Electronic Healthcare Records Data Research Articles

Communicating exploratory unsupervised machine learning analysis in age clustering for paediatric disease

BackgroundDespite the increasing availability of electronic healthcare record (EHR) data and wide availability of plug-and-play machine learning (ML) Application Programming Interfaces, the adoption of data-driven decision-making within routine hospital workflows...

Accuracy of the recording of pneumonia events in English electronic healthcare record data in patients with chronic obstructive pulmonary disease

BackgroundIn primary care, identifying pneumonia events in people with chronic obstructive pulmonary disease (COPD) may be challenging due to similarities in symptoms with COPD exacerbations and lack of diagnostic testing. This study explored the accuracy of pneumonia diagnosis coded in primary care by comparing diagnosis in primary care with diagnosis in hospital.MethodsA study population of people with COPD in England was created using the Clinical Practice Research Datalink Aurum database linked with Hospital Episode Statistics inpatient data. Pneumonia codes only, and pneumonia code with associated clinical and/or treatment codes (chest x-ray, symptoms, antibiotics, sputum and blood culture) were used to determine pneumonia events in primary care. Events that were followed by hospitalisation within 7 days were used to estimate the positive predictive value (PPV) of pneumonia coding in primary care, using primary diagnosis of pneumonia in secondary care as the gold standard. The PPV of primary care recording of hospitalised pneumonia was also calculated.ResultsTwo hundred seventy-four thousand one hundred fifty-six COPD patients were eligible for inclusion, of whom 7,560 had an eligible pneumonia event in primary care diagnosed between 2015–2019 which was not ‘hospital-acquired’ and was diagnosed and entered on the same day. Of the 2,094 events which were followed by hospitalisation within 7 days, 1,208 had a primary diagnosis of pneumonia in hospital, representing a PPV of pneumonia coding in primary care of 57.7% (95% CI 55.6%-59.8%). Another 284 (13.6%) were diagnosed as a COPD exacerbation and 114 (5.4%) were diagnosed as another respiratory disease. Use of additional pneumonia clinical and treatment codes had a modest effect on the PPV but substantially lowered the number of events. Of the 33,603 eligible pneumonia events identified in secondary care, only 11,445 were recorded in primary care within 42 days, representing a sensitivity of 34.1% (95% CI 33.6%-34.6%).ConclusionsUse of primary care pneumonia codes and associated clinical and treatment codes to determine pneumonia is not recommended due to significant levels of misdiagnosis and many hospitalised events failing to be recorded in primary care.

Trends and Futuristic Applications of Big Data and Electronic Health Record Data in Empowering Constructive Clinical Decision Support Systems

Modern healthcare benefits significantly from big data, electronic healthcare record data technologies, and artificial intelligence, which provide robust tools for the collection and analysis of vast and diverse datasets originating from various sources, such as clinical care, administration, and research. This advancement enables the creation of information technology infrastructures that facilitate the realization of the "Learning Healthcare System Cycle," wherein healthcare practice and research seamlessly intertwine in a synergistic manner. This review focuses on illustrating how the integration of extensive data collections, empowered by big data, can enhance clinical decision-making and advance biomedical research. Most importantly, electronic health records offer several benefits, including heightened accessibility to patient information, enhanced interdisciplinary communication, improved continuity of care, legible documentation, minimized duplication, and increased efficiency. The incorporation of computerized physician order entry within electronic medical records contributes to patient safety by mitigating medication errors and offering clinical guidance through prompts and alerts during electronic order entry. Furthermore, when evidence-based clinical decision support is integrated with electronic health records, it serves as a valuable tool for guiding healthcare providers and clinicians in aligning their clinical practices with meaningful use standards and compliance with quality metrics. This review also outlines the contemporary application of clinical decision support systems in the field of medicine. Its types, their existing practical applications with documented effectiveness, prevalent challenges, and potential adverse consequences. The review concludes by offering evidence-based guidelines aimed at mitigating risks associated with support systems design, implementation, evaluation, and ongoing maintenance.

Development of an Asthma Exacerbation Risk Prediction Model for Conversational Use by Adults in England.

Improving accurate risk assessment of asthma exacerbations, and reduction via relevant behaviour change among people with asthma could save lives and reduce health care costs. We developed a simple personalised risk prediction model for asthma exacerbations using factors collected in routine healthcare data for use in a risk modelling feature for automated conversational systems. We used pseudonymised primary care electronic healthcare records from the Clinical Practice Research Datalink (CPRD) Aurum database in England. We combined variables for prediction of asthma exacerbations using logistic regression including age, gender, ethnicity, Index of Multiple Deprivation, geographical region and clinical variables related to asthma events. We included 1,203,741 patients divided into three cohorts to implement temporal validation: 898,763 (74.7%) in the training sample, 226,754 (18.8%) in the testing sample and 78,224 (6.5%) in the validation sample. The Area under the ROC curve (AUC) for the full model was 0.72 and for the restricted model was 0.71. Using a cut-off point of 0.1, approximately 27 asthma reviews by clinicians per 100 patients would be prevented compared with a strategy that all patients are regarded as high risk. Compared with patients without an exacerbation, patients who exacerbated were older, more likely to be female, prescribed more SABA and ICS in the preceding 12 months, have history of GORD, COPD, anxiety, depression, live in very deprived areas and have more severe disease. Using information available from routinely collected electronic healthcare record data, we developed a model that has moderate ability to separate patients who had an asthma exacerbation within 3 months from their index date from patients who did not. When comparing this model with a simplified model with variables that can easily be self-reported through a WhatsApp chatbot, we have shown that the predictive performance of the model is not substantially different.

Enhanced phenotypes for identifying opioid overdose in emergency department visit electronic health record data.

Accurate identification of opioid overdose (OOD) cases in electronic healthcare record (EHR) data is an important element in surveillance, empirical research, and clinical intervention. We sought to improve existing OOD electronic phenotypes by incorporating new data types beyond diagnostic codes and by applying several statistical and machine learning methods. We developed an EHR dataset of emergency department visits involving OOD cases or patients considered at risk for an OOD and ascertained true OOD status through manual chart reviews. We developed and validated prediction models using Random Forest, Extreme Gradient Boost, and Elastic Net models that incorporated 717 features involving primary and second diagnoses, chief complaints, medications prescribed, vital signs, laboratory results, and procedural codes. We also developed models limited to single data types. A total of 1718 records involving 1485 patients were manually reviewed; 541 (36.4%) patients had one or more OOD. Prediction performance was similar for all models; sensitivity varied from 94% to 97%; and area under the receiver operating characteristic curve (AUC) was 98% for all methods. The primary diagnosis and chief complaint were the most important contributors to AUC performance; primary diagnoses and medication class contributed most to sensitivity; chief complaint, primary diagnosis, and vital signs were most important for specificity. Models limited to decision support data types available in real time demonstrated robust prediction performance. Substantial prediction performance improvements were demonstrated for identifying OODs in EHR data. Our e-phenotypes could be applied in surveillance, retrospective empirical applications, or clinical decision support systems.

Prevalence, risk factors and outcomes of cardiac disease in cystic fibrosis: a multinational retrospective cohort study.

Although people living with cystic fibrosis (PwCF) often have some risk factors for cardiovascular disease, including diabetes and chronic inflammation, little is known about the long-term cardiac risk in this condition. We aimed to determine the characteristics, rates and outcomes for cardiac disease in CF. We looked at rates and outcomes for cardiac disease in 5649 adult PwCF in the UK CF Registry and 6265 adult PwCF in TriNetX (a global federated database of electronic healthcare record data). We used propensity matching to compare risk of major adverse cardiac events (MACE) (myocardial infarction, left-sided heart failure and atrial fibrillation) in PwCF against matched non-CF comparators in the general population and other inflammatory diseases. PwCF had a high prevalence of diabetes but low rates of hypertension and obesity. Some cardiac risk factors (age, diabetes and hypertension) were associated with MACE, but relationships between disease-specific risk factors (lung function and intravenous antibiotic days) were also observed. In propensity score-matched analyses, PwCF had higher risk of MACE than matched general population comparators (hazard ratio (HR) 1.65, 95% CI 1.40-1.95; p<0.001) and an equivalent or higher relative risk compared with other inflammatory conditions considered "high risk" for cardiovascular disease, including systemic lupus erythematosus (HR 0.95, 95% CI 0.82-1.09; p=0.44), rheumatoid arthritis (HR 1.21, 95% CI 1.00-1.48; p<0.001) and HIV (HR 0.93, 95% CI 0.82-1.06; p=0.29). PwCF are at increased risk of adverse cardiac disease events. Future work should focus on defining determinants of cardiovascular risk such that appropriate risk stratification can be employed.

Determining prescriptions in electronic healthcare record data: methods for development of standardized, reproducible drug codelists.

To develop a standardizable, reproducible method for creating drug codelists that incorporates clinical expertise and is adaptable to other studies and databases. We developed methods to generate drug codelists and tested this using the Clinical Practice Research Datalink (CPRD) Aurum database, accounting for missing data in the database. We generated codelists for: (1) cardiovascular disease and (2) inhaled Chronic Obstructive Pulmonary Disease (COPD) therapies, applying them to a sample cohort of 335931 COPD patients. We compared searching all drug dictionary variables (A) against searching only (B) chemical or (C) ontological variables. In Search A, we identified 165150 patients prescribed cardiovascular drugs (49.2% of cohort), and 317963 prescribed COPD inhalers (94.7% of cohort). Evaluating output per search strategy, Search C missed numerous prescriptions, including vasodilator anti-hypertensives (A and B:19696 prescriptions; C:1145) and SAMA inhalers (A and B:35310; C:564). We recommend the full search (A) for comprehensiveness. There are special considerations when generating adaptable and generalizable drug codelists, including fluctuating status, cohort-specific drug indications, underlying hierarchical ontology, and statistical analyses. Methods must have end-to-end clinical input, and be standardizable, reproducible, and understandable to all researchers across data contexts.

EE454 Cost-Effectiveness of Screening for Atrial Fibrillation Utilizing the Unafied-7 Algorithm Versus Usual Care in Individuals Aged 65 from a US Payer Perspective

There is limited evidence on the clinical and cost benefits of screening for non-valvular atrial fibrillation (NVAF). Predictive algorithms that identify individuals at high risk of NVAF may lead to earlier diagnosis and treatment in eligible patients. A recently published predictive algorithm, UNAFIED-7, was effective in identifying incident AF using Optum’s electronic healthcare record data. We aimed to assess the cost-effectiveness of one-time screening for NVAF utilizing UNAFIED-7 versus usual care in individuals aged 65 from a US payer perspective.

The quality of vital signs measurements and value preferences in electronic medical records varies by hospital, specialty, and patient demographics

We aimed to assess the frequency of value preferences in recording of vital signs in electronic healthcare records (EHRs) and associated patient and hospital factors. We used EHR data from Oxford University Hospitals, UK, between 01-January-2016 and 30-June-2019 and a maximum likelihood estimator to determine the prevalence of value preferences in measurements of systolic and diastolic blood pressure (SBP/DBP), heart rate (HR) (readings ending in zero), respiratory rate (multiples of 2 or 4), and temperature (readings of 36.0 °C). We used multivariable logistic regression to investigate associations between value preferences and patient age, sex, ethnicity, deprivation, comorbidities, calendar time, hour of day, days into admission, hospital, day of week and speciality. In 4,375,654 records from 135,173 patients, there was an excess of temperature readings of 36.0 °C above that expected from the underlying distribution that affected 11.3% (95% CI 10.6–12.1%) of measurements, i.e. these observations were likely inappropriately recorded as 36.0 °C instead of the true value. SBP, DBP and HR were rounded to the nearest 10 in 2.2% (1.4–2.8%) and 2.0% (1.3–5.1%) and 2.4% (1.7–3.1%) of measurements. RR was also more commonly recorded as multiples of 2. BP digit preference and an excess of temperature recordings of 36.0 °C were more common in older and male patients, as length of stay increased, following a previous normal set of vital signs and typically more common in medical vs. surgical specialities. Differences were seen between hospitals, however, digit preference reduced over calendar time. Vital signs may not always be accurately documented, and this may vary by patient groups and hospital settings. Allowances and adjustments may be needed in delivering care to patients and in observational analyses and predictive tools using these factors as outcomes or exposures.

Generic Patient-Centered Blockchain-Based EHR Management System

Accessing healthcare services by several stakeholders for diagnosis and treatment has become quite prevalent owing to the improvement in the industry and high levels of patient mobility. Due to the confidentiality and high sensitivity of electronic healthcare records (EHR), the majority of EHR data sharing is still conducted via fax or mail because of the lack of systematic infrastructure support for secure and reliable health data transfer, delaying the process of patient care. As a result, it is critically essential to provide a framework that allows for the efficient exchange and storage of large amounts of medical data in a secure setting. The objective of this research is to develop a Patient-Centered Blockchain-Based EHR Management (PCEHRM) system that allows patients to manage their healthcare records across multiple stakeholders and to facilitate patient privacy and control without the need for a centralized infrastructure by means of granting or revoking access or viewing one’s records. We used an Ethereum blockchain and IPFS (inter-planetary file system) to store records because of its advantage of being distributed and ensuring the immutability of records and allowing for the decentralized storage of medical metadata, such as medical reports. To achieve secure a distributed, and trustworthy access control policy, we proposed an Ethereum smart contract termed the patient-centric access control protocol. We demonstrate how the PCEHRM system design enables stakeholders such as patients, labs, researchers, etc., to obtain patient-centric data in a distributed and secure manner and integrate utilizing a web-based interface for the patient and all users to initiate the EHR sharing transactions. Finally, we tested the proposed framework in the Windows environment by compiling a smart contract prototype using Truffle and deploy on Ethereum using Web3. The proposed system was evaluated in terms of the projected medical data storage costs for the IPFS on blockchain, and the execution time for a different number of peers and document sizes. The findings of the study indicate that the proposed strategy is both efficient and practicable.

Predicting Kidney Transplant Recipient Cohorts’ 30-Day Rehospitalization Using Clinical Notes and Electronic Health Care Record Data

Rehospitalization after kidney transplant is costly to patients and health care systems and is associated with poor outcomes. Few prediction model studies have examined whether inclusion of clinical notes data from the electronic medical record (EMR) enhances prediction of rehospitalization. In a retrospective, observational study of first-time, adult kidney transplant recipients at a large, urban hospital in southeastern United States (2005-2015), we examined 30-day rehospitalization (30DR) using structured EMR and unstructured (i.e., clinical notes) data. We used natural language processing (NLP) methods on 8 types of clinical notes and included terms in predictive models using unsupervised machine learning approaches. Both the area under the receiver operating curve and precision-recall curve (ROC and PRC, respectively) were used to determine and compare model accuracy, and 5-fold cross-validation tested model performance. Among 2060 kidney transplant recipients, 30.7% were readmitted within 30 days. Predictive models using clinical notes did not meaningfully improve performance over previous models using structured data alone (ROC 0.6821; 95% confidence interval [CI]: 0.6644, 0.6998). Predictive models built using solely clinical notes performed worse than models using both clinical notes and structured data. The data that contributed to the top performing models were not identical but both included structured data and progress notes (ROC 0.6902; 95% CI: 0.6699, 0.7105). Including new features from clinical notes in risk prediction models did not substantially increase predictive accuracy for 30DR for kidney transplant recipients. Future research should consider pooling data from multiple institutions to increase sample size and avoid overfitting models.

Neural-signature methods for structured EHR prediction

Models that can effectively represent structured Electronic Healthcare Records (EHR) are central to an increasing range of applications in healthcare. Due to the sequential nature of health data, Recurrent Neural Networks have emerged as the dominant component within state-of-the-art architectures. The signature transform represents an alternative modelling paradigm for sequential data. This transform provides a non-learnt approach to creating a fixed vector representation of temporal features and has shown strong performances across an increasing number of domains, including medical data. However, the signature method has not yet been applied to structured EHR data. To this end, we follow recent work that enables the signature to be used as a differentiable layer within a neural architecture enabling application in high dimensional domains where calculation would have previously been intractable. Using a heart failure prediction task as an exemplar, we provide an empirical evaluation of different variations of the signature method and compare against state-of-the-art baselines. This first application of neural-signature methods in real-world healthcare data shows a competitive performance when compared to strong baselines and thus warrants further investigation within the health domain.

A National Registry for People With All Stages of Kidney Disease: The National Kidney Foundation (NKF) Patient Network

The National Kidney Foundation (NKF) launched the first national US kidney disease patient registry, the NKF Patient Network, that is open to patients throughout the continuum of chronic kidney disease (CKD). The Network provides individualized education and will facilitate patient-centered research, clinical care, and health policy decisions. Here, we present the overall design and the results of a feasibility study that was conducted July through December2020. Longitudinal observational cohort study of patient-entered data with or without electronic health care record (EHR) linkage in collaboration with health systems. People with CKD, age≥18 years, are invited through their provider, NKF communications, or national outreach campaign. People self-enroll and share their data through a secure portal that offers individualized education and support. The first health system partner is Geisinger. Any cause and stage of CKD, including dialysis and kidney transplant recipients. Feasibility of the EHR data transfer, participants' characteristics, and their perspectives on usability and content. Data were collected and analyzed through the registry portal powered by the Pulse Infoframe healthie 2.0 platform. During the feasibility study, 80 participants completed their profile, and 42 completed a satisfaction survey. Mean age was 57.5 years, 51% were women, 83% were White, and 89% were non-Hispanic or Latino. Of the participants, 60% were not aware of their level of estimated glomerular filtration rate and 91% of their urinary albumin-creatinine ratio. Challenges for the Network are lack of awareness of kidney disease for many with CKD, difficulty in recruiting vulnerable populations or those with low digital readiness, and loss to follow-up, all leading to selection bias. The Network is positioned to become a national and international platform for real-world data that can inform the development of patient-centered research, care, and treatments.

215.3: Prediction of Post-transplant Hospitalization Among Kidney Transplant Recipients With Clinical Notes and Electronic Healthcare Record Data

215.3: Prediction of Post-transplant Hospitalization Among Kidney Transplant Recipients With Clinical Notes and Electronic Healthcare Record Data

Standard Vocabularies to Improve Machine Learning Model Transferability With Electronic Health Record Data: Retrospective Cohort Study Using Health Care-Associated Infection.

BackgroundWith the widespread adoption of electronic healthcare records (EHRs) by US hospitals, there is an opportunity to leverage this data for the development of predictive algorithms to improve clinical care. A key barrier in model development and implementation includes the external validation of model discrimination, which is rare and often results in worse performance. One reason why machine learning models are not externally generalizable is data heterogeneity. A potential solution to address the substantial data heterogeneity between health care systems is to use standard vocabularies to map EHR data elements. The advantage of these vocabularies is a hierarchical relationship between elements, which allows the aggregation of specific clinical features to more general grouped concepts.ObjectiveThis study aimed to evaluate grouping EHR data using standard vocabularies to improve the transferability of machine learning models for the detection of postoperative health care–associated infections across institutions with different EHR systems.MethodsPatients who underwent surgery from the University of Utah Health and Intermountain Healthcare from July 2014 to August 2017 with complete follow-up data were included. The primary outcome was a health care–associated infection within 30 days of the procedure. EHR data from 0-30 days after the operation were mapped to standard vocabularies and grouped using the hierarchical relationships of the vocabularies. Model performance was measured using the area under the receiver operating characteristic curve (AUC) and F1-score in internal and external validations. To evaluate model transferability, a difference-in-difference metric was defined as the difference in performance drop between internal and external validations for the baseline and grouped models.ResultsA total of 5775 patients from the University of Utah and 15,434 patients from Intermountain Healthcare were included. The prevalence of selected outcomes was from 4.9% (761/15,434) to 5% (291/5775) for surgical site infections, from 0.8% (44/5775) to 1.1% (171/15,434) for pneumonia, from 2.6% (400/15,434) to 3% (175/5775) for sepsis, and from 0.8% (125/15,434) to 0.9% (50/5775) for urinary tract infections. In all outcomes, the grouping of data using standard vocabularies resulted in a reduced drop in AUC and F1-score in external validation compared to baseline features (all P<.001, except urinary tract infection AUC: P=.002). The difference-in-difference metrics ranged from 0.005 to 0.248 for AUC and from 0.075 to 0.216 for F1-score.ConclusionsWe demonstrated that grouping machine learning model features based on standard vocabularies improved model transferability between data sets across 2 institutions. Improving model transferability using standard vocabularies has the potential to improve the generalization of clinical prediction models across the health care system.

Assessment of an institutional guideline for vancomycin dosing and identification of predictive factors associated with dose and drug trough levels

To evaluate the effectiveness of an antimicrobial guideline for vancomycin prescribing deployed using electronic prescribing aid and web/phone-based app. To define factors associated with guideline compliance and drug levels, and to investigate if antimicrobial dosing recommendations can be refined using routinely collected electronic healthcare record data. We used data from Oxford University Hospitals between 01-January-2016 and 01-June-2021 and multivariable regression models to investigate factors associated with dosing compliance, drug levels and acute kidney injury (AKI). 3767 patients received intravenous vancomycin for ≥24h. Compliance with recommended loading and initial maintenance doses reached 84% and 70% respectively; 72% of subsequent maintenance doses were correctly adjusted. However, only 26% first and 32% subsequent levels reached the target range, and for patients with ongoing vancomycin treatment, 55-63% achieved target levels at 5 days. Drug levels were independently higher in older patients. Incidence of AKI was low (5.7%). Model estimates were used to propose updated age, weight and eGFR specific guidelines. Despite good compliance with guidelines for vancomycin dosing, the proportion of drug levels achieving the target range remained suboptimal. Routinely collected electronic data can be used at scale to inform pharmacokinetic studies and could improve vancomycin dosing.

Long COVID burden and risk factors in 10 UK longitudinal studies and electronic health records

The frequency of, and risk factors for, long COVID are unclear among community-based individuals with a history of COVID-19. To elucidate the burden and possible causes of long COVID in the community, we coordinated analyses of survey data from 6907 individuals with self-reported COVID-19 from 10 UK longitudinal study (LS) samples and 1.1 million individuals with COVID-19 diagnostic codes in electronic healthcare records (EHR) collected by spring 2021. Proportions of presumed COVID-19 cases in LS reporting any symptoms for 12+ weeks ranged from 7.8% and 17% (with 1.2 to 4.8% reporting debilitating symptoms). Increasing age, female sex, white ethnicity, poor pre-pandemic general and mental health, overweight/obesity, and asthma were associated with prolonged symptoms in both LS and EHR data, but findings for other factors, such as cardio-metabolic parameters, were inconclusive.

Predicting hospitalization of COVID-19 positive patients using clinician-guided machine learning methods.

ObjectivesThe coronavirus disease 2019 (COVID-19) is a resource-intensive global pandemic. It is important for healthcare systems to identify high-risk COVID-19-positive patients who need timely health care. This study was conducted to predict the hospitalization of older adults who have tested positive for COVID-19.MethodsWe screened all patients with COVID test records from 11 Mass General Brigham hospitals to identify the study population. A total of 1495 patients with age 65 and above from the outpatient setting were included in the final cohort, among which 459 patients were hospitalized. We conducted a clinician-guided, 3-stage feature selection, and phenotyping process using iterative combinations of literature review, clinician expert opinion, and electronic healthcare record data exploration. A list of 44 features, including temporal features, was generated from this process and used for model training. Four machine learning prediction models were developed, including regularized logistic regression, support vector machine, random forest, and neural network.ResultsAll 4 models achieved area under the receiver operating characteristic curve (AUC) greater than 0.80. Random forest achieved the best predictive performance (AUC = 0.83). Albumin, an index for nutritional status, was found to have the strongest association with hospitalization among COVID positive older adults.ConclusionsIn this study, we developed 4 machine learning models for predicting general hospitalization among COVID positive older adults. We identified important clinical factors associated with hospitalization and observed temporal patterns in our study cohort. Our modeling pipeline and algorithm could potentially be used to facilitate more accurate and efficient decision support for triaging COVID positive patients.

P183 Establishing the epidemiology of rheumatoid arthritis, psoriatic arthritis, and axial spondyloarthritis in England using primary care electronic health record data

Background/AimsThe substantial personal and socioeconomic costs associated with rheumatoid arthritis (RA), psoriatic arthritis (PsA), and axial spondyloarthritis (SpA) make understanding their epidemiology crucial. The Clinical Practice Research Datalink (Aurum) is an electronic healthcare record (EHR) database, containing primary care records from ∼20% of English practices (>13 million patients currently registered). To determine RA/PsA/axial SpA epidemiology using EHR data, validated methods need to be applied to ascertain patients with these diagnoses. To address this, we updated and applied approaches validated in other primary care EHR databases in Aurum and described the annual incidence/point-prevalence of RA/PsA/axial SpA alongside patient characteristics (providing indirect evidence of coding accuracy).MethodsDiagnosis and synthetic disease-modifying anti-rheumatic drug (DMARD) prescription code lists were constructed, and pre-defined approaches for ascertaining patients with RA/axial SpA/PsA applied. The annual incidence and point-prevalence of RA/PsA/axial SpA were calculated from 2004-2020. Samples were stratified by age/gender, and mean age and gender/ethnic-group relative frequencies described. The study was approved by the CPRD Independent Scientific Advisory Committee (reference 20_000244).ResultsFrom 2004-2019 the point-prevalence of RA/PsA increased annually, peaking in 2019 (RA 7.79/1,000; PsA 2.87/1,000) then falling slightly. From 2004-2020 the point-prevalence of axial SpA increased annually (except in 2018/2019), peaking in 2020 (1.13/1,000). Annual RA incidence was higher between 2013-2019 (when included in the Quality Outcomes Framework, ranging 0.491 to 0.521/1,000 person-years) than 2004-2012 (ranging 0.345 to 0.400/1,000 person-years). The annual incidence of PsA and axial SpA increased from 2006 (0.108 to a peak of 0.172/1,000 person-years) and 2010 (0.025 to a peak of 0.045/1,000 person-years), respectively. These years were when new disease classification criteria were introduced. Marked falls in the annual incidence of RA, PsA and axial SpA between 2019 and 2020 were seen, reducing by 40.1%, 67.4% and 38.1%, respectively, reflecting the impact of the COVID-19 pandemic on arthritis diagnoses. Stratifying incidence/prevalence by age/gender broadly showed expected patterns (although the incidence of axial SpA/PsA in women increased over time), and the mean age and gender proportions followed those previously reported.ConclusionThe approaches we used to determine patients with RA, PsA, and axial SpA in Aurum led to incidence/prevalence estimates broadly consistent with published studies, and patient characteristics as would be expected. These data support the potential of the Aurum-updated ascertainment approaches for use in further studies of RA, PsA and axial SpA.Disclosure I. Scott: None. R. Whittle: None. J. Bailey: None. H. Twohig: None. S. Hider: None. C. Mallen: None. S. Muller: None. K. Jordan: None.

Fusion of sequential visits and medical ontology for mortality prediction

The goal of mortality prediction task is to predict the future death risk of patients according to their previous Electronic Healthcare Records (EHR). The main challenge of mortality prediction is how to design an accurate and robust predictive model with sequential, multivariate, sparse and irregular EHR data. In addition, the performance of model may be affected by lack of sufficient information of some patients with rare diseases in EHRs. To address these challenges, we propose a model to fuse Sequential visits and Medical Ontology to predict patients’ death risk. SeMO not only learns reasonable embeddings for medical concepts from sequential and irregular visits, but also exploits medical ontology to improve the prediction performance. With integration of multivariate features, SeMO learns robust representations of medical codes, mitigating data insufficiency and insightful sequential dependencies among patient’s visits. Experimental results on real world datasets prove that the proposed SeMO improves the prediction performance compared with the baseline approaches. Our model achieves an precision of up to 0.975. Compared with RNN, the precision has been improved up to 2.204%.