Comprehensive Electronic Health Record Research Articles

The digital health landscape at children’s hospitals in Canada

ABSTRACT Objectives Canadian hospitals have historically lagged behind peer nations in terms of adoption of digital health tools. The aim of this study was to assess the current state of adoption of digital health tools at children’s hospitals in Canada. Methods We conducted an online survey of Canadian pediatric tertiary-care hospitals between January and July 2023. The 35-item questionnaire was administered in English and French. Hospital characteristics, informatics infrastructure data and electronic health record (EHR) functionality data were summarized using descriptive statistics. Results The survey was completed by 15 of the 17 (88.2%) pediatric hospitals in Canada. All institutions had an EHR with 10 (66.6%) being fully digitized and five (33.3%) being partially digitized. Funding and availability of clinicians with expertise in clinical informatics were cited as barriers to implementing digital health tools. The availability of core EHR functionalities ranged from 53.3% for medication to 100% for the presence of a laboratory and radiology information system. Only five (33.3%) institutions reported that they had a patient portal. Discussion While all hospitals in this study had an EHR, functionalities varied greatly between centers. Canada lags behind the United States in terms of adoption of digital health tools such as patient portals likely due to governmental mandates and incentives, and a focus on cultivating a physician informatics workforce. Conclusion Canadian pediatric hospitals lag behind peer institutions in digital maturity. This study highlights the perceived need from hospitals for policies, funding, and resources from the Canadian government to help increase the uptake of comprehensive EHRs.

Who transitions to bipolar disorder? A comparison of major depressive disorder, anxiety, and ADHD

BackgroundDiagnostic delays in Bipolar Disorder (BD) are common and can contribute to worse outcomes. While most studies focus on depression as a primary precursor, both anxiety and attention deficit disorders are also frequent initial diagnoses. In the current study, we utilized a large, diverse electronic health record (EHR) dataset to quantify the rates and correlates of conversion to BD from these major precursor diagnoses. MethodsOur study analyzed a comprehensive ten-year EHR dataset from Johns Hopkins Medicine, a diverse urban medical center, to assess and compare the rates and correlates of conversion to BD from Major Depressive Disorder (MDD), anxiety disorders, and ADHD. Risk factors for transition were assessed as time-varying variables in proportional hazards models. ResultsOf the 21,341 patients initially included, 1232 later transitioned to a diagnosis of BD. Adjusted-one-year conversion rates for patients with MDD, anxiety disorders, and ADHD were 4.2 %, 3.4 %, and 4.0 %, respectively, with ten-year rates at 11.4 %, 9.4 %, and 10.9 %, respectively. Age (19–29 years), treatment setting (emergency and inpatient), and psychotropic medications were associated with conversion to BD across all precursor diagnoses. Severe and psychotic forms of MDD were among the strongest risk factors for transitioning to BD. Although transition rates were lower in children, risk factors were similar, with the exception of ADHD, which showed a stronger association with BD conversion in adults. ConclusionsThe highest risk of transitioning to BD was observed in patients initially diagnosed with MDD, though significant risk was also noted in those with initial diagnoses of anxiety disorders and adult ADHD.

B-137 Sepsis risk prediction using machine learning (ML) and routine blood markers up to 1 week before emergency admission

Abstract Background Sepsis is a major global health concern, responsible for approximately 50 million cases and 11 million deaths annually, representing 20% of global mortality. It is the leading cause of hospital readmissions and deaths, with associated costs exceeding $38 billion per year. Alarmingly, 80% of sepsis cases occur outside the hospital, and 70% of patients visit an outpatient facility within a week of admission. Early diagnosis and appropriate treatment could prevent 80% of sepsis-related deaths. Our study demonstrates that machine learning models hold significant promise in identifying sepsis risks using routine blood tests up to 1 week before admission. Methods For our study, we utilized a dataset of over 25,000 patient records, including both sepsis and non-sepsis cases, sourced from MIMIC-IV (version 2.2), a comprehensive electronic health record dataset from Beth Israel Deaconess Medical Center, Boston, MA. We employed a gradient boosted model with 300 trees, a maximal depth of 30 layers, and gain ratio as the criterion for attribute selection. The model's performance was evaluated using 10-fold cross-validation, demonstrating optimal performance. Input parameters included age, gender, and results of routine blood markers such as complete blood counts, differential counts, comprehensive metabolic panels, and lipid panels recorded up to 1 week before sepsis diagnosis. Results Our model achieved remarkable predictive accuracy for sepsis risk, with an area under the receiver operating characteristic curve (AUC) of 0.99 and accuracy of 0.99. Notably, the model achieved 1.0 sensitivity, 0.99 specificity, 0.99 positive predictive value, 1.0 negative predictive value, 0.99 F measure, and 0.99 Youden index. Calcium, protein, liver enzymes, hematocrit, white blood cells, and cholesterol were identified as key contributors to sepsis risk prediction. Importantly, prediction accuracy remained consistent up to 1 week prior to diagnosis. Conclusions In conclusion, dysregulated intracellular calcium handling during sepsis is associated with an intensified inflammatory response, cellular death, and subsequent organ dysfunction. Low serum protein levels, particularly albumin, are linked to poor outcomes in sepsis. Recent studies have identified liver dysfunction as an early event in sepsis, underscoring its significance in the disease process. Elevated bilirubin levels are associated with an increased risk of mortality. Low hematocrit (HCT) levels are linked to higher 30-day mortality and serve as an independent prognostic biomarker in sepsis. Monocytes play a significant role in sepsis pathophysiology, undergoing various changes including preserved phagocytic activity, increased reactive oxygen species and nitric oxide generation, and decreased production of inflammatory cytokines such as IL-6 and TNF-alpha. Septic patients often exhibit lymphopenia, affecting CD4 and CD8 T cells, B cells, and natural killer (NK) cells. Hypocholesterolemia is common in sepsis and serves as an important early prognostic factor, with low plasma concentrations associated with poor outcomes. Since every hour of delayed treatment increases the risk of death by 4%-9%, changes in routine blood markers detected by machine learning models can provide early indications of impending sepsis up to 1 week before diagnosis.

Women’s health in focus: Real-world data on valproate prescriptions during pregnancy – a cohort study in Catalonia (Spain)

ObjectivesTo characterise the exposure to valproate within a cohort of pregnant women using electronic health records (EHRs) from Catalonia (System for the Development of Research in Primary Care, SIDIAP).DesignDrug-utilisation cohort...

Advancing Healthcare Infrastructure: The Features of Iran’s Sina Electronic Health Record System

Background and aims: The Sina Electronic Health Record (SinaEHR) system was developed to support primary care delivery across Iran by enabling chronic disease monitoring, infectious disease surveillance, health policy decision-making, and quality improvement initiatives. Methods: The SinaEHR system was implemented by Mashhad University of Medical Sciences which contains health records of over 5 million people across various regions of Iran. It utilizes the International Classification of Diseases (ICD) version 10, integrating laboratory results through a laboratory information system to facilitate real-time clinical decision-making. Results: Widespread implementation of the SinaEHR system has established an invaluable healthcare infrastructure resource for the improvement of primary care services across Iran. Strict security and confidentiality measures are enforced to protect sensitive patient information. Conclusion: The SinaEHR system significantly advances Iran’s healthcare infrastructure by providing a comprehensive electronic health record platform to improve healthcare delivery and promote evidence-based decision-making.

Electronic Health Record Adoption in Oral and Dental Health Centres in Turkey

The digitalisation studies in public hospitals in Turkey started with the Health Transformation Program in 2003. The Ministry of Health has been measuring the dissemination of meaningful usage and adoption of electronic health record (EHR) using Electronic Medical Record Adoption Model (EMRAM) and Outpatient Electronic Medical Record Adoption Model (O-EMRAM). Using a comprehensive survey, our study intends to measure the stage of EHR adoption in oral and dental health centres (ODHCs) providing primary healthcare services in Turkey. The results obtained were classified by basic and comprehensive EHR functions and benchmarked against the studies addressing EHR adoption in primary healthcare institutions of other countries. In 2021, 167 ODHCs/oral and dental health hospitals (ODHHs) actively operating in Turkey completed the O-EMRAM survey. The data obtained from the survey results were analysed with QlikView Personal Edition. The availability and prevalence of medical information systems and EHR functions and their use were measured. It is observed that 59 (40.7%) of ODHCs/ODHHs have comprehensive EHR functions, while the remaining 86 (59.3%) have basic EHR functions. Under these circumstances, it can be stated that all ODHCs/ODHHs in Turkey use at least basic EHR functions. All ODHCs/ODHHs in Turkey have at least basic EHR functions. The integration of e-prescription, electronic authorisation system and e-Pulse used throughout Turkey and Health Information Management System Society (HIMSS) used in ODHCs/ODHHs can be listed as one of the decisive factors in this result. ODHCs/ODHHs still have a long way to go in adopting EHR, but the current circumstances are also promising. It is considered that this study will serve as a baseline for future measurements and produce a notable result for decision-makers.

Patient Engagement Functionalities' Influence on Quality Outcomes: The Road via EHR Presence.

Patients engaged in self-care through information technology can potentially improve the quality of healthcare they receive. This study aimed to examine how electronic health record (EHR) system functionalities help hospitals mediate the impact of patient engagement on quality outcomes-notably, readmission rates. A pooled cross-sectional study design employed data containing 3,547 observations from general acute care hospitals (2014-2018). The breadth of patient engagement functionalities adopted by a hospital was used as the independent variable and the degree of EHR presence was used as the mediating variable. Mean time to readmission for acute myocardial infarction (AMI), pneumonia, and heart failure were the dependent variables. The Baron and Kenny method was used to test mediation. Patient engagement was associated with reduced AMI readmission rates both directly and via EHR system presence. Mediation effects were present, in that a 1-unit increase in patient engagement through EHR system presence was associated with a 0.33% decrease in AMI readmission rates (p <.05). For other disease categories (heart failure and pneumonia), a significant effect was not found. For hospitals with a comprehensive EHR system, patient engagement through information technology can potentially reduce readmission rates for some diseases. More research is needed to determine which specific clinical conditions are amenable to quality improvement through patient engagement. Synergies between patient engagement functionalities and an EHR system positively affect quality outcomes. Therefore, practitioners and hospital managers should leverage hospital investments made in their EHR system infrastructure and use it to engage patients in self-care.

Pregnancy-Related Complications During the COVID-19 Pandemic in Iran.

The COVID-19 pandemic has profoundly affected healthcare systems worldwide, with significant collateral damage to vulnerable populations, including the perinatal population. This study sought to compare pregnancy-related complications before and during the COVID-19 pandemic in Iran. This retrospective data analysis was performed from February 20 to August 20, 2019 (prior to the onset of the COVID-19 pandemic) and from February 20 to August 20, 2020 (during the pandemic), encompassing the initial wave of the pandemic and the subsequent lockdown. To collect data, we utilized the medical records of 168,358 women obtained from the Iranian Maternal and Neonatal Network, which is a comprehensive electronic health record database management system specifically designed to store information pertaining to maternal and neonatal health. A total of 168,358 medical records were analyzed, with 87388 (51.9%) and 80970 (48.1%) before and during the pandemic, respectively. The occurrence of pregnancy complications was found to be significantly more frequent during the pandemic compared to the pre-pandemic period. Notably, there was a higher likelihood of experiencing preeclampsia (odds ratio [OR]=1.14, 95% confidence interval [CI]: 1.07‒1.22, P=0.0001) and gestational diabetes (OR=1.14, 95% CI: 1.09‒1.19, P=0.0001) during the pandemic. Furthermore, cesarean section (CS) became more prevalent during the pandemic in comparison to vaginal delivery (OR=1.19, 95% CI: 1.17‒-1.22, P=0.0001). Our findings demonstrated a significant association between the COVID-19 pandemic and an escalation in adverse pregnancy outcomes, notably preeclampsia, gestational diabetes, and CS deliveries. However, further research is warranted to gain a richer understanding of the intricate interplay between the COVID-19 pandemic and pregnancy complications. This is particularly crucial in light of the evolving landscape of new coronavirus variants.

A Natural Language Processing Model for COVID-19 Detection Based on Dutch General Practice Electronic Health Records by Using Bidirectional Encoder Representations From Transformers: Development and Validation Study.

Natural language processing (NLP) models such as bidirectional encoder representations from transformers (BERT) hold promise in revolutionizing disease identification from electronic health records (EHRs) by potentially enhancing efficiency and accuracy. However, their practical application in practice settings demands a comprehensive and multidisciplinary approach to development and validation. The COVID-19 pandemic highlighted challenges in disease identification due to limited testing availability and challenges in handling unstructured data. In the Netherlands, where general practitioners (GPs) serve as the first point of contact for health care, EHRs generated by these primary care providers contain a wealth of potentially valuable information. Nonetheless, the unstructured nature of free-text entries in EHRs poses challenges in identifying trends, detecting disease outbreaks, or accurately pinpointing COVID-19 cases. This study aims to develop and validate a BERT model for detecting COVID-19 consultations in general practice EHRs in the Netherlands. The BERT model was initially pretrained on Dutch language data and fine-tuned using a comprehensive EHR data set comprising confirmed COVID-19 GP consultations and non-COVID-19-related consultations. The data set was partitioned into a training and development set, and the model's performance was evaluated on an independent test set that served as the primary measure of its effectiveness in COVID-19 detection. To validate the final model, its performance was assessed through 3 approaches. First, external validation was applied on an EHR data set from a different geographic region in the Netherlands. Second, validation was conducted using results of polymerase chain reaction (PCR) test data obtained from municipal health services. Lastly, correlation between predicted outcomes and COVID-19-related hospitalizations in the Netherlands was assessed, encompassing the period around the outbreak of the pandemic in the Netherlands, that is, the period before widespread testing. The model development used 300,359 GP consultations. We developed a highly accurate model for COVID-19 consultations (accuracy 0.97, F1-score 0.90, precision 0.85, recall 0.85, specificity 0.99). External validations showed comparable high performance. Validation on PCR test data showed high recall but low precision and specificity. Validation using hospital data showed significant correlation between COVID-19 predictions of the model and COVID-19-related hospitalizations (F1-score 96.8; P<.001; R2=0.69). Most importantly, the model was able to predict COVID-19 cases weeks before the first confirmed case in the Netherlands. The developed BERT model was able to accurately identify COVID-19 cases among GP consultations even preceding confirmed cases. The validated efficacy of our BERT model highlights the potential of NLP models to identify disease outbreaks early, exemplifying the power of multidisciplinary efforts in harnessing technology for disease identification. Moreover, the implications of this study extend beyond COVID-19 and offer a blueprint for the early recognition of various illnesses, revealing that such models could revolutionize disease surveillance.

Basic electronic health record (EHR) adoption in **Türkiye is nearly complete but challenges persist

BackgroundThe digitalization studies in public hospitals in Türkiye started with the Health Transformation Program in 2003. As digitalization was accomplished, the policymakers needed to measure hospitals’ electronic health record (EHR) usage and adoptions. The ministry of health has been measuring the dissemination of meaningful usage and adoption of EHR since 2013 using Electronic Medical Record Adoption Model (EMRAM). The first published study about this analysis covered the surveys applied between 2013 and 2017. The results showed that 63.1% of all hospitals in Türkiye had at least basic EHR functions, and 36% had comprehensive EHR functions. Measuring the countrywide EHR adoption level is becoming popular in the world. This study aims to measure adoption levels of EHR in public hospitals in Türkiye, indicate the change to the previous study, and make a benchmark with other countries measuring national EHR adoption levels. The research question of this study is to reveal whether there has been a change in the adoption level of EHR in the three years since 2018 in Türkiye. Also, make a benchmark with other countries such as the US, Japan, and China in country-wide EHR adoption in 2021.MethodsIn 2021, 717 public hospitals actively operating in Türkiye completed the EMRAM survey. The survey results, deals with five topics (General Stage Status, Information Technology Security, Electronic Health Record/Clinical Data Repository, Clinical Documentation, Closed-Loop Management), was reviewed by the authors. Survey data were compared according to hospital type (Specialty Hospitals, General Hospitals, Teaching and Research Hospitals) in terms of general stage status. The data obtained from the survey results were analyzed with QlikView Personal Edition. The availability and prevalence of medical information systems and EHR functions and their use were measured.ResultsWe found that 33.7% of public hospitals in Türkiye have only basic EHR functions, and 66.3% have extensive EHR functions, which yields that all hospitals (100%) have at least basic EHR functions. That means remarkable progress from the previous study covering 2013 and 2017. This level also indicates that Türkiye has slightly better adoption from the US (96%) and much better than China (85.3%) and Korea (58.1%).ConclusionsAlthough there has been outstanding (50%) progress since 2017 in Turkish public hospitals, it seems there is still a long way to disseminate comprehensive EHR functions, such as closed-loop medication administration, clinical decision support systems, patient engagement, etc. Measuring the stage of EHR adoption at regular intervals and on analytical scales is an effective management tool for policymakers. The bottom-up adoption approach established for adopting and managing EHR functions in the US has also yielded successful results in Türkiye.

Utilization of electronic health records and associated factors among nurses in a faith-based teaching hospital, Ilishan, Nigeria.

It has been documented that nurses' use of electronic health records (EHRs) impacts clients' health outcomes positively. Some health facilities, primarily privately owned institutions, introduced EHRs for optimal healthcare. Evidence of such and associated factors among nurses must be documented to improve utilization and quality. The study assessed the utilization of EHRs and associated factors among nurses in a faith-based teaching hospital. This sequential explanatory mixed-methods study involved a sample of all 240 nurses from a teaching hospital where EHRs have been introduced. Quantitative data through semistructured questionnaires were collected and analyzed using Chi-square and logistic regression. Qualitative data were collected from 10 purposively selected nurses using an in-depth interview guide and analyzed through content analysis. The majority of participants reported availability of EHR computer software (62.8%), internet facility (84.2%), and desktops (76.3%), but EHR was poorly utilized (27.3%). Factors significantly associated were nurses who were females [OR (odds ratio) = 1.5, 95% CI (confidence interval), 0.21-11.24], BNSc degrees holders [OR = 4.3; 95% CI, 1.06-17.43]; had computer EHR software [OR = 7.4, 95% CI, 0.83-3.81], and sponsored EHR training [OR = 2.10; 95% CI, 0.24-18.6]. Noncapturing of nursing tasks and nursing standardized language by EHR software, lack of institutional enforcement on EHR use, and absence of clear EHR policies were the main identified themes for the key barriers to using EHRs. EHR was poorly utilized among nurses. Gender, educational qualification, EHR resources, and sponsored training were factors significantly associated with the use. There is an urgent need for comprehensive EHR packages, sustained sponsored training, and formulation of EHR policy for effective EHR implementation.

Prevalence of concealed arrhythmogenic substrate for sudden cardiac arrest in the general population

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): US National Heart Lung and Blood institute National Institute of Health. Background Individuals who present with SCA as the first manifestation of heart disease (Concealed Substrate) present the greatest challenge for clinical risk stratification but population-based information in the current era is lacking. We hypothesized that improved awareness and access to health care has resulted in a reduction in prevalence of Concealed Substrate. Purpose We compared the overall prevalence, secular trends and SCA etiologies of Concealed Substrate in two geographically distinct US populations. Methods Consecutive cases of SCA were ascertained prospectively from 2 large ongoing US studies, the Population Cohort A (PopulationCohortA, 2002-2018) and Population Cohort B (PopulationCohortB, 2015-2022), total population 1.85 million residents. All out-of-hospital SCA events attended by emergency medical services were determined following an established multi-source ascertainment and adjudication process, and detailed lifetime clinical record evaluation conducted for each individual. This included the comprehensive electronic health record, evaluation of SCA survivors as well as all available autopsy information. Concealed Substrate cases were defined as SCA with no prior heart disease diagnosis. Results From a total of 5708 SCA cases (age 66.5 ± 18.2, female 35%), n=3390 (age 64.3 ± 18.8, female 34%) were identified in PopulationCohortA and n=2318 (age 71.6 ± 15.7, female 36%) in PopulationCohortB. In PopulationCohortA, prevalence of Concealed Substrate was 39% (n=1282, age 58.6 ± 21.7, 32% female) and decreased modestly over time (Figure; 37% in 2002-06 vs. to 33% in 2014-18; p=0.001). In PopulationCohortB, overall prevalence of Concealed Substrate was 41% (n=949, age 68.3 ± 17.6, 33% female) and did not change significantly over time (Figure; 2015-2022; p=0.06). Etiology of SCA was determined in a subgroup of overall cases (n=3210, 56%) of which n=2727 (85%) were ischemic SCA and n=483 (15%) non-ischemic SCA. Non-ischemic SCA was more commonly identified in Concealed Substrate (29% vs. 11%; p&amp;lt;0001) and ischemic SCA was more common in Revealed Substrate (89% vs. 72%; p&amp;lt;0.0001). Conclusion There was a modest reduction in prevalence over time, 33-41% of individuals continued to present with Concealed Substrate, with higher likelihood of non-ischemic SCA. These findings suggest that widely available and inexpensive tests such as the 12-lead ECG warrant further evaluation as a pre-screening method in the general population.

Patient-specific comorbidities as prognostic variables for survival inmyelofibrosis.

Treatment decisions in primary myelofibrosis (PMF) are guided by numerous prognostic systems. Patient-specific comorbidities have influence on treatment-related survival and are considered in clinical contexts but have not been routinely incorporated into current prognostic models. We hypothesized that patient-specific comorbidities would inform prognosis and could be incorporated into a quantitative score. All patients with PMF or secondary myelofibrosis with available DNA and comprehensive electronic health record (EHR) data treated at Vanderbilt University Medical Center between 1995 and 2016 were identified within Vanderbilt's Synthetic Derivative and BioVU Biobank. We recapitulated established PMF risk scores (eg, Dynamic International Prognostic Scoring System [DIPSS], DIPSS plus, Genetics-Based Prognostic Scoring System, Mutation-Enhanced International Prognostic Scoring System 70+) and comorbidities through EHR chart extraction and next-generation sequencing on biobanked peripheral blood DNA. The impact of comorbidities was assessed via DIPSS-adjusted overall survival using Bonferroni correction. Comorbidities associated with inferior survival include renal failure/dysfunction (hazard ratio [HR], 4.3; 95% confidence interval [95% CI], 2.1-8.9; P= .0001), intracranial hemorrhage (HR, 28.7; 95% CI, 7.0-116.8; P= 2.83e-06), invasive fungal infection (HR, 41.2; 95% CI,7.2-235.2; P= 2.90e-05), and chronic encephalopathy (HR, 15.1; 95% CI, 3.8-59.4; P=.0001). The extended DIPSS model including all 4 significant comorbidities showed a significantly higher discriminating power (C-index 0.81; 95% CI, 0.78-0.84) than the original DIPSS model (C-index 0.73; 95% CI, 0.70-0.77). In summary, we repurposed an institutional biobank to identify and risk-classify an uncommon hematologic malignancy by established (eg, DIPSS) and other clinical and pathologic factors (eg, comorbidities) in an unbiased fashion. The inclusion of comorbidities into risk evaluation may augment prognostic capability of future genetics-based scoring systems.

Machine Learning Model of Emergency Department Use for Patients Undergoing Treatment for Head and Neck Cancer Using Comprehensive Multifactor Electronic Health Records

To use a hybrid method, combining statistical profiling, machine learning (ML), and clinical evaluation to predict emergency department (ED) visits among patients with head and neck cancer undergoing radiotherapy. Patients with head and neck cancer treated with radiation therapy from 2015 to 2019 were identified using electronic health record data. Records from 60 days before 90 days after treatment were analyzed. Statistical profiling and ML were used to create a predictive model for ED visits during or after radiation therapy. A comprehensive set of variables were studied. Multiple ML models were developed including extreme gradient-boosted decision tree and generalized logistic regression with comparison of multiple predictive performance metrics. Of the 1,355 patients studied, 13% had an ED visit during or after treatment. Our hybrid methodology enabled evidence-based winnowing of candidate features from 141 to 11 with clinically applicable, evidence-based thresholds. Extreme gradient boosting had the highest area under the curve (0.81 ± 0.06) with a sensitivity of 0.89 ± 0.10 and exceeded generalized logistic regression (area under the curve 0.64 ± 0.02). Significant predictors of ED visits during treatment included increasingly complex opioid use, number of prior ED visits, tumor volume, rate of change of blood urea nitrogen, total bilirubin, body mass index, and distance from hospital. Our approach combining bootstrapped statistical profiling and ML importance analysis supported integration of clinician input to identify a distilled set of phenotypical characteristics for developing ML models predicting which patients undergoing head and neck cancer radiation therapy were at risk for ED visits.

Abstract 242: Assessing Variation In Data Captured By Clinical Registries And The Electronic Health Record: Clinical Phenotypes Of Patients Undergoing Cardiac Catheterization

Background: Cardiovascular registries are built through expert abstraction of summary clinical information often represented as binary flags used to define diagnosis status. We considered underlying variation in clinical data obtained from the electronic health record (EHR) with respect to binary flags denoting presence or absence of cardiometabolic disease in a clinical data registry. Methods: We linked data abstracted for percutaneous coronary intervention (PCI) visits registered in NCDR CathPCI® with EHR data from Yale New Haven Hospital. We identified registry flags indicating diagnoses of hyperlipidemia (HLD), hypertension (HTN) and diabetes mellitus (DM) and evaluated EHR data with respect to registry diagnosis status. Results: Among 3,473 patients, mean age was 67 ± 12 years and 960 (28%) patients were female. HLD, HTN and DM registry diagnoses were present (+) in 2694 (78%), 2833 (82%) and 1358 (39%) patients, and absent (-) in 779 (22%), 640 (18%) and 2115 (61%) patients. EHR data revealed mean baseline total cholesterol of 168 ± 51 mg/dL (+HLD) vs 182 ± 44 mg/dL (-HLD), presenting systolic blood pressure of 130 ± 19 mmHg (+HTN) vs 120 ± 17 mmHg (-HTN), and A1c of 7.9% ± 1.8% (+DM) vs 5.8% ± 0.7% (-DM). Prescription histories of patients with HLD, HTN or DM diagnoses included 1 or more relevant drug class in 2409 (89%), 2620 (92%) and 1222 (90%) patients, and 2 or more drug classes in 450 (17%), 2020 (71%) and 820 (60%) patients. Patients without these diagnoses had been prescribed 1 or more drug class in 268 (34%), 223 (35%) and 55 (3%) patients, and 2 or more drug classes in 28 (4%), 71 (11%) and 9 (0.4%) patients (Figure 1). Conclusions: Binary diagnosis flags used in cardiovascular registries do not capture the full variation of clinical phenotypes underlying cardiometabolic disease processes. Harnessing comprehensive EHR data may provide a fuller picture of cardiometabolic disease, enhance quality and utility of national registries and improve outcomes research.

Proposed Model of Management for Patients with Multi-Morbidity in Iranian Hospitals.

Background: Today, multi-morbidity (MM), the presence of more than one disease in the same person at the same time, has been prevalent. This is while the healthcare delivery systems are formed based on a single-disease-oriented approach. Hence, this study intended to address presenting a model for the management of patients with multi-morbidity in Iranian hospitals. Methods: This was a mixed-method study. The data was gathered from 54 semi-structured interviews with the participation of experts in inpatient care management who were purposefully selected. The qualitative data were analyzed using content analysis. The Interpretive Structural Modeling (ISM) via STATA and Excel software was exploited in the quantitative phase. Results: The factors affecting the management of patients with multi-morbidity were identified in 26 main themes and 142 sub-themes, and ultimately, a model for improving the management of patients with multi-morbidity in Iranian hospitals at six different levels was offered. The "Comprehensive Health Care Information System (CHIS) and Electronic Health Record (EHR)" had the greatest influence and the lowest dependency. "Efforts to remove patients' confusion" had the highest dependency and the lowest influence. The results of employing the Cross Impact Matrix Multiplication Applied to Classification (MICMAC) analysis demonstrated that most of the variables are placed in the third group of linkage variables that have high driving power and dependence power. Conclusion: Concerning the sophisticated needs of patients with multi-morbidity for the management of their clinical conditions, the presented model could be provided to policymakers and health care managers as a beneficial performance guideline for improving the quality of care.

Is Electronic Health Record Safety a Paradox?

Is Electronic Health Record Safety a Paradox?

Machine Learning Applied to Electronic Health Records: Identification of Chemotherapy Patients at High Risk for Preventable Emergency Department Visits and Hospital Admissions.

PURPOSEAcute care use (ACU) is a major driver of oncologic costs and is penalized by a Centers for Medicare & Medicaid Services quality measure, OP-35. Targeted interventions reduce preventable ACU; however, identifying which patients might benefit remains challenging. Prior predictive models have made use of a limited subset of the data in the electronic health record (EHR). We aimed to predict risk of preventable ACU after starting chemotherapy using machine learning (ML) algorithms trained on comprehensive EHR data.METHODSChemotherapy patients treated at an academic institution and affiliated community care sites between January 2013 and July 2019 who met inclusion criteria for OP-35 were identified. Preventable ACU was defined using OP-35 criteria. Structured EHR data generated before chemotherapy treatment were obtained. ML models were trained to predict risk for ACU after starting chemotherapy using 80% of the cohort. The remaining 20% were used to test model performance by the area under the receiver operator curve.RESULTSEight thousand four hundred thirty-nine patients were included, of whom 35% had preventable ACU within 180 days of starting chemotherapy. Our primary model classified patients at risk for preventable ACU with an area under the receiver operator curve of 0.783 (95% CI, 0.761 to 0.806). Performance was better for identifying admissions than emergency department visits. Key variables included prior hospitalizations, cancer stage, race, laboratory values, and a diagnosis of depression. Analyses showed limited benefit from including patient-reported outcome data and indicated inequities in outcomes and risk modeling for Black and Medicaid patients.CONCLUSIONDense EHR data can identify patients at risk for ACU using ML with promising accuracy. These models have potential to improve cancer care outcomes, patient experience, and costs by allowing for targeted, preventative interventions.

Prerequisites of Personal Health Record for Chronic Kidney Disease: A Scoping Review and Evaluation of the Content Validity.

It is obvious that the Personal Health Record (PHR) is a major cornerstone for "improving the self-management of patient". However, lack of an effective and comprehensive personal health record system prohibits the widespread use of PHRs. The aim of this study was to identify the core data sets and required functionalities for designing a PHRs for chronic kidney disease (CKD) management and assess their validity. It was a study including two phases. In the initial phase, a scoping review was conducted with the aim of determination the core data sets and required functionalities for designing PHRs. Then in the second phase, the validity of data items and functionalities was determined by 25 multidisciplinary experts. 22 studies were eligible after screening 1335 titles and abstracts and reviewing 88 full texts. We determined 20 core data set and 8 required functionalities of PHRs. From the perspective of experts, 'health maintenance' and 'advance directives' were most often marked as useful but not essential, while 'test and examination', 'medication list' and 'diagnosis and comorbid conditions" were predominantly considered as essential by all experts (n=25,100%). This research is a step that we have taken to identify prerequisites that could be used for the design, development, and implementation of an effective and comprehensive electronic personal health record.

Anticipating Poor Outcomes: A Prognostic Machine Learning Model of Unplanned Visits to the Emergency Department for Patients Undergoing Treatment for Head and Neck Cancer Using Comprehensive Multi-Factor Electronic Health Records

Unplanned visits to the emergency department (ED) and hospital admissions (HA) are a significant problem in head neck cancer (HNC) with the literature suggesting that 20-50% patients will visit the ED during active cancer treatment or within 90 days of treatment completion. These unplanned visits compromise outcomes for patients with interruptions or discontinuation of care, and contribute large expense to the healthcare system. Previous work has identified static pretreatment clinical factors associated with ED visits and HA. Here, we used our custom warehouse system to evaluate pre- and during-treatment clinical, dosimetric and toxicity factors.The records of 1285 HNC cancer patients treated in our department between 2014 and 2019 were examined to profile details of a comprehensive set of factors including: demographics, location, diagnosis and staging, chemotherapy, radiation therapy, opioid use, provider reported toxicities and laboratory values and changes to create a prognostic model for ED use during or within 90 days of end of radiation therapy. For continuous variables, Youden's Index was used to identify discriminatory thresholds. Odds ratios were calculated to gauge statistical evidence of relevance. Machine learning models were constructed using gradient boosting (GB), random forest (RF), support vector machines (SVM) and naïve bayes (NB) using 10-fold cross validation.Within our cohort, 17% visited the ED during or within 90 days of the end of radiation therapy. Factors associated with ED visits included female sex (OR = 1.2), single marital status (1.6), smoking history recorded in EHR (1.3), lack of significant alcohol history recorded in EHR (1.3), living within 34 miles of hospital (1.8), age < 59 (1.2), stage IV (1.4), chemotherapy (1.3) including cisplatin (1.5), ≥ 3 changes in opioid Rx (3.4) including morphine (2.0) or fentanyl (2.2), any provider graded toxicity ≥ 3 (2.1). RF had the highest accuracy 0.87 ± 0.005, and AUC of 0.75 ± 0.05. Accuracy was significantly (P < 0.01) less for NB than for RF or GB. Highest importance scoring factors in the model were number of opioid prescription records, body mass index values and changes, volume of low risk PTV volume, age, creatinine, changes in platelet and neutrophil counts, albumin-bilirubin score, and distance.We were able to combine a comprehensive set of clinical, dosimetric and toxicity grading factors to identify an actionable clinical profile for anticipating ED. This forms the basis for next steps to develop and test interventions based on the model as reduction in unplanned acute care is a priority for clinical transformation in oncology. The scope of high-ranking inputs underscores the importance of constructing data ware house systems, that enable constructing multi-domain models. Significance of "distance of hospital" in our model and similar limitations for all ED studies underscores the need for integrated EHR across all hospital systems.