Integrating Deep Learning with Electronic Health Records for Early Glaucoma Detection: A Multi-Dimensional Machine Learning Approach.

Multicenter research of geriatric emergency department (GED) care remains limited. Our objectives were to: 1. Prospectively collect data prioritized by the Geriatric Emergency care Applied Research (GEAR) network, a transdisciplinary taskforce for GED care, and create a multicenter GED research repository of prospective and electronic health record (EHR) data, 2. Assess concordance between prospective and EHR data. The GEAR Standardization Study (GEARSS) is a multicenter, prospective study of older emergency department (ED) patients (65+) focusing on the 4Ms of age-friendly care (mobility, medication safety, mentation, what matters) and elder mistreatment. Demographic and clinical data were collected via interviews by trained research assistants (RA) on Days 0, 4, 30, and 90 and linked to EHR. Prevalence of chronic comorbidities and incident delirium were measured and reported using descriptive statistics. Prospective and EHR data concordance was assessed with Cohen's Kappa. 999 participants were recruited from 5 EDs (3/25/2021-6/30/2022) across 3 institutions: Grady Health System, Northwestern Memorial Hospital, and Yale New Haven Health. The cohort was 57.0% female, 55.2% White, 39.1% Black, and 3.4% Hispanic, and the mean age was 75.1 years. For rheumatologic disease, peptic ulcer disease, diabetes, renal disease, and cancer, prevalence differed between prospective and EHR data by > 10%. About two-thirds of participants were at risk for falls. Concordance between prospective and EHR data was good for ethnicity (K = 0.73); excellent for sex (K = 1.00), age (K = 1.00), and race (K = 0.98); fair for disposition (K = 0.53); slight for ED observation status (K = 0.33) and dementia diagnosis (K = 0.24); poor for delirium presence (K = 0.07). In GEARSS, demographic variables aligned strongly between prospective and EHR data, while diagnosis, disposition, and mentation factors did not. This multicenter data source provides preliminary findings for common geriatric syndromes and conditions. Choice of measures using these data should be driven by GED research questions.

IntroductionThe ability to utilize the electronic health record (EHR) without compromising the doctor-patient relationship (DPR) is an essential skill of all physicians and trainees, yet little time is spent on educating or assessing learners on needed techniques. To address this gap, we developed a conventional OSCE station coupled with a simulated patient chart within the Epic program in order to assess our students' skills utilizing the EHR during a patient encounter.MethodsOf third-year medical students, 119 were given full access to the patient's simulated chart 24 hours in advance of their OSCE to review clinical data. During an in-person OSCE with a standardized patient (SP), students performed a focused history and physical, using the EHR to verify allergies and medications. Students completed an electronic patient note graded by faculty. SPs evaluated the students on communication and interpersonal skills with specific rubric elements. Faculty graded the students' notes to evaluate their expression of clinical reasoning in the assessment and plan.ResultsTraining SPs and faculty to assess students on EHR skills was feasible. After implementation of a comprehensive curriculum focused on EHR and DPR, there was a significant difference on EHR-related communication skills (M = 76.4, SD = 17.6) versus (M = 37, SD = 28.9) before curriculum enhancement t (117.9) = −12.4, p <.001.DiscussionThe EHR OSCE station provided a standardized method of assessing students' EHR skills during a patient encounter. Challenges still exist in the technological requirements to develop and deliver cases in today's EHR platform.

Objective: To evaluate the continuity and completeness of electronic health record (EHR) data, and the concordance of select clinical outcomes and baseline comorbidities between EHR and linked claims data, from three healthcare delivery systems in Taiwan. Methods: We identified oral hypoglycemic agent (OHA) users from the Integrated Medical Database of National Taiwan University Hospital (NTUH-iMD), which was linked to the National Health Insurance Research Database (NHIRD), from June 2011 to December 2016. A secondary evaluation involved two additional EHR databases. We created consecutive 90-day periods before and after the first recorded OHA prescription and defined patients as having continuous EHR data if there was at least one encounter or prescription in a 90-day interval. EHR data completeness was measured by dividing the number of encounters in the NTUH-iMD by the number of encounters in the NHIRD. We assessed the concordance between EHR and claims data on three clinical outcomes (cardiovascular events, nephropathy-related events, and heart failure admission). We used individual comorbidities that comprised the Charlson comorbidity index to examine the concordance of select baseline comorbidities between EHRs and claims. Results: We identified 39,268 OHA users in the NTUH-iMD. Thirty-one percent (n = 12,296) of these users contributed to the analysis that examined data continuity during the 6-month baseline and 24-month follow-up period; 31% (n = 3,845) of the 12,296 users had continuous data during this 30-month period and EHR data completeness was 52%. The concordance of major cardiovascular events, nephropathy-related events, and heart failure admission was moderate, with the NTU-iMD capturing 49–55% of the outcome events recorded in the NHIRD. The concordance of comorbidities was considerably different between the NTUH-iMD and NHIRD, with an absolute standardized difference >0.1 for most comorbidities examined. Across the three EHR databases studied, 29–55% of the OHA users had continuous records during the 6-month baseline and 24-month follow-up period. Conclusion: EHR data continuity and data completeness may be suboptimal. A thorough evaluation of data continuity and completeness is recommended before conducting clinical and translational research using EHR data in Taiwan.

Healthcare data management has advanced with Electronic Health Records (EHRs), enhancing the efficiency of medical procedures. Machine learning applied to EHRs transitions healthcare from reactive to proactive, supporting the cost-efficiency and sustainability goals of smart cities. However, digitizing medical records introduces security risks, especially from internal threats, necessitating strong detection systems. Research into machine learning techniques, such as decision trees, random forests, and support vector machines (SVMs), shows their effectiveness in detecting EHR breaches. Balancing system usability with patient privacy remains a key challenge amid widespread data sharing. This study highlights SVMs and deep learning models as promising for improving EHR data accuracy, enhancing detection efficiency, and supporting clinical decisions. Despite advancements in AI, deep learning continues to play a crucial role in refining clinical decision systems, including translating EHR data using technologies like natural language processing (NLP). The study provides a qualitative analysis of how deep learning can optimize EHR processes while addressing security and functional challenges.

BackgroundGlaucoma is a leading cause of irreversible blindness worldwide. Predicting a patient’s future clinical trajectory would help physicians personalize management. We present a novel approach for predicting patient visual field (VF) progression by combining Functional Principal Component Analysis (FPCA) with electronic health record (EHR) data.MethodsWe identified glaucoma patients using diagnosis codes who had >=3 VF tests. We developed a 2-stage modeling pipeline: 1) Patients were split 80:10:10 into train, validation, and test sets and classified as fast-progressors or slow-progressors. 2) FPCA was used to predict mean deviation (MD) trajectories over 10 years after the baseline year of VF exams, using the first 2 principal components. To make predictions, the model uses up to one year of baseline VF and EHR data as input, but can flexibly make predictions from as few as a single VF test.Results15,764 VF tests belonging to 2,372 patients were included, of which 8.92% of eyes were fast progressors. On the held-out test set, predictions over 10 years of future MD trajectories using baseline VF and EHR clinical data yielded an R2 of 0.646 and an RMSE of 3.67 for fast-progressors, and an R2 of 0.728 and an RMSE of 3.09 for slow-progressors. Performance was higher when predicting over the near term (fast progressors: year 1 R2 0.920, RMSE 1.83; year 5 R2 0.515, RMSE 4.26; slow progressors: year 1 R2 0.918, RMSE 1.771; year 5 R2 0.717, RMSE 3.12).ConclusionA novel modeling approach combining FPCA with clinical data from EHR was able to model longitudinal clinical trajectories of glaucoma patients. This method is well-suited for modeling longitudinal healthcare data, handling sparse and irregular observation schedules with a varying number of inputs.

While deep learning models have achieved remarkable performance, their adoption in healthcare faces a critical challenge due to a lack of interpretability. Interpretability is a serious issue in high-stakes environments like Intensive Care Units (ICUs), where transparency in decision-making is not just mandatory but also essential. Several studies have ascertained the trade-off between performance and interpretability, with interpretability being sacrificed for performance and vice versa. Therefore, this study aims to demonstrate that performance and interpretability need not be mutually exclusive by proposing a hybrid framework that integrates LSTM, a deep learning architecture, with explainable models such as SHAP for ICU mortality prediction using Electronic Health Record (EHR) data. The study employs publicly available ICU datasets such as MIMIC-III or MIMIC-IV, which contain comprehensive EHR data for ICU patients. The LSTM achieved an accuracy of 98.6% and a recall of 87.5% on unseen data, but recorded low Precision, indicating that the model was biased toward the majority class (No Mortality). When the LSTM results were compared with baseline models (Random Forest and Logistic Regression), it generally outperformed the baseline models. The major limitation is the presence of class imbalance within the dataset, as shown by the precision. Despite this, the LSTM model successfully maintained interpretability through SHAP without compromising predictive performance, thereby achieving a balance between accuracy, transparency, and clinical relevance.

Background: It is known that oral diseases such as periodontal (gum) disease are closely linked to various systemic diseases and disorders. Deep learning advances have the potential to make major contributions to healthcare, particularly in the domains that rely on medical imaging. Incorporating non-imaging information based on clinical and laboratory data may allow clinicians to make more comprehensive and accurate decisions. Methods: Here, we developed a multimodal deep learning method to predict systemic diseases and disorders from oral health conditions. A dual-loss autoencoder was used in the first phase to extract periodontal disease-related features from 1188 panoramic radiographs. Then, in the second phase, we fused the image features with the demographic data and clinical information taken from electronic health records (EHR) to predict systemic diseases. We used receiver operation characteristics (ROC) and accuracy to evaluate our model. The model was further validated by an unseen test dataset. Findings: According to our findings, the top three most accurately predicted chapters, in order, are the Chapters III, VI and IX. The results indicated that the proposed model could predict systemic diseases belonging to Chapters III, VI and IX, with AUC values of 0.92 (95% CI, 0.90-94), 0.87 (95% CI, 0.84-89) and 0.78 (95% CI, 0.75-81), respectively. To assess the robustness of the models, we performed the evaluation on the unseen test dataset for these chapters and the results showed an accuracy of 0.88, 0.82 and 0.72 for Chapters III, VI and IX, respectively. Interpretation: The present study shows that the combination of panoramic radiograph and clinical oral features could be considered to train a fusion deep learning model for predicting systemic diseases and disorders.

Introduction: Arthritis is one of the most common chronic diseases. Early detection of arthritis and its progression can facilitate early intervention measures, lowering disease severity in patients. As electronic health records (EHR) become more accessible, this study assesses whether general health information and arthritis-related questionnaires can be used in arthritis diagnosis, without the involvement of costly imaging methods. Therefore, we created deep learning (DL) and machine learning (ML) models to explore the feasibility of combining EHR and modern computational tools to diagnose arthritis. Methods: A total of 782 arthritis patients and 4014 control patients were identified from the Osteoarthritis Initiative (OAI) – a ten-year-long observational study that included patient EHR in five time points. Six hundred variables were filtered by random forest classifier followed by manual filtering. Data were split properly to training, testing and validation set, and the training set was balanced. Sequential, nonsequential DL models, and five independent DL models for each time points were used. The accuracy, positive prevalence value (PPV), negative prevalence value (NPV), and area under curve (AUC), were assessed and compared with four classical ML models. SHAP (SHapley Additive exPlanations) summary analysis was also conducted. Results: Sequential and non-sequential deep learning models showed accuracies of ~ 0.97, and the four classical machine learning approaches showed accuracies of above 0.9. High positive and negative predicted values (&gt; 0.90) for all of the models suggested the potential clinical applicability of the model, while the SHAP analysis demonstrated its interpretability. Discussion: We tested various models and showed the ability to use machine learning methods for early diagnosis of arthritis with EHR. The models can be used as a screening tool to select susceptible patients for confirmatory tests such as X-ray and MRI. Identification of early disease states could facilitate protective measures that slow disease progression.

BackgroundMedical referral is the transfer of a patient’s care from one physician to another upon request. This process involves multiple steps that require provider-to-provider and provider-to-patient communication. In Taiwan, the National Health Insurance Administration (NHIA) has implemented a national medical referral (NMR) system, which encourages physicians to refer their patients to different health care facilities to reduce unnecessary hospital visits and the financial stress on the national health insurance. However, the NHIA’s NMR system is a government-based electronic medical referral service, and its referral data access and exchange are limited to authorized clinical professionals using their national health smart cards over the NHIA virtual private network. Therefore, this system lacks scalability and flexibility and cannot establish trusting relationships among patients, family doctors, and specialists.ObjectiveTo eliminate the existing restrictions of the NHIA’s NMR system, this study developed a scalable, flexible, and blockchain-enabled framework that leverages the NHIA’s NMR referral data to build an alliance-based medical referral service connecting health care facilities.MethodsWe developed a blockchain-enabled framework that can integrate patient referral data from the NHIA’s NMR system with electronic medical record (EMR) and electronic health record (EHR) data of hospitals and community-based clinics to establish an alliance-based medical referral service serving patients, clinics, and hospitals and improve the trust in relationships and transaction security. We also developed a blockchain-enabled personal health record decentralized app (DApp) based on our blockchain-enabled framework for patients to acquire their EMR and EHR data; DApp access logs were collected to assess patients’ behavior and investigate the acceptance of our personal authorization-controlled framework.ResultsThe constructed iWellChain Framework was installed in an affiliated teaching hospital and four collaborative clinics. The framework renders all medical referral processes automatic and paperless and facilitates efficient NHIA reimbursements. In addition, the blockchain-enabled iWellChain DApp was distributed for patients to access and control their EMR and EHR data. Analysis of 3 months (September to December 2018) of access logs revealed that patients were highly interested in acquiring health data, especially those of laboratory test reports.ConclusionsThis study is a pioneer of blockchain applications for medical referral services, and the constructed framework and DApp have been applied practically in clinical settings. The iWellChain Framework has the scalability to deploy a blockchain environment effectively for health care facilities; the iWellChain DApp has potential for use with more patient-centered applications to collaborate with the industry and facilitate its adoption.

Clinical encounter data are heterogeneous and vary greatly from institution to institution. These problems of variance affect interpretability and usability of clinical encounter data for analysis. These problems are magnified when multisite electronic health record (EHR) data are networked together. This article presents a novel, generalizable method for resolving encounter heterogeneity for analysis by combining related atomic encounters into composite "macrovisits." Encounters were composed of data from 75 partner sites harmonized to a common data model as part of the NIH Researching COVID to Enhance Recovery Initiative, a project of the National Covid Cohort Collaborative. Summary statistics were computed for overall and site-level data to assess issues and identify modifications. Two algorithms were developed to refine atomic encounters into cleaner, analyzable longitudinal clinical visits. Atomic inpatient encounters data were found to be widely disparate between sites in terms of length-of-stay (LOS) and numbers of OMOP CDM measurements per encounter. After aggregating encounters to macrovisits, LOS and measurement variance decreased. A subsequent algorithm to identify hospitalized macrovisits further reduced data variability. Encounters are a complex and heterogeneous component of EHR data and native data issues are not addressed by existing methods. These types of complex and poorly studied issues contribute to the difficulty of deriving value from EHR data, and these types of foundational, large-scale explorations, and developments are necessary to realize the full potential of modern real-world data. This article presents method developments to manipulate and resolve EHR encounter data issues in a generalizable way as a foundation for future research and analysis.

Depression and anxiety are prevalent mental health conditions among individuals with type 2 diabetes mellitus (T2DM), who exhibit unique vulnerabilities and etiologies. However, existing approaches fail to fully utilize regional heterogeneous electronic health record (EHR) data. Integrating this data can provide a more comprehensive understanding of depression and anxiety in T2DM patients, leading to more personalized treatment strategies. This study aims to develop and validate a deep learning model, the Regional EHR for Depression and Anxiety Prediction Model (REDAPM), using regional EHR data to predict depression and anxiety in patients with T2DM. A case-control development and validation study was conducted using regional EHR data from the Nanjing Health Information Center (NHIC). Two retrospective, matched (1:3) datasets were constructed from the full cohort for the model's internal and external validation. These two datasets were selected from the NHIC data of 2020 and 2022, respectively. The REDAPM incorporates both structured and unstructured EHR data, capturing the temporal dependency of clinical events. The performance of REDAPM was compared to a set of baseline models, evaluated using the area under the receiver operating characteristic curve (ROC-AUC) and the area under the precision-recall curve (PR-AUC). Subgroup, ablation, and interpretation analyses were conducted to identify relevant clinical features available from EHRs. The internal and external validation datasets comprised 24,724 and 34,340 patients, respectively. The REDAPM outperformed baseline models in both datasets, achieving ROC-AUC scores of 0.9029±0.008 and 0.7360±0.005, and PR-AUC scores of 0.8124±0.011 and 0.5504±0.009. Ablation and subgroup experiments confirmed the significant contribution of patients' medical history text to the model's performance. Integrated gradient score analysis identified the predictive importance of other mental disorders. The REDAPM effectively leverages the heterogeneous characteristics of regional EHR data, demonstrating strong predictive performance for depression onset in diabetic patients. It also shows potential for discovering significant clinical features, indicating considerable promise for clinical utility.

The past decade has seen a tremendous advancement in using Electronic Health Records (EHRs) to offer clinical decision support and provide personalized healthcare to patients. Despite the potential benefits offered by EHR data, it is challenging to represent and analyze large EHRs for predictive modeling due to heterogeneity, high dimensionality, and sparsity. This article proposes a novel supervised Deep Similarity Learning approach that learns the patient representations and also finds the relationship between the patients using pairwise similarity learning to facilitate predictive analysis for personalized healthcare. We develop CNN_Softmax which is a Siamese-based neural network for multi-class classification methods corresponding to the prediction of disease. It uses Convolutional Neural Network (CNN) to study the vector representation of raw EHRs and capture essential information of patient features, and a Softmax-based supervised classification method that learns the similarity between pairs of patients and performs disease prediction using this similarity information. Our approach uses data type mapping to handle heterogeneity and the polynomial interpolation method to handle sparsity existing in EHR data. ORBDA, which is an openEHR (standard) benchmark dataset, is used for evaluating this study. Experimental results show that CNN_Softmax achieves an accuracy of 97.8%, a recall of 98.1%, a precision of 96.02%, and an F1 score of 97.82%. The comparative results show that our proposed novel methodology performs disease prediction with highly promising results and outperforms state-of-the-art similarity learning methods. The current study is the first attempt to perform disease prediction on standardized EHRs, to the best of the authors' knowledge. The deep similarity learning approach provides support for clinical decision making that is more reliable and generalizable than previous approaches and focuses on dealing with heterogeneous and sparse data. The concept also serves as a new implementation of artificial intelligence technologies for the application of clinical big data.

Electronic health record (EHR) data are widely used in public health research, including in HIV-related studies, but are limited by potential bias due to incomplete and inaccurate information, lack of generalizability, and lack of representativeness. This study explores how workflow processes among HIV health care providers (HCPs), data scientists, and state health department professionals may potentially introduce or minimize bias within EHR data. One focus group with 3 health department professionals working in HIV surveillance and 16 in-depth interviews (ie, 5 people with HIV, 5 HCPs, 5 data scientists, and 1 health department professional providing retention-in-care services) were conducted with participants purposively sampled in South Carolina from August 2023 to April 2024. All interviews were transcribed verbatim and analyzed using a constructivist grounded theory approach, where transcripts were first coded and then focused, axial, and theoretically coded. The EHR data lifecycle originates with people with HIV and HCPs in the clinical setting. Data scientists then curate EHR data and health department professionals manage and use the data for surveillance and policy decision-making. Throughout this lifecycle, the three primary stakeholders (ie, HCPs, data scientists, and health department professionals) identified challenges with EHR processes and provided their recommendations and accommodations in addressing the related challenges. HCPs reported the influence of socio-structural biases on their inquiry, interpretation, and documentation of social determinants of health (SDOH) information of people living with HIV, the influence of which is proposed to be mitigated through people living with HIV access to their EHRs. Data scientists identified limited data availability and representativeness as biasing the data they manage. Health department professionals face challenges with delayed and incomplete data, which may be addressed statistically but require consideration of the data's limitations. Overall, bias within the EHR data lifecycle persists because workflows are not intentionally structured to minimize bias and there is a diffusion of responsibility for data quality between the various stakeholders. From the perspective of various stakeholders, this study describes the EHR data lifecycle and its associated challenges as well as stakeholders' accommodations and recommendations for mitigating and eliminating bias in EHR data. Based upon these findings, studies reliant on EHR data should adequately consider its challenges and limitations. Throughout the EHR data lifecycle, bias could be reduced through an inclusive, supportive health care environment, people living with HIV verification of SDOH information, the customization of data collection systems, and EHR data inspection for completeness, accuracy, and timeliness. Future research is needed to further identify instances where bias is introduced and how it can best be mitigated and eliminated across the EHR data lifecycle. Systematic changes are necessary to reduce instances of bias between data workflows and stakeholders.

In the 2015 Medicare Access and CHIP Reauthorization Act, the Centers for Medicare and Medicaid Services (CMS) began the transition to a value-based payment system through implementation of the Quality Payment Program (QPP) ([1][1]). Representing a paradigm shift away from traditional volume-based

Despite the extensive use of real-world data for retrospective, observational clinical research, our understanding of how real-world data might increase the efficiency of data collection in patient-level randomized clinical trials is largely unknown. The structure of real-world data is inherently heterogeneous, with each source electronic health record and claims database different from the next. Their fitness-for-use as data sources for multisite trials in the United States has not been established. For a subset of participants in the HARMONY Outcomes Trial, we obtained electronic health record data from recruiting sites or Medicare claims data from the Centers for Medicare & Medicaid Services. For baseline characteristics and follow-up events, we assessed the level of agreement between these real-world data and data documented in the trial database. Real-world data-derived demographic information tended to agree with trial-reported demographic information, although real-world data were less accurate in identifying medical history. The ability of real-world data to identify baseline medication usage differed by real-world data source, with claims data demonstrating substantially better performance than electronic health record data. The limited number of lab results in the collected electronic health record data matched closely with values in the trial database. There were not enough follow-up events in the ancillary study population to draw meaningful conclusions about the performance of real-world data for identification of events. Based on the conduct of this ancillary study, the challenges and opportunities of using real-world data within clinical trials are discussed. Based on a subset of participants from the HARMONY Outcomes Trial, our results suggest that electronic health record or claims data, as currently available, are unlikely to be a complete substitute for trial data collection of medical history or baseline lab results, but that Medicare claims were able to identify most medications. The limited size of the study population prevents us from drawing strong conclusions based on these results, and other studies are clearly needed to confirm or refute these findings.