Search

Frailty is associated with poor post-operative outcomes. However, frailty assessments in clinical practice are challenging due to the need for more resources and pragmatic complexities. We aimed to create a pre-operative frailty ascertainment using machine learning (ML) from electronic health record (EHR) data that is built on the Fried frailty phenotype. We leveraged a research database of 8,999 individuals aged 65 years and above who underwent a surgery. Healthcare providers administered a pre-operative frailty assessment using the Fried frailty phenotype. We built ML models to predict pre-operative frailty as a whole and by surgical service. We used the SHapley Additive exPlanations (SHAP) to interpret the results. Comparisons with the accumulated deficit approach were made using Pearson's correlation coefficients and McNemar's test. The ML model achieved an AUC of 0.74 in predicting pre-operative frailty phenotype. The ML models' predictive power varied by surgical services, with AUC ranging from 0.63 to 0.81. SHAP showed that advanced age, anemia, chronic pulmonary disease, and literacy deficit were the most important features of pre-operative frailty. The ML model had similar performance to the accumulated deficit approach, with positive correlation (r: 0.36-0.58, P> 0.001) between frailty risk scores predicted by ML and frailty indices obtained from the accumulated deficit approach. Our approach provided new insights into the importance of specific EHR features in pre-operative frailty assessment. ML may offer an alternative method to predict frailty in pre-operative settings. Further validation on external data sources is required to generalize our ML models. [ABSTRACT FROM AUTHOR]

Objective To develop and evaluate a data-driven process to generate suggestions for improving alert criteria using explainable artificial intelligence (XAI) approaches. Methods We extracted data on alerts generated from January 1, 2019 to December 31, 2020, at Vanderbilt University Medical Center. We developed machine learning models to predict user responses to alerts. We applied XAI techniques to generate global explanations and local explanations. We evaluated the generated suggestions by comparing with alert's historical change logs and stakeholder interviews. Suggestions that either matched (or partially matched) changes already made to the alert or were considered clinically correct were classified as helpful. Results The final dataset included 2 991 823 firings with 2689 features. Among the 5 machine learning models, the LightGBM model achieved the highest Area under the ROC Curve: 0.919 [0.918, 0.920]. We identified 96 helpful suggestions. A total of 278 807 firings (9.3%) could have been eliminated. Some of the suggestions also revealed workflow and education issues. Conclusion We developed a data-driven process to generate suggestions for improving alert criteria using XAI techniques. Our approach could identify improvements regarding clinical decision support (CDS) that might be overlooked or delayed in manual reviews. It also unveils a secondary purpose for the XAI: to improve quality by discovering scenarios where CDS alerts are not accepted due to workflow, education, or staffing issues. [ABSTRACT FROM AUTHOR]

Objective To measure pediatrician adherence to evidence-based guidelines in the treatment of young children with attention-deficit/hyperactivity disorder (ADHD) in a diverse healthcare system using natural language processing (NLP) techniques. Materials and Methods We extracted structured and free-text data from electronic health records (EHRs) of all office visits (2015-2019) of children aged 4-6 years in a community-based primary healthcare network in California, who had ≥1 visits with an ICD-10 diagnosis of ADHD. Two pediatricians annotated clinical notes of the first ADHD visit for 423 patients. Inter-annotator agreement (IAA) was assessed for the recommendation for the first-line behavioral treatment (F-measure = 0.89). Four pre-trained language models, including BioClinical Bidirectional Encoder Representations from Transformers (BioClinicalBERT), were used to identify behavioral treatment recommendations using a 70/30 train/test split. For temporal validation, we deployed BioClinicalBERT on 1,020 unannotated notes from other ADHD visits and well-care visits; all positively classified notes (n  = 53) and 5% of negatively classified notes (n  = 50) were manually reviewed. Results Of 423 patients, 313 (74%) were male; 298 (70%) were privately insured; 138 (33%) were White; 61 (14%) were Hispanic. The BioClinicalBERT model trained on the first ADHD visits achieved F1 = 0.76, precision = 0.81, recall = 0.72, and AUC = 0.81 [0.72-0.89]. Temporal validation achieved F1 = 0.77, precision = 0.68, and recall = 0.88. Fairness analysis revealed low model performance in publicly insured patients (F1 = 0.53). Conclusion Deploying pre-trained language models on a variable set of clinical notes accurately captured pediatrician adherence to guidelines in the treatment of children with ADHD. Validating this approach in other patient populations is needed to achieve equitable measurement of quality of care at scale and improve clinical care for mental health conditions. [ABSTRACT FROM AUTHOR]

Objective: Positive antinuclear antibodies (ANAs) cause diagnostic dilemmas for clinicians. Currently, no tools exist to help clinicians interpret the significance of a positive ANA in individuals without diagnosed autoimmune diseases. We developed and validated a risk model to predict risk of developing autoimmune disease in positive ANA individuals. Methods: Using a de-identified electronic health record (EHR), we randomly chart reviewed 2,000 positive ANA individuals to determine if a systemic autoimmune disease was diagnosed by a rheumatologist. A priori, we considered demographics, billing codes for autoimmune disease-related symptoms, and laboratory values as variables for the risk model. We performed logistic regression and machine learning models using training and validation samples. Results: We assembled training (n = 1030) and validation (n = 449) sets. Positive ANA individuals who were younger, female, had a higher titer ANA, higher platelet count, disease-specific autoantibodies, and more billing codes related to symptoms of autoimmune diseases were all more likely to develop autoimmune diseases. The most important variables included having a diseasespecific autoantibody, number of billing codes for autoimmune disease-related symptoms, and platelet count. In the logistic regression model, AUC was 0.83 (95% CI 0.79-0.86) in the training set and 0.75 (95% CI 0.68-0.81) in the validation set. Conclusion: We developed and validated a risk model that predicts risk for developing systemic autoimmune diseases and can be deployed easily within the EHR. The model can risk stratify positive ANA individuals to ensure high-risk individuals receive urgent rheumatology referrals while reassuring low-risk individuals and reducing unnecessary referrals. [ABSTRACT FROM AUTHOR]

This paper explores the balance between fairness and performance in machine learning classification, predicting the likelihood of a patient receiving antimicrobial treatment using structured data in community nursing wound care electronic health records. The data includes two important predictors (gender and language) of the social determinants of health, which we used to evaluate the fairness of the classifiers. At the same time, the impact of various groupings of language codes on classifiers' performance and fairness is analyzed. Most common statistical learning-based classifiers are evaluated. The findings indicate that while K-Nearest Neighbors offers the best fairness metrics among different grouping settings, the performance of all classifiers is generally consistent across different language code groupings. Also, grouping more variables tends to improve the fairness metrics over all classifiers while maintaining their performance. [ABSTRACT FROM AUTHOR]

Licensed practical nurses (LPNs) are the second largest occupational group and the largest group in the social and healthcare sector in Finland, and they have an extensive working environment. Like other health and social care professionals, LPNs also use health information systems (HIS) and client information systems (CIS) in their daily work. The aim of this study was to describe LPNs' perceptions of the benefits of information systems in daily patient care. The information systems include the main HIS or CIS that the respondents mainly use in their work. The data comprised 3 866 LPNs' responses were collected via an online survey in 2022. Most of the LPNs work in social care using the Lifecare system. ESKO is used in public health care and was rated as the most popular system that LPNs use regarding the benefits of information systems. Highly experienced LPNs seem to rate the benefits of information systems higher than LPNs who have just started working. [ABSTRACT FROM AUTHOR]

Situation awareness (SA) is an important non-technical skill for nurses. Nurses interact directly with patients and review their clinical signs. If we improve nurses' SA, they will likely detect clinical changes and prevent patient harm. A clinical endeavor that can benefit from improved nurses' SA is the prevention of Healthcare-Acquired Urinary Tract Infection (HAUTI). Electronic Health Records contain comprehensive nursing assessment data that researchers can use to analyze trends and provide a context-based understanding of the infection risk factors. We conducted a study that involved extracting nursing assessment data and preparing it for supervised learning algorithms and predicting HAUTI. In this paper, we share the methods we used to prepare the data for supervised learning algorithms and present the challenges related to data missingness. [ABSTRACT FROM AUTHOR]

Only ten percent of nursing schools in the US are currently using EHRs in their curriculum. This is seen to decrease students' effectiveness as new nurses. We began implementation of EHR use in Spring 2024 and have started collecting data as to the effectiveness of this process. This is a starting point using semester 1 junior (J1) students and we plan to have this process follow these students throughout all 4 semesters and assess them in their first year of practice to gain full understanding of the benefits. [ABSTRACT FROM AUTHOR]

This interpretative case study demonstrates how the work system properties and roles of clinical departments are perceived to shape their adoption of an organisation-wide electronic health record. In hospitals, with their heterogeneous and powerful clinical departments, partial or non-adoption by some departments may jeopardise effective overall use. Viewing a hospital's EHR adoption from a sociotechnical fit perspective, we unravel how clinical departments' adoption of this organisation-wide system was primarily shaped by their critical work system dependencies rather than their role during its implementation. From the interviewees' accounts, three orientations emerge: (1) organisation-oriented resulting in unconditional adoption; (2) department-oriented resulting in problematic or delayed adoption with high compromise costs; (3) environment-oriented resulting in at least partial non-adoption or extra design costs. Our subsequent analysis leads to the identification of four types of subunit-level characteristics influencing adoption. Their interplay informs departments in balancing between accepting compromise costs and negotiating design costs. We develop a diagnostic system model of subunit-level adoption that takes the embeddedness of the subunit into account and thereby complements individual- and organisation-level adoption theories. An accompanying heuristic may enable hospital managers to better anticipate the heterogeneity in departments' adoption. [ABSTRACT FROM AUTHOR]

Purpose: The application of artificial intelligence (AI) in ophthalmology has shown significant promise across various clinical domains. This study addresses the need for assessing the predictive value of AI models utilizing electronic health records (EHRs) for diagnosis, prognostication and management of ocular diseases. Methods: A search was conducted using Ovid MEDLINE, Ovid EMBASE, and Cochrane Central for relevant studies published between January 2010 to February 2023 on predictive value of AI algorithms in ophthalmic EHRs. The study followed the Preferred Reporting Items for a Systematic Review and Meta-analysis (PRISMA) guidelines, with a protocol registered on Prospero (registration number: CRD42022303128). A bivariate random effects model was used to perform the meta-analysis. The ROBINS-I tool was used to assess methodological quality and applicability of the included studies. Results: Out of 4968 initial records, 41 studies met the inclusion criteria, comprising a total of 639,637 patients, with an average disease prevalence of 11%. The studies exhibited a diagnostic odds ratio of 18.527 (95% CI: 9.654–35.556), sensitivity of 0.811 (95% CI: 0.751−0.859), specificity of 0.812 (95% CI: 0.736−0.87) and Grading of Recommendations, Assessment, Development and Evaluation (GRADE) moderate. Likelihood ratios (LR+ and LR−) were 4.316 (95% CI: 2.938–6.339) and 0.233 (95% CI: 0.169−0.322), respectively. False positive rate was 0.188 (95% CI: 0.13−0.264). Inter-rate concordance for ROBINS-I scoring had a kappa score of 0.83. Out of the 41 studies, 22 had an overall low risk of bias, and 19 had a moderate risk of bias. There was a low to moderate quality of body of evidence for the reported outcomes. Conclusion: This meta-analysis affirms the substantial potential of AI models utilizing EHRs for predictive modeling and clinical management of ocular diseases. Future research should emphasize external validation and standardized reporting for better implementation of AI in ophthalmic practice.

Objective: To assess the impact of a multi-pronged educational approach on the knowledge, attitudes, and behaviors regarding Familial Hypercholesterolemia (FH) management at a large academic medical center with the aim of empowering primary care clinicians (PCC) to diagnose and treat FH. Methods: A comprehensive educational program for PCCs on FH management was developed and piloted from July 2022 to March 2024. Components of our intervention included: 1. Implementation of a novel clinical decision support tool in the electronic medical record for FH management, 2. Development and dissemination of an interactive educational website focused on FH and its management, 3. Delivery of virtual instructional sessions to increase awareness of the tool, provide education on its use, and obtain support from institutional leadership, and 4. Direct outreach to a pilot subset of PCCs whose patients had been detected using the validated FIND FH® machine learning algorithm. Participating clinicians were surveyed at baseline before the intervention and after the educational session. Results: 70 PCC consented to participate in the study with a survey completion rate of 79 % (n = 55) and 42 % (n = 23) for the baseline and follow-up surveys, respectively. Objective PCC knowledge scores improved from 40 to 65 % of responders correctly responding to at least 2/3rds of survey questions. Despite the fact that 87 % identified PCC's as most effective for early detection of FH, 100 % of PCCs who received direct outreach chose to defer care to an outpatient cardiologist over pursuing workup in the primary care setting. Conclusion: Empowering PCCs in management of FH serves as a key strategy in addressing this underdiagnosed and undertreated potentially life-threatening condition. A systems-based approach to addressing these aims may include leveraging EMR-based clinical decision support models and cross-disciplinary educational partnerships with medical specialists.

In recent years, adolescents have experienced rising rates of suicidal ideation. Though pediatricians in our clinic were screening for depression, suicide risk screens were not completed. We designed a quality improvement (QI) project with the goal to increase the rate of suicide risk screening by 20 %, and improve clinicians’ attitudes toward the electronic health record (EHR) for mental health screening by 20 %. Baseline analyses included chart review and a provider survey. Suicide risk screen tools were built with decision-support into the EHR. After provider education was completed, the EHR-based interventions were implemented. Post-intervention, suicide risk screening completion increased from 8.8 % to 95 %. Provider attitudes towards the EHR also improved. This QI project demonstrates that implementing EHR tools that are provider-informed can help to improve suicide risk screening rates, streamline workflows, and help improve clinicians’ attitudes towards the EHR.

Aim: Although the US FDA encourages manufacturers of medical devices to submit real-world evidence (RWE) to support regulatory decisions, the ability of real-world data (RWD) to generate evidence suitable for decision making remains unclear. The 2017 Medical Device User Fee Amendments (MDUFA IV), authorized the National Evaluation System for health Technology Coordinating Center (NESTcc) to conduct pilot projects, or ‘Test-Cases’, to assess whether current RWD captures the information needed to answer research questions proposed by industry stakeholders. We synthesized key lessons about the challenges conducting research with RWD and the strategies used by research teams to enhance their ability to generate evidence from RWD based on 18 Test-Cases conducted between 2020 and 2022. Materials & methods: We reviewed study protocols and reports from each Test-Case team and conducted 49 semi-structured interviews with representatives of participating organizations. Interview transcripts were coded and thematically analyzed. Results: Challenges that stakeholders encountered in workingwith RWD included the lack of unique device identifiers, capturing key data elements and their appropriate meaning in structured data, limited reliability of diagnosis and procedure codes in structured data, extracting information from unstructured electronic health record (EHR) data, limited capture of long-term study end points, missing data and data sharing. Successful strategies included using manufacturer and supply chain data, leveraging clinical registries and registry reporting processes to collect and aggregate data, querying standardized EHR data, implementing natural language processing algorithms and using multidisciplinary research teams. Conclusion: The Test-Cases identified numerous challenges working with RWD but also opportunities to address these challenges and improve researchers’ ability to use RWD to generate evidence on medical devices.

ObjectiveThis report describes our experience in electronic health record (EHR) note modification and creation of an external dashboard to create a local learning health system that contributes to quality improvement and patient care within our pediatric rheumatology clinic.MethodsWe applied quality improvement methodology to develop a more reliable and accurate system to identify patients with juvenile idiopathic arthritis and track important measures that aide in improving patient care and performance outcomes. From 2019 to 2021, we iteratively modified our outpatient clinic EHR note to include structured data elements to improve longitudinal monitoring. We then validated data transferred to an electronic dashboard external to the EHR and demonstrated utility for identifying an accurate patient population and tracking quality improvement initiatives.ResultsCreation of the structured data elements improved the identification of patients with JIA with >99% accuracy and without requiring manual review of the chart. Using the dashboard to monitor performance, we improved documentation of critical disease activity measures that resulted in improvement in those scores across the local population of patients with JIA. The structured data elements also enabled us to automate electronic data transfer to a multicenter learning network registry.ConclusionThe structured data element modifications made to our outpatient EHR note populate a local dashboard that allows real time access to critical information for patient care, population management, and improvement in quality metrics. The collection and monitoring of structured data can be scaled to other quality improvement initiatives in our clinic and shared with other centers.

Abstract INTRODUCTION Alzheimer's disease (AD) is often misclassified in electronic health records (EHRs) when relying solely on diagnosis codes. This study aimed to develop a more accurate, computable phenotype (CP) for identifying AD patients using structured and unstructured EHR data. METHODS We used EHRs from the University of Florida Health (UFHealth) system and created rule‐based CPs iteratively through manual chart reviews. The CPs were then validated using data from the University of Texas Health Science Center at Houston (UTHealth) and the University of Minnesota (UMN). RESULTS Our best‐performing CP was “patient has at least 2 AD diagnoses and AD‐related keywords in AD encounters,” with an F1‐score of 0.817 at UF, 0.961 at UTHealth, and 0.623 at UMN, respectively. DISCUSSION We developed and validated rule‐based CPs for AD identification with good performance, which will be crucial for studies that aim to use real‐world data like EHRs. Highlights Developed a computable phenotype (CP) to identify Alzheimer's disease (AD) patients using EHR data. Utilized both structured and unstructured EHR data to enhance CP accuracy. Achieved a high F1‐score of 0.817 at UFHealth, and 0.961 and 0.623 at UTHealth and UMN. Validated the CP across different demographics, ensuring robustness and fairness.

Background: Ascertaining accurately the exposure to antithrombotic medications for both research and quality initiatives has been challenging due to a multitude of reasons: aspirin, the most commonly used antithrombotic, is available over the counter in the United States. Additionally, antithrombotic medications are frequently interrupted for bleeding and procedures. Objectives: We aimed to develop and validate an algorithm to capture accurately the longitudinal exposure to antithrombotic medications including aspirin using the electronic health record. Methods: We used the Medical Inpatient Thrombosis and Hemostasis cohort, which consists of primary care patients at a university medical center followed for a median of 6.2 years. Exposure to antithrombotic medications was captured using the medication reconciliation data linked to each ambulatory encounter. We developed an algorithm that used the taking “yes” or “no” tab as well as start and stop dates to define the duration of exposure for each medication. Eighty charts were reviewed and compared with results of the algorithm for validation. We estimated the sensitivity, specificity, and positive and negative predictive values. Results: The algorithm was 97% (95% CI, 94%-100%) sensitive and 95% (95% CI, 90%-100%) specific in identifying exposure to any antithrombotic medication. This translated to a 93% (95% CI, 85%-100%) positive predictive value and 98% (95% CI, 96%-100%) negative predictive value. When looking at aspirin alone, the sensitivity and the positive predictive value were 95% (95% CI, 88%-100%) and 87% (95% CI, 71%-100%). Conclusion: This current algorithm provides a new and easily adaptable strategy to capture accurately exposure to aspirin and other antithrombotic medications.

Artificial Intelligence is extensively applied in heartcare to analyze patient data, detect anomalies, and provide personalized treatment recommendations, ultimately improving diagnosis and patient outcomes. In a field where accountability is indispensable, the prime reason why medical practitioners are still reluctant to utilize AI models, is the reliability of these models. However, explainable AI (XAI) was a game changing discovery where the so-called back boxes can be interpreted using Explainability algorithms. The proposed conceptual model reviews the existing recent researches for AI in heartcare that have found success in the past few years. The various techniques explored range from clinical history analysis, medical imaging to the nonlinear dynamic theory of chaos to metabolomics with specific focus on machine learning, deep learning and Explainability. The model also comprehensively surveys the different modalities of datasets used in heart disease prediction focusing on how results differ based on the different datasets along with the publicly available datasets for experimentation. The review will be an eye opener for medical researchers to quickly identify the current progress and to identify the most reliable data and AI algorithm that is appropriate for a particular technology for heartcare along with the Explainability algorithm suitable for the specific task.

ObjectivesElectronic health records (German: elektronische Patientenakte - ePA) are an important healthcare tool. However, in Germany, current participation remains low for their national ePA. To rectify this, the German government recently adopted an opt-out approach to their national ePA system. The objective of this study is to investigate and provide a brief overview of German public attitudes towards this approach to inform policymakers with evidence-based insights.MethodsFour public focus groups were conducted with 12 German citizens to discuss their opinions on the German governments new opt-out approach to the ePA.ResultsThree major thematic categories were identified (Contributors to Opt-Out Implementation, Barriers to Opt-Out Implementation, and Contingent Factors) to describe citizen views on the opt-out approach for the ePA.ConclusionThe public is generally supportive of an opt-out approach to ePAs in Germany, as they see the benefits ePAs can provide to German society; but they are skeptical on how successful this approach might be due to extant issues that policymakers must be aware of in order to successfully implement an opt-out approach for Germany’s national ePA system.

The availability of vast amounts of longitudinal data from electronic health records (EHRs) and personal wearable devices opens the door to numerous new research questions. In many studies, individual variability of a longitudinal outcome is as important as the mean. Blood pressure fluctuations, glycemic variations, and mood swings are prime examples where it is critical to identify factors that affect the within-individual variability. We propose a scalable method, within-subject variance estimator by robust regression (WiSER), for the estimation and inference of the effects of both time-varying and time-invariant predictors on within-subject variance. It is robust against the misspecification of the conditional distribution of responses or the distribution of random effects. It shows similar performance as the correctly specified likelihood methods but is 103 ∼ 105 times faster. The estimation algorithm scales linearly in the total number of observations, making it applicable to massive longitudinal data sets. The effectiveness of WiSER is evaluated in extensive simulation studies. Its broad applicability is illustrated using the accelerometry data from the Women's Health Study and a clinical trial for longitudinal diabetes care.

Post-acute sequelae of SARS-CoV-2 (PASC) is defined as persistent symptoms after apparent recovery from acute COVID-19 infection, also known as COVID-19 long-haul. We performed a retrospective review of electronic health records (EHR) from the University of California COvid Research Data Set (UC CORDS), a de-identified EHR of PCR-confirmed SARS-CoV-2-positive patients in California. The purposes were to (1) describe the prevalence of PASC, (2) describe COVID-19 symptoms and symptom clusters, and (3) identify risk factors for PASC. Data were subjected to non-negative matrix factorization to identify symptom clusters, and a predictive model of PASC was developed. PASC prevalence was 11% (277/2,153), and of these patients, 66% (183/277) were considered asymptomatic at days 0-30. Five PASC symptom clusters emerged and specific symptoms at days 0-30 were associated with PASC. Women were more likely than men to develop PASC, with all age groups and ethnicities represented. PASC is a public health priority.

Abstract E-health has become a top priority for healthcare organizations focused on advancing healthcare services. Thus, medical organizations have been widely adopting cloud services, resulting in the effective storage of sensitive data. To prevent privacy and security issues associated with the data, attribute-based encryption (ABE) has been a popular choice for encrypting private data. Likewise, the attribute-based access control (ABAC) technique has been widely adopted for controlling data access. Researchers have proposed electronic health record (EHR) systems using ABE techniques like ciphertext policy attribute-based encryption (CP-ABE), key policy attribute-based encryption (KP-ABE), and multi authority attribute-based encryption (MA-ABE). However, there is a lack of rigorous comparison among the various ABE schemes used in healthcare systems. To better understand the usability of ABE techniques in medical systems, we performed a comprehensive review and evaluation of the three popular ABE techniques by developing EHR systems using knowledge graphs with the same data but different encryption mechanisms. We have used the MIMIC-III dataset with varying record sizes for this study. This paper can help healthcare organizations or researchers using ABE in their systems to comprehend the correct usage scenario and the prospect of ABE deployment in the most recent technological evolution.

Abstract Background Electronic health records (EHRs) are digital records of individual health information. However, their adoption and utilization remain low. This study explores the factors influencing the implementation of EHRs through a questionnaire survey to enhance individual awareness and adoption of EHRs. Methods A questionnaire and an expert rating scale were developed sequentially, and the consistency of the scores from five experts was calculated using Kendall’s W to generate a final questionnaire. A non-parametric test was utilized to analyze differences in continuous data that did not follow a normal distribution. Categorical variables were expressed as percentages (%), the chi-square test was employed for group comparisons, and multiple logistic regression was implemented to assess individuals’ awareness and adoption of EHRs. Results In total, 1,341 survey questionnaires were distributed between January and December 2022, with 1,337 valid responses (99.7%). The results indicated that the proportion of participants who were aware of EHRs and had a bachelor’s degree or higher education, an income of ≥$700 per month, residence in urban areas, possessed self-care abilities, and underwent annual physical examinations was significantly higher than that without awareness of EHRs (P

Medical information systems could benefit from electronic health records management using openEHR. On the other hand, such a standard adds an additional software layer to the system, which might impact performance. In this article, we present an in-depth comparison of open-source openEHR servers and propose tools for testing them. Load tests for selected opensource servers were prepared using Apache JMeter. Statistics of elapsed time of requests and throughput of each solution were calculated. Results show that open-source openEHR servers significantly differ in performance and stability and prove that load testing should be a crucial part of a development process.

Background: Hypertension is frequently studied in epidemiological studies that have been conducted using retrospective observational data, either as an outcome or a variable. However, there are few validation studies investigating the accuracy of hypertension phenotyping algorithms in aggregated electronic health record (EHR) data. Methods: Utilizing a centralized repository of inpatient EHR data from Singapore for the period of 2019–2020, a new algorithm that incorporates both diagnostic codes and medication details (Diag+Med) was devised. This algorithm was intended to supplement and improve the diagnostic code-only model (Diag-Only) for the classification of hypertension. We computed various metrics (sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV)) to assess the algorithm’s effectiveness in identifying hypertension on 2813 chart-reviewed records. This pool was composed of two patient cohorts: a random sampling of all inpatient admissions (Random Cohort) and a targeted group with atrial fibrillation diagnoses (AF Cohort). Results: The Diag+Med algorithm was more sensitive at detecting hypertension patients in both cohorts compared to the Diag-Only algorithm (83.8 and 87.6% vs. 68.2 and 66.5% in the Random and AF Cohorts, respectively). These improvements in sensitivity came at minimal costs in terms of PPV reductions (88.2 and 90.3% vs. 91.4 and 94.2%, respectively). Conclusion: The combined use of diagnosis codes and specific antihypertension medication exposure patterns facilitates a more accurate capture of patients with hypertension in a database of aggregated EHRs from diverse healthcare institutions in Singapore. The results presented here allow for the bias correction of risk estimates derived from observational studies involving hypertension.

Abstract Background and objective Clinical trials are of high importance for medical progress. This study conducted a systematic review to identify the applications of EHRs in supporting and enhancing clinical trials. Materials and methods A systematic search of PubMed was conducted on 12/3/2023 to identify relevant studies on the use of EHRs in clinical trials. Studies were included if they (1) were full-text journal articles, (2) were written in English, (3) examined applications of EHR data to support clinical trial processes (e.g. recruitment, screening, data collection). A standardized form was used by two reviewers to extract data on: study design, EHR-enabled process(es), related outcomes, and limitations. Results Following full-text review, 19 studies met the predefined eligibility criteria and were included. Overall, included studies consistently demonstrated that EHR data integration improves clinical trial feasibility and efficiency in recruitment, screening, data collection, and trial design. Conclusions According to the results of the present study, the use of Electronic Health Records in conducting clinical trials is very helpful. Therefore, it is better for researchers to use EHR in their studies for easy access to more accurate and comprehensive data. EHRs collects all individual data, including demographic, clinical, diagnostic, and therapeutic data. Moreover, all data is available seamlessly in EHR. In future studies, it is better to consider the cost-effectiveness of using EHR in clinical trials.

Abstract The encryption of user data is crucial when employing electronic health record services to guarantee the security of the data stored on cloud servers. Attribute-based encryption (ABE) scheme is considered a powerful encryption technique that offers flexible and fine-grained access control capabilities. Further, the multi-user collaborative access ABE scheme additionally supports users to acquire access authorization through collaborative works. However, the existing multi-user collaborative access ABE schemes do not consider the different weights of collaboration users. Therefore, using these schemes for weighted multi-user collaborative access results in redundant attributes, which inevitably reduces the efficiency of the ABE scheme. This paper proposes a revocable and verifiable weighted attribute-based encryption with collaborative access scheme (RVWABE-CA), which can provide efficient weighted multi-user collaborative access, user revocation, and data integrity verification, as the fundamental cornerstone for establishing a robust framework to facilitate secure sharing of electronic health records in a public cloud environment. In detail, this scheme employs a novel weighted access tree to eliminate redundant attributes, utilizes encryption version information to control user revocation, and establishes Merkle Hash Tree for data integrity verification. We prove that our scheme is resistant against chosen plaintext attack. The experimental results demonstrate that our scheme has significant computational efficiency advantages compared to related works, without increasing storage or communication overhead. Therefore, the RVWABE-CA scheme can provide an efficient and flexible weighted collaborative access control and user revocation mechanism as well as data integrity verification for electronic health record systems.

Abstract Background An increasing number of studies have described new and persistent symptoms and conditions as potential post-acute sequelae of SARS-CoV-2 infection (PASC). However, it remains unclear whether certain symptoms or conditions occur more frequently among persons with SARS-CoV-2 infection compared with those never infected with SARS-CoV-2. We compared the occurrence of specific COVID-associated symptoms and conditions as potential PASC 31- to 150-day following a SARS-CoV-2 test among adults and children with positive and negative test results. Methods We conducted a retrospective cohort study using electronic health record (EHR) data from 43 PCORnet sites participating in a national COVID-19 surveillance program. This study included 3,091,580 adults (316,249 SARS-CoV-2 positive; 2,775,331 negative) and 675,643 children (62,131 positive; 613,512 negative) who had a SARS-CoV-2 laboratory test during March 1, 2020–May 31, 2021 documented in their EHR. We used logistic regression to calculate the odds of having a symptom and Cox models to calculate the risk of having a newly diagnosed condition associated with a SARS-CoV-2 positive test. Results After adjustment for baseline covariates, hospitalized adults and children with a positive test had increased odds of being diagnosed with ≥ 1 symptom (adults: adjusted odds ratio[aOR], 1.17[95% CI, 1.11–1.23]; children: aOR, 1.18[95% CI, 1.08–1.28]) or shortness of breath (adults: aOR, 1.50[95% CI, 1.38–1.63]; children: aOR, 1.40[95% CI, 1.15–1.70]) 31–150 days following a SARS-CoV-2 test compared with hospitalized individuals with a negative test. Hospitalized adults with a positive test also had increased odds of being diagnosed with ≥ 3 symptoms or fatigue compared with those testing negative. The risks of being newly diagnosed with type 1 or type 2 diabetes (adjusted hazard ratio[aHR], 1.25[95% CI, 1.17–1.33]), hematologic disorders (aHR, 1.19[95% CI, 1.11–1.28]), or respiratory disease (aHR, 1.44[95% CI, 1.30–1.60]) were higher among hospitalized adults with a positive test compared with those with a negative test. Non-hospitalized adults with a positive test also had higher odds or increased risk of being diagnosed with certain symptoms or conditions. Conclusions Patients with SARS-CoV-2 infection, especially those who were hospitalized, were at higher risk of being diagnosed with certain symptoms and conditions after acute infection.

Abstract Background Rare diseases affect approximately 400 million people worldwide. Many of them suffer from delayed diagnosis. Among them, NPHP1-related renal ciliopathies need to be diagnosed as early as possible as potential treatments have been recently investigated with promising results. Our objective was to develop a supervised machine learning pipeline for the detection of NPHP1 ciliopathy patients from a large number of nephrology patients using electronic health records (EHRs). Methods and results We designed a pipeline combining a phenotyping module re-using unstructured EHR data, a semantic similarity module to address the phenotype dependence, a feature selection step to deal with high dimensionality, an undersampling step to address the class imbalance, and a classification step with multiple train-test split for the small number of rare cases. The pipeline was applied to thirty NPHP1 patients and 7231 controls and achieved good performances (sensitivity 86% with specificity 90%). A qualitative review of the EHRs of 40 misclassified controls showed that 25% had phenotypes belonging to the ciliopathy spectrum, which demonstrates the ability of our system to detect patients with similar conditions. Conclusions Our pipeline reached very encouraging performance scores for pre-diagnosing ciliopathy patients. The identified patients could then undergo genetic testing. The same data-driven approach can be adapted to other rare diseases facing underdiagnosis challenges.

In the context of patients with complex conditions, the capacity to assist physicians in making appropriate prescribing decisions is of paramount importance. The field of prescription recommendation has attracted a growing body of research interest, given its significant clinical value. However, existing research has not taken into account the patients’ intention to visit the doctor and the similarities between patients’ visit intentions. In this study, we employ medical order data in both prescription intention modeling and patient similarity analysis, subsequently integrating these to generate more personalized and accurate prescription recommendations. The cross-attention mechanism is first employed to focus the representation of the patient’s health status on the patient’s chief complaints, with the objective of predicting treatment intention and, subsequently, the prescriptions. Subsequently, when we learn from the successful treatment experiences of similar patients, we focus more on the similarities of their chief complaints. The experimental results on the MIMIC-III dataset demonstrate that our method achieves optimal performance in terms of Jaccard score, PRAUC, F1 score and average precision, with relative improvements of 7%, 5%, 4%, and 8% respectively, in comparison to state-of-the-art approaches.

Early identification of patients’ clinical status plays a critical role in intensive care unit (ICU) care. The increased adoption of electronic health records (EHRs) in the ICU creates prospects for deep learning (DL) application systems in this discipline. However, monitoring and prediction systems in the ICU encounter problems with security, alarm errors, and interpretation. This research presents deep contrastive federated learning (Deep-CFL), an approach that leverages explainable AI (XAI), CFL, and imbalanced supervised learning techniques to address these problems. CFL introduces an innovative approach to minimize the difference in local and global model prediction ability while increasing the gap in prediction performance of the current local model and its previous model in a communication round. When paired with imbalanced learning, this strategy substantially mitigates error alarm problems while ensuring data security. The XAI technique, specifically integrated gradient, is employed to refine the DL-based model architecture to enhance system interpretability. Extensive experiments and in-depth analyses across three significant clinical datasets highlight the superior performance of Deep-CFL over local and centralized learning-based approaches. The results involving $25,329$ patients admitted to Chonnam National University Hospital reveal that Deep-CFL, with an area under the receiver operating characteristics curve of 0.879, an area under the precision-recall curve of 0.886, and an average precision of 0.884, surpasses systems based on centralized learning while reducing the late alarm rate by up to 10.3%.

Background and Aims: Gastric cancer (GC) is a leading cause of cancer incidence and mortality globally. Population screening is limited by the low incidence and prevalence of GC in the United States. A risk prediction algorithm to identify high-risk patients allows for targeted GC screening. We aimed to determine the feasibility and performance of a logistic regression model based on electronic health records to identify individuals at high risk for noncardia gastric cancer (NCGC). Methods: We included 614 patients who had a diagnosis of NCGC between ages 40 and 80 years and who were seen at our large tertiary medical center in multiple states between 2010 and 2021. Controls without a diagnosis of NCGC were randomly selected in a 1:10 ratio of cases to controls. Multiple imputation by chained equations for missing data followed by logistic regression on imputed datasets was used to estimate the probability of NCGC. Area under the curve and the 0.632 estimator was used as the estimate for discrimination. Results: The 0.632 estimator value was 0.731, indicating robust model performance. Probability of NCGC was higher with increasing age (odds ratio [OR] = 1.16, 95% confidence interval [CI]: 1.04–1.3), male sex (OR = 1.97; 95% CI: 1.64–2.36), Black (OR = 3.07; 95% CI: 2.46–3.83) or Asian race (OR = 4.39; 95% CI: 2.60–7.42), tobacco use (OR = 1.61; 95% CI: 1.34–1.94), anemia (OR = 1.35; 95% CI: 1.09–1.68), and pernicious anemia (OR = 6.12, 95% CI: 3.42–10.95). Conclusion: We demonstrate the feasibility and good performance of an electronic health record–based logistic regression model for estimating the probability of NCGC. Future studies will refine and validate this model, ultimately identifying a high-risk cohort who could be eligible for NCGC screening.

Abstract Background Researchers often use publicly available data sources to describe injuries occurring in professional athletes, developing and testing hypotheses regarding athletic-related injury. It is reasonable to question whether publicly available data sources accurately indicate athletic-related injuries resulting from professional sport participation. We compared sport-related concussion (SRC) clinical incidence using data from publicly available sources to a recent publication reporting SRC using electronic health records (EHR) from the National Football League (NFL). We hypothesize publicly available data sources will underrepresent SRC in the NFL. We obtained SRCs reported from two publicly available data sources (NFL.com, pro-football-reference.com) and data reported from the NFL’s published EHR. We computed SRC per 100 unique player signings from 2015–2019 and compared the clinical incidence from publicly available data sources to EHR rates using clinical incidence ratios (CIR) and 95% confidence intervals (CI). Findings From 2015–2019, SRC counts from published EHR record data ranged from 135–192 during the regular season, whereas SRC counts ranged from 102–194 and 69–202 depending on the publicly available data source. In NFL.com the SRC clinical incidence was significantly and progressively lower in 2017 (CIR: 0.73, 95% CI: 0.58–0.91), 2018 (CIR: 0.66, 95% CI: 0.50–0.87), and 2019 (CIR: 0.48, 95% CI: 0.35–0.64) relative to the gold-standard EHR. In the pro-football-reference.com data, the documented SRCs in publicly available data sources for other years were ~ 20–30% lower than the gold-standard EHR numbers (CIRs 0.70–0.81). Conclusions Publicly available data for SRCs per 100 unique player signings did not match published data from the NFL’s EHR and in several years were significantly lower. Researchers should use caution before using publicly available data sources for injury research.