished spinal health, decrease LBP-related disability, and improve function for individuals with LLA. [ABSTRACT FROM AUTHOR]

Background: Electronic medical records (EMRs) have the potential to improve and streamline the quality and safety of patient care. Harnessing the full benefits of EMR implementation depends on the utilisation of advanced features, defined as "mature usage." At present, little is known about the maturity of EMR usage by allied health professionals (AHPs). Objective: To examine current maturity of EMR use by AHPs and explore perceived barriers to mature EMR utilisation and optimisation. Method: AHPs were recruited from three health services. Participants completed a 27-question electronic questionnaire based on the EMR Adoption Framework, which measures clinician EMR utilisation (0 = paper chart, 5 = theoretical maximum) across 10 EMR feature categories. Interviews were conducted with both clinicians and managers to explore the nature of current EMR utilisation and perceived facilitators and barriers to mature usage. Results: Questionnaire responses were obtained from 193 participants AHPs. The majority of questions (74%) showed a mean score of <3, indicating a lack of mature EMR use. Pockets of mature usage were identified in the categories of health information, referrals and administration processes. Interviews with 21 clinicians and managers revealed barriers to optimisation across three themes: (1) limited understanding of EMR opportunities; (2) complexity of the EMR change process and (3) end-user and environmental factors. Conclusion: Mature usage across EMR feature categories of the EMR Adoption Framework was low. However, questionnaire and qualitative interview data suggested pockets of mature utilisation. Implications: Achieving mature allied health EMR use will require strategies implemented at the clinician, EMR support, and service levels. [ABSTRACT FROM AUTHOR]

Background: Primary ciliary dyskinesia (PCD) is a rare, genetically heterogeneous disease. Due to difficulty accessing diagnostic services and a lack of awareness of the syndrome, clinicians often fail to recognize the classic phenotype, leading to missed diagnoses. Methods: Relevant medical records were accessed through The BIG DATA QUERY AND ANALYSIS SYSTEM of Peking Union Medical College Hospital from September 1, 2012 to March 31, 2024. The search strategy included the following key terms: (bronchiectasis OR atelectasis OR recurrent cough OR recurrent expectoration OR hemoptysis) AND (sinusitis OR nasal polyps OR otitis media OR neonatal pneumonia OR neonatal respiratory distress OR ectopic pregnancy OR infertility OR artificial insemination OR assisted reproduction OR hydrocephalus OR congenital heart disease OR organ laterality defect OR right‐sided heart OR semen OR consanguineous marriage). Patients were filtered according to inclusion and exclusion criteria, and those with clinical suspicion of PCD were invited for screening, which included nasal nitric oxide and whole exome sequencing. Results: A total of 874 medical records were retrieved. After filtering based on inclusion and exclusion criteria, 65 patients with clinical suspicion of PCD were identified, 21 of whom accepted our invitation to complete PCD‐related screening. Among them, four were diagnosed with PCD, one was diagnosed with cystic fibrosis, and one was diagnosed with immunodeficiency‐21. Conclusions: This is the first study to use an electronic medical record retrieval system to identify missed diagnoses PCD. We believe that the methods used in this study can be extended to other rare diseases in the future. [ABSTRACT FROM AUTHOR]

The Chikungunya virus (CHIKV) presents substantial public health challenges in the Eastern Mediterranean Region (EMR), with its prevalence and interaction with other arboviruses (ABVs) remaining poorly understood. This systematic review and meta-analysis aimed to assess the prevalence of CHIKV and its association with other ABVs, such as dengue virus (DENV), Rift Valley fever virus (RVFV), malaria, and yellow fever virus (YFV), in the EMR. We systematically searched databases including PubMed, Embase, Web of Science, Scopus, Cochrane Library, CINAHL, PsycINFO, and ScienceDirect to identify epidemiological studies that report CHIKV prevalence and provide odds ratios (ORs) for CHIKV compared to other ABVs. Data analysis was performed using a random-effects model. Heterogeneity was evaluated using the χ2 test and I 2 statistic. The GRADE approach was used to evaluate the quality of the studies while the AXIS tool, NOS tool, and AHRQ checklist assessed the risk of bias. The meta-analysis revealed a significant prevalence of CHIKV in the EMR. However, the studies exhibited heterogeneity, indicating variability in the results. A comparison of CHIKV with other ABVs did not show any statistically significant differences in prevalence. The meta-analysis found a notable prevalence of CHIKV in the EMR. The results also indicated that the prevalence of CHIKV is comparable to that of other ABVs in the region. These findings provide an overview of the burden of CHIKV in the EMR. [ABSTRACT FROM AUTHOR]

Bartonella is a vector-borne zoonotic pathogen, which could also be transmitted directly and cause a variety of clinical illnesses. This study aimed to investigate the prevalence of Bartonella in countries in the WHO Eastern Mediterranean Region (WHO-EMR) region. We searched using the keywords Bartonella and the name of each country in the WHO-EMR in databases such as PubMed, ISI (Web of Science), Scopus, and Google Scholar, with a publication date range of 1990–2022 and limited to English articles. We evaluated the quality of the studies using the STROBE 6-item checklist and used the random effects model to integrate the findings of the included studies. A total of 45 papers out of 240 were included in the analysis. The results showed the prevalence of Bartonella infection among endocarditis patients was 3.8% (95% CI: 0.2–7.4) and the seroprevalence of Bartonella among other people was 27.5% (95% CI: 13.5–41.5). The overall prevalence of Bartonella spp. among animals, as determined by molecular, serological, and culture methods, was 11.9% (95% CI: 5.7–18.2), 38.9% (95% CI: 27.5–50.2), and 1.7% (95% CI: 0.5–2.9), respectively. Furthermore, the prevalence of Bartonella spp. in ectoparasites was 3.9% (95% CI: 3.5–5.2), with fleas (6.2%) showing a higher prevalence compared to lice (4.9%) and ticks (1.0%). The detection of Bartonella in all animal and ectoparasites species and human populations in the WHO-EMR with prevalence ranging from 0.3% to 23% is concerning, emphasizes the importance of conducting more comprehensive studies to gain a deeper understanding of the spread of Bartonella in these areas. [ABSTRACT FROM AUTHOR]

Introduction: While an association between hidradenitis suppurativa (HS) and inflammatory arthritis has been reported in clinical studies, the potential link between HS and gout remains uncertain. As HS and gout share common immunological pathways, we conducted a retrospective cohort study to determine whether HS patients are at an increased risk of developing gout in the future. Methods: This retrospective multicenter cohort study obtained information through the US collaborative network, a subset of the TriNetX research network. Patients diagnosed with HS between January 01, 2005, and December 31, 2017, were recruited, and a 1:1 propensity score matching was conducted to identify appropriate controls. The hazard ratio (HR) for the new-onset gout in HS patients was subsequently calculated. Results: Compared to individuals without HS, those with HS were associated with a 1.39-fold higher risk (95% confidence interval [CI], 1.20, 1.62) of developing new-onset gout within 5 years after the index date. This association remained significant in shorter follow-up times and sensitivity analyses utilizing different matching models. For both male and female HS patients, the risk of developing new-onset gout within 5 years after the index date was statistically significant, with respective HRs of 1.61 (95% CI, 1.28, 2.02) for males and 1.41 (95% CI, 1.11,1.78) for females. Conclusion: HS patients are at a high risk of developing gout within 5 years after an HS diagnosis while comparing with non-HS controls. [ABSTRACT FROM AUTHOR]

Patient privacy data security is a pivotal area of research within the burgeoning field of smart healthcare. This study proposes an innovative hybrid blockchain-based framework for the secure sharing of electronic medical record (EMR) data. Unlike traditional privacy protection schemes, our approach employs a novel tripartite blockchain architecture that segregates healthcare data across distinct blockchains for patients and healthcare providers while introducing a separate social blockchain to enable privacy-preserving data sharing with authorized external entities. This structure enhances both security and transparency while fostering collaborative efforts across different stakeholders. To address the inherent complexity of managing multiple blockchains, a unique cross-chain signature algorithm is introduced, based on the Boneh-Lynn-Shacham (BLS) signature aggregation technique. This algorithm not only streamlines the signature process across chains but also strengthens system security and optimizes storage efficiency, addressing a key challenge in multi-chain systems. Additionally, our external sharing algorithm resolves the prevalent issue of medical data silos by facilitating better data categorization and enabling selective, secure external sharing through the social blockchain. Security analyses and experimental results demonstrate that the proposed scheme offers superior security, storage optimization, and flexibility compared to existing solutions, making it a robust choice for safeguarding patient data in smart healthcare environments. [ABSTRACT FROM AUTHOR]

Background: The fourth universal definition of myocardial infarction (MI) introduced the differentiation of acute myocardial injury from MI. In this study, we developed a computational phenotype for distinct identification of acute myocardial injury and MI within electronic medical records (EMRs). Methods: Two cohorts were used from a Calgary-wide EMR system: a chart review of 3042 randomly selected inpatients from Dec 2014 to Jun 2015; and 11,685 episodes of care that included cardiac catheterization from Jan 2013 to Apr 2017. Electrocardiogram (ECG) reports were processed using natural language processing and combined with high-sensitivity troponin lab results to classify patients as having an acute myocardial injury, MI, or neither. Results: For patients with an MI diagnosis, only 64.0% (65.7%) in the catheterized cohorts (chart review cohort) had two troponin measurements within 6 h of each other. For patients with two troponin measurements within 6 h; of those with an MI diagnosis, our phenotype classified 25.2% (31.3%) with an acute myocardial injury and 62.2% (55.2%) with an MI in the catheterized cohort (chart review cohort); and of those without an MI diagnosis, our phenotype classified 12.9% (12.4%) with an acute myocardial injury and 10.0% (13.1%) with an MI in the catheterized cohort (chart review cohort). Conclusions: Patients with two troponin measurements within 6 h, identified by our phenotype as having either an acute myocardial injury or MI, will at least meet the diagnostic criteria for an acute myocardial injury (barring lab errors) and indicate many previously uncaptured cases. Myocardial infarctions are harder to be certain of because ECG report findings might be superseded by evidence not included in our phenotype, or due to errors with the natural language processing. [ABSTRACT FROM AUTHOR]

Introduction Methods Results Discussion A learning health system (LHS) approach is a collaborative model that continuously examines, evaluates, and re‐evaluates data eventually transforming it into knowledge. High quantity of high‐quality data are needed to establish this model. The purpose of this article is to describe the collaborative discovery process used to identify and standardize clinical data documented during daily multidisciplinary inpatient rehabilitation that would then allow access to these data to conduct comparative effectiveness research.CARE4TBI is a prospective observational research study designed to capture clinical data within the standard inpatient rehabilitation documentation workflow at 15 TBI Model Systems Centers in the US. Three groups of stakeholders guided project development: therapy representative work group (TRWG) consisting of frontline therapists from occupational, physical, speech‐language, and recreational therapies; rehabilitation leader representative group (RLRG); and informatics and information technology team (IIT). Over a 12‐month period, the three work groups and research leadership team identified the therapeutic components captured within daily documentation throughout the duration of inpatient TBI rehabilitation.Data brainstorming among the groups created 98 distinct categories of data with each containing a range of data elements comprising a total of 850 discrete data elements. The free‐form data were sorted into three large categories and through review and discussion, reduced to two categories of prospective data collection—session‐level and therapy activity‐level data. Twelve session data elements were identified, and 54 therapy activities were identified, with each activity containing discrete sub‐categories for activity components, method of delivery, and equipment or supplies. A total of 561 distinct meaningful data elements were identified across the 54 activities.The CARE4TBI data discovery process demonstrated feasibility in identifying and capturing meaningful high quantity and high‐quality treatment data across multiple disciplines and rehabilitation sites, setting the foundation for a LHS coalition for acute traumatic brain injury rehabilitation. [ABSTRACT FROM AUTHOR]

This article presents a thorough examination of the progress and limitations in the application of Natural Language Processing (NLP) and Machine Learning (ML), particularly Deep Learning (DL), in the healthcare industry. This paper examines the progress and limitations in the utilisation of Natural Language Processing (NLP) and Machine Learning (ML) in the healthcare field, specifically in relation to Electronic Medical Records (EMRs). The review also examines the incorporation of Natural Language Processing (NLP) and Machine Learning (ML) in medical imaging as a supplementary field, emphasising the transformative impact of these technologies on the analysis of healthcare data and patient care. This review attempts to analyse both fields in order to offer insights into the current state of research and suggest potential chances for future advancements. The focus is on the use of these technologies in Electronic Medical Records (EMRs) and medical imaging. The review methodically detects, chooses, and assesses literature published between 2015 and 2023, utilizing keywords pertaining to natural language processing (NLP) and healthcare in databases such as SCOPUS. After applying precise inclusion criteria, 100 papers were thoroughly examined. The paper emphasizes notable progress in utilizing NLP and ML methodologies to improve healthcare decision-making, extract information from unorganized data, and evaluate medical pictures. The key findings highlight the successful combination of natural language processing (NLP) and image processing to enhance the accuracy of diagnoses and improve patient care. The study also demonstrates the effectiveness of deep learning-based NLP pipelines in extracting valuable information from electronic medical records (EMRs). Additionally, the research suggests that NLP has the potential to optimize the allocation of medical imaging resources. The identified gaps encompass the necessity for scalable and practical implementations, improved interdisciplinary collaboration, the consideration of ethical factors, the analysis of longitudinal patient data, and the customization of approaches for specific medical situations. Subsequent investigations should focus on these deficiencies in order to fully exploit the capabilities of natural language processing (NLP) and machine learning (ML) in the healthcare sector, consequently enhancing patient outcomes and the delivery of healthcare services. [ABSTRACT FROM AUTHOR]

Background: There are limited hospital‐acquired malnutrition (HAM) studies among the plethora of malnutrition literature, and a few studies utilise electronic medical records to assist with malnutrition care. This study therefore aimed to determine the point prevalence of HAM in long‐stay adult patients across five facilities, whether any descriptors could assist in identifying these patients and whether a digital Dashboard accurately reflected 'real‐time' patient nutritional status. Methods: HAM was defined as malnutrition first diagnosed >14 days after hospital admission. Eligible patients were consenting adult (≥18 years) inpatients with a length of stay (LOS) >14 days. Palliative, mental health and intensive care patients were excluded. Descriptive, clinical and nutritional data were collected, including nutritional status, and whether a patient had hospital‐acquired malnutrition to determine point prevalence. Descriptive Fisher's exact and analysis of variance (ANOVA) tests were used. Results: Eligible patients (n = 134) were aged 68 ± 16 years, 52% were female and 92% were acute admissions. HAM and malnutrition point prevalence were 4.5% (n = 6/134) and 19% (n = 26/134), respectively. Patients with HAM had 72 days greater LOS than those with malnutrition present on admission (p < 0.001). A high proportion of HAM patients were inpatients at a tertiary facility and longer‐stay wards. The Dashboard correctly reflected recent ward dietitian assessments in 94% of patients at one facility (n = 29/31). Conclusions: HAM point prevalence was 4.5% among adult long‐stay patients. Several descriptors may be suitable to screen for at‐risk patients in future studies. Digital Dashboards have the potential to explore factors related to HAM. Key points: Hospital‐acquired malnutrition (HAM) is a significant issue among malnourished long‐stay patients. Future research should explore whether factors such as length of stay, tertiary facilities and rehabilitation or subacute wards could assist with screening for HAM. With effective governance digital Dashboards have the potential to explore factors related to HAM. [ABSTRACT FROM AUTHOR]

The construction of medical heterogeneous map for disease diagnosis using electronic medical records is a research hotspot of medical artificial intelligence. However, existing disease diagnosis networks based on message passing mechanisms have certain limitations. For instance, these models exhibit limited expressiveness and suffer from issues such as over-compression and oversmoothing, which subsequently affect the accuracy of disease diagnosis. To address these issues, a disease diagnosis framework named Trans4DD is proposed, based on the medical heterogeneous graph Transformer. In Trans4DD's medical heterogeneous graph encoder, edge type embeddings and residual connections are introduced. Edge type embeddings effectively capture the node structure and heterogeneous information in the graph. Residual connections aid in avoiding oversmoothing and gradient vanishing problems. A node-level graph Transformer is adopted to overcome the limitations of the message passing mechanism. By employing a multi-hop node context sampling strategy, a broader range of global attention mechanisms is introduced to obtain more accurate patient representations. Experimental results on the MIMIC-IV dataset demonstrate that Trans4DD outperforms other baseline methods in terms of disease diagnosis performance, effectively enhancing the accuracy of disease diagnosis. [ABSTRACT FROM AUTHOR]

Purpose: In the context of ophthalmologic practice, there has been a rapid increase in the amount of data collected using electronic health records (EHR). Artificial intelligence (AI) offers a promising means of centralizing data collection and analysis, but to date, most AI algorithms have only been applied to analyzing image data in ophthalmologic practice. In this review we aimed to characterize the use of AI in the analysis of EHR, and to critically appraise the adherence of each included study to the CONSORT-AI reporting guideline. Methods: A comprehensive search of three relevant databases (MEDLINE, EMBASE, and Cochrane Library) from January 2010 to February 2023 was conducted. The included studies were evaluated for reporting quality based on the AI-specific items from the CONSORT-AI reporting guideline. Results: Of the 4,968 articles identified by our search, 89 studies met all inclusion criteria and were included in this review. Most of the studies utilized AI for ocular disease prediction (n = 41, 46.1%), and diabetic retinopathy was the most studied ocular pathology (n = 19, 21.3%). The overall mean CONSORT-AI score across the 14 measured items was 12.1 (range 8–14, median 12). Categories with the lowest adherence rates were: describing handling of poor quality data (48.3%), specifying participant inclusion and exclusion criteria (56.2%), and detailing access to the AI intervention or its code, including any restrictions (62.9%). Conclusions: In conclusion, we have identified that AI is prominently being used for disease prediction in ophthalmology clinics, however these algorithms are limited by their lack of generalizability and cross-center reproducibility. A standardized framework for AI reporting should be developed, to improve AI applications in the management of ocular disease and ophthalmology decision making. [ABSTRACT FROM AUTHOR]

Aim: To assess a commercially available electronic whiteboard's usability and acceptability in isolation rooms, focusing on improving nurse–patient communication and supporting data input. Design: A cross‐sectional study with quantitative and qualitative mixed methods. Methods: We evaluated the usability and acceptability of electronic whiteboards among nurses using scenarios in a virtual isolation room environment. Results: Nurses recognised the electronic whiteboard as a valuable tool for communication and error reductions in record‐keeping but noted a learning curve for less tech‐savvy users. Positive correlations were found between perceived usefulness, ease of use and adoption intent. Despite challenges, electronic whiteboards show promise for enhancing patient care, requiring comprehensive training and management systems. Time allocation in patient wards and nurse–patient interactions are crucial considerations. Conclusion: Electronic whiteboards have usability and acceptability as a tool to improve nurse–patient communication. However, considering technical issues and staff resistance, a management system and user training are necessary. Implications for the Profession and/or Patient Care: Nurses perceive electronic whiteboards as user‐friendly and as facilitating data input. Reporting Method: TREND (Nonrandomised evaluations of behavioural and public health interventions). Patient or Public Contribution: No patient or public contribution. [ABSTRACT FROM AUTHOR]

Background The complexities of diagnosing cancer in general practice has driven the development of quality improvement (QI) interventions, including clinical decision support (CDS) and auditing tools. Future Health Today (FHT) is a novel QI tool, consisting of CDS at the point-of-care, practice population-level auditing, recall, and the monitoring of QI activities. Objectives Explore the acceptability and usability of the FHT cancer module, which flags patients with abnormal test results that may be indicative of undiagnosed cancer. Methods Interviews were conducted with general practitioners (GPs) and general practice nurses (GPNs), from practices participating in a randomized trial evaluating the appropriate follow-up of patients. Clinical Performance Feedback Intervention Theory (CP-FIT) was used to analyse and interpret the data. Results The majority of practices reported not using the auditing and QI components of the tool, only the CDS which was delivered at the point-of-care. The tool was used primarily by GPs; GPNs did not perceive the clinical recommendations to be within their role. For the CDS, facilitators for use included a good workflow fit, ease of use, low time cost, importance, and perceived knowledge gain. Barriers for use of the CDS included accuracy, competing priorities, and the patient population. Conclusions The CDS aligned with the clinical workflow of GPs, was considered non-disruptive to the consultation and easy to implement into usual care. By applying the CP-FIT theory, we were able to demonstrate the key drivers for GPs using the tool, and what limited the use by GPNs. [ABSTRACT FROM AUTHOR]

Treatment of periprosthetic joint infections (PJIs) typically requires more resource utilization than primary total joint arthroplasty. This study quantifies the amount of time spent in the electronic medical record (EMR) for patients who have PJI requiring surgical intervention. A retrospective analysis of EMR activity for 165 hip and knee PJIs was performed to capture work during the preoperative and postoperative time periods. Independent sample t tests were conducted to compare total time based on procedure, age, insurance, health literacy, sex, race, and ethnicity. The EMR work performed by the orthopaedic team was 338.4 minutes (min) (SD 130.3), with 119.4 minutes (SD 62.8) occurring preoperatively and 219.0 minutes (SD 112.9) postoperatively. Preoperatively, the surgeon's work accounted for 35.7 minutes (SD 25.4), mid-level providers 21.3 minutes (SD 15.9), nurses 38.6 minutes (SD 36.8), and office staff 32.7 minutes (SD 29.9). Infectious disease colleagues independently performed 158.9 minutes (SD 108.5) of postoperative work. Overall, PJI of the knees required more postoperative work. Secondary analysis revealed that patients who have hip PJI and a body mass index <30 and patients <65 years of age required more work when compared to the PJI of heavier and older individuals. There was no difference in total work based on insurance, health literacy, race, or ethnicity. Over 8 hours of administrative work is required for surgical management of PJI. Surgeons alone performed 451% more work for PJI during the preoperative period (7.9 versus 35.7 min) compared to primary total joint arthroplasty. In efforts to provide best care for our sickest patients, much work is required perioperatively. This work is necessary to consider when assigning value and physician reimbursement. [ABSTRACT FROM AUTHOR]

Background: General practice (GP) is crucial to primary care delivery in the Republic of Ireland and is almost fully computerised. General practice teams were the first point of contact for much COVID-19-related care and there were concerns routine healthcare activities could be disrupted due to COVID-19 and related restrictions. Aims: The study aimed to assess effects of the pandemic on GP activity through analysis of electronic medical record data from general practice clinics in the Irish Midwest. Methods: A retrospective, descriptive study of electronic medical record data relating to patient record updates, appointments and medications prescribed across 10 GP clinics over the period 2019–2021 inclusive. Results: Data relating to 1.18 million record transactions for 32 k patients were analysed. Over 500 k appointments were examined, and demographic trends presented. Overall appointment and prescribing activity increased over the study period, while a dip was observed immediately after the pandemic's arrival in March 2020. Delivery of non-childhood immunisations increased sixfold as a result of COVID-19, childhood immunisation activity was maintained, while cervical smears decreased in 2020 as the screening programme was halted. A quarter of consultations in 2020 and 2021 were teleconsultations, and these were more commonplace for younger patients. Conclusions: General practice responded robustly to the pandemic by taking on additional activities while maintaining routine services where possible. The shift to teleconsulting was a significant change in workflow. Analysing routinely collected electronic medical record data can provide valuable insights for service planning, and access to these insights would be beneficial for future pandemic responses. [ABSTRACT FROM AUTHOR]

This study aims to evaluate the readiness of Class D Primary Hospital Sendawar in implementing Electronic Medical Records (EMR) using DOQ-IT, focusing on four aspects: human resources (HR), organizational work culture, governance and leadership, and technology infrastructure. The method used was qualitative through in-depth interviews with health workers and hospital staff, and thematic analysis. The results showed that HR had a positive understanding of the benefits of EMR, but lacked information technology skills and adequate training. Organizational work culture supports the implementation of EMR, but needs strengthening in the formation of special teams and policy development. Management support for EMR is strong, but a clear strategy for internal coordination is needed. Technology infrastructure is currently inadequate, with problems in hardware, software and internet networks. The research conclusions emphasize the need for increased HR training, policy development, and infrastructure improvements. RSUD Kelas D Pratama Sendawar is advised to adopt a comprehensive approach to managing change to facilitate an effective transition to the EMR system. [ABSTRACT FROM AUTHOR]

Copyright of International Journal of Professional Business Review (JPBReview) is the property of Open Access Publications LLC and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Abstract Background Patients with transient ischemic attack (TIA) face a significantly increased risk of stroke. However, TIA screening and early detection rates are low, especially in developing countries. This study aims to develop an inclusive and practical TIA risk prediction model using machine learning (ML) that performs well in both hospital and resource-limited clinic settings. This model is essential for initiating the first ML-enabled learning health system (LHS) unit designed for routine and equitable TIA screening and early detection across broad populations. Methods Employing a novel protocol, this study first standardized data from a hospital’s electronic medical records (EMR) to construct inclusive TIA risk prediction ML models using a data-centric approach. Subsequently, a quantitative distribution of TIA risk factors was applied in feature engineering to reduce the number of variables for a practical ML model. This refined model initiated a TIA ML-LHS unit that is capable of continuously updating with new EMR data from hospitals and clinics. Additionally, the practical model underwent external validation using data from another hospital. Results The inclusive 150-variable ML models, derived from all available EMR variables for TIA, achieved a recall of 0.868 and an accuracy of 0.886 in predicting TIA risk. Further feature engineering produced a practical XGBoost model with 20 variables, maintaining acceptable performance of 0.855 recall and 0.796 accuracy. The initialized TIA ML-LHS unit, based on the practical model, achieved performance metrics of 0.830 recall, 0.726 precision, 0.816 ROC-AUC, and 0.812 accuracy. The model also performed well in external validation, confirming its effectiveness with patient data from different clinical settings. Conclusions This study developed the first inclusive and practical TIA XGBoost model from full hospital EHR and initiated the first TIA risk prediction ML-LHS unit. This TIA model, which requires only 20 variables, enables the ML-LHS to serve not only patients in hospitals but also those in resource-limited clinics. These results have significant implications for expanding risk-based TIA screening in community and rural clinics, thereby enhancing early detection of TIA among underserved populations and improving health equity. The novel protocol used in this study is also applicable for initiating ML-LHS units for various preventable diseases, providing a new system-level approach to responsible AI development and applications.

Objective To explore the construction of a real-world data annotation platform, and compare the real-world data extraction performance of retrieval augmented generation (RAG) combined with large language models and pre-training fine-tuning methods for pre-trained language models.Methods Taking the pathological records of bladder cancer in the real world electronic medical record data as an example, a real-world data annotation platform was built. Based on the platform annotation data, the effects of automatic extraction of cancer typing and staging of bladder cancer using RAG combined with GPT-3.5, and the pre- training fine tuning method based on BERT and RoBERTa models were compared. Results The extraction effects of the pre-training and fine-tuning model based on the fine-tuning of the full-training set were better than that of RAG combined with large model method and pre-training and fine-tuning model with the few-shot fine-tuning, and the effects of RoBERTa model were generally better than that of BERT model, but the extraction effects of these methods needs to be improved totally. The F1 scores for extracting bladder cancer typing, T staging, and N staging in the test set, using the RoBERTa model fine-tuned with the entire training set, were 71.06%, 50.18%, and 73.65% respectively. Conclusion Pre-trained language models have the application potential in processing clinical unstructured data, but there is still room for improvement in the information extraction effect of existing methods. Future work requires further optimization of models or training strategies to accelerate data empowerment.

Abstract Background Rheumatoid Arthritis (RA) is a chronic inflammatory disease that is primarily diagnosed and managed by rheumatologists; however, it is often primary care providers who first encounter RA-related symptoms. This study developed and validated a case definition for RA using national surveillance data in primary care settings. Methods This cross-sectional validation study used structured electronic medical record (EMR) data from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN). Based on the reference set generated by EMR reviews by five experts, three machine learning steps: ‘bag-of-words’ approach to feature generation, feature reduction using a feature importance measure coupled with recursive feature elimination and clustering, and classification using tree-based methods (Decision Tree, Random Forest, and Extreme Gradient Boosting). The three tree-based algorithms were compared to identify the procedure that generated the optimal evaluation metrics. Nested cross-validation was used to allow evaluation and comparison and tuning of models simultaneously. Results Of 1.3 million patients from seven Canadian provinces, 5,600 people aged 19 + were randomly selected. The optimal algorithm for selecting RA cases was generated by the XGBoost classification method. Based on feature importance scores for features in the XGBoost output, a human-readable case definition was created, where RA cases are identified when there are at least 2 occurrences of text “rheumatoid” in any billing, encounter diagnosis, or health condition table of the patient chart. The final case definition had sensitivity of 81.6% (95% CI, 75.6–86.4), specificity of 98.0% (95% CI, 97.4–98.5), positive predicted value of 76.3% (95% CI, 70.1–81.5), and negative predicted value of 98.6% (95% CI, 98.0-98.6). Conclusion A case definition for RA in using primary care EMR data was developed based off the XGBoost algorithm. With high validity metrics, this case definition is expected to be a reliable tool for future epidemiological research and surveillance investigating the management of RA in CPCSSN dataset.