Health Records Research Articles

Enhancing early detection of cognitive decline in the elderly: a comparative study utilizing large language models in clinical notes

BackgroundLarge language models (LLMs) have shown promising performance in various healthcare domains, but their effectiveness in identifying specific clinical conditions in real medical records is less explored. This study evaluates LLMs for detecting signs of cognitive decline in real electronic health record (EHR) clinical notes, comparing their error profiles with traditional models. The insights gained will inform strategies for performance enhancement. MethodsThis study, conducted at Mass General Brigham in Boston, MA, analysed clinical notes from the four years prior to a 2019 diagnosis of mild cognitive impairment in patients aged 50 and older. We developed prompts for two LLMs, Llama 2 and GPT-4, on Health Insurance Portability and Accountability Act (HIPAA)-compliant cloud-computing platforms using multiple approaches (e.g., hard prompting, retrieval augmented generation, and error analysis-based instructions) to select the optimal LLM-based method. Baseline models included a hierarchical attention-based neural network and XGBoost. Subsequently, we constructed an ensemble of the three models using a majority vote approach. Confusion-matrix-based scores were used for model evaluation. FindingsWe used a randomly annotated sample of 4,949 note sections from 1,969 patients (women: 1,046 [53.1%]; age: mean, 76.0 [SD, 13.3] years), filtered with keywords related to cognitive functions, for model development. For testing, a random annotated sample of 1,996 note sections from 1,161 patients (women: 619 [53.3%]; age: mean, 76.5 [SD, 10.2] years) without keyword filtering was utilised. GPT-4 demonstrated superior accuracy and efficiency compared to Llama 2, but did not outperform traditional models. The ensemble model outperformed the individual models in terms of all evaluation metrics with statistical significance (p<0.01), achieving a precision of 90.2% [95% CI: 81.9%-96.8%], a recall of 94.2% [95% CI: 87.9%-98.7%], and an F1-score of 92.1% [95% CI: 86.8%-96.4%]. Notably, the ensemble model showed a significant improvement in precision, increasing from a range of 70%-79% to above 90%, compared to the best-performing single model. Error analysis revealed that 63 samples were incorrectly predicted by at least one model; however, only 2 cases (3.2%) were mutual errors across all models, indicating diverse error profiles among them. InterpretationLLMs and traditional machine learning models trained using local EHR data exhibited diverse error profiles. The ensemble of these models was found to be complementary, enhancing diagnostic performance. Future research should investigate integrating LLMs with smaller, localised models and incorporating medical data and domain knowledge to enhance performance on specific tasks.

Health Care Delivery Site- and Patient-Level Factors Associated With COVID-19 Primary Vaccine Series Completion in a National Network of Community Health Centers.

Objectives. To assess multilevel factors associated with variation in COVID-19 vaccination rates in a US network of community health centers. Methods. Using multilevel logistic regression with electronic health record data from ADVANCE (Accelerating Data Value Across a National Community Health Center Network; January 1, 2022-December 31, 2022), we assessed associations between health care delivery site-level (n = 1219) and patient-level (n = 1 864 007) characteristics and COVID-19 primary vaccine series uptake. Results. A total of 1 337 440 patients completed the COVID-19 primary vaccine series. Health care delivery site characteristics were significantly associated with lower series completion rates, including being located in non-Medicaid expansion states and isolated or rural communities and serving fewer patients. Patient characteristics associated with significantly lower likelihood of completing the vaccine series included being Black/African American or American Indian/Alaska Native (vs White), younger age, lower income, being uninsured or publicly insured (vs using private insurance), and having fewer visits. Conclusions. Both health care delivery site- and patient-level factors were significantly associated with lower COVID-19 vaccine uptake. Community health centers have been a critical resource for vaccination during the pandemic. (Am J Public Health. 2024;114(11):1242-1251. https://doi.org/10.2105/AJPH.2024.307773).

Comparing Provider and ChatGPT Responses to Breast Reconstruction Patient Questions in the Electronic Health Record.

Patient-directed Electronic Health Record (EHR) messaging is used as an adjunct to enhance patient-physician interactions but further burdens the physician. There is a need for clear electronic patient communication in all aspects of medicine, including plastic surgery. We can potentially utilize innovative communication tools like ChatGPT. This study assesses ChatGPT's effectiveness in answering breast reconstruction queries, comparing its accuracy, empathy, and readability with healthcare providers' responses. Ten deidentified questions regarding breast reconstruction were extracted from electronic messages. They were presented to ChatGPT3, ChatGPT4, plastic surgeons, and advanced practice providers for response. ChatGPT3 and ChatGPT4 were also prompted to give brief responses. Using 1-5 Likert scoring, accuracy and empathy were graded by 2 plastic surgeons and medical students, respectively. Readability was measured using Flesch Reading Ease. Grades were compared using 2-tailed t tests. Combined provider responses had better Flesch Reading Ease scores compared to all combined chatbot responses (53.3 ± 13.3 vs 36.0 ± 11.6, P < 0.001) and combined brief chatbot responses (53.3 ± 13.3 vs 34.7 ± 12.8, P < 0.001). Empathy scores were higher in all combined chatbot than in those from combined providers (2.9 ± 0.8 vs 2.0 ± 0.9, P < 0.001). There were no statistically significant differences in accuracy between combined providers and all combined chatbot responses (4.3 ± 0.9 vs 4.5 ± 0.6, P = 0.170) or combined brief chatbot responses (4.3 ± 0.9 vs 4.6 ± 0.6, P = 0.128). Amid the time constraints and complexities of plastic surgery decision making, our study underscores ChatGPT's potential to enhance patient communication. ChatGPT excels in empathy and accuracy, yet its readability presents limitations that should be addressed.

Advancements in health information technology and their role in enhancing cancer care: Innovations in early detection, treatment, and data privacy

The rapid advancement of Health Information Technology (HIT) has revolutionized cancer care, introducing new capabilities in early detection, individualized treatment, and data security. This paper reviews the key HIT innovations that have reshaped oncology, including the widespread adoption of Electronic Health Records (EHRs), the use of artificial intelligence (AI) in diagnostic processes, and the application of telemedicine and predictive analytics to improve treatment planning. These technologies have enabled earlier and more accurate diagnoses, expanded access to care, and facilitated data-driven, personalized treatment strategies. In parallel, this review addresses the importance of safeguarding sensitive patient data, focusing on encryption, anonymization techniques, and the regulatory frameworks like HIPAA and GDPR that ensure data privacy. Finally, the paper discusses future directions for HIT, particularly the potential of AI-driven decision support systems, real-time data analytics, and privacy-enhancing algorithms to further advance cancer care, creating a more secure, efficient, and patient-centered healthcare system.

Sociodemographic Disparities in 1-Year Outcomes of Children With Community-Acquired Acute Kidney Injury.

Racial disparities have been identified in pediatric community-acquired acute kidney injury (CA-AKI), and they are associated with increased risk of child mortality, morbidity, and progression of kidney disease. To assess clinical outcomes at 1 year among children with CA-AKI, stratified by age, race, and ethnicity. This retrospective cohort study is a population-based analysis of deidentified, aggregated electronic health record data collected by 61 large health care organizations from 2003 to 2023 and accessed through the TriNetX platform. Outcomes were assessed at 1 year after a CA-AKI episode. Participants included pediatric patients (aged <18 years) with AKI. Data were accessed in January 2024. A diagnosis of CA-AKI and sociodemographic factors such as race, ethnicity, and age, as reported in electronic health records. The primary end point of this study was to assess differences in clinical outcomes within 1 year of an episode of CA-AKI, including all-cause emergency department (ED) visits, intensive care unit (ICU) admissions, mechanical intubation and ventilation, and mortality. Risk was compared between White children and Asian (including Asian, Native Hawaiian, and Other Pacific Islander), Black, and Hispanic children, stratified by age group. Measures of association, Cox proportional hazard analyses, and Kaplan-Meier survival curves were performed within the TriNetX Advanced Analytics Platform between racial and ethnic groups for each analysis. From the total sample of 18 152 children, those with hospital-acquired AKI, chronic kidney disease, end-stage kidney failure, or dialysis dependence were excluded, leaving a final cohort of 17 125 children (mean [SD] age, 11.2 [5.2] years; 9424 male [55.3%]). Eligible patients were divided into racial and ethnic groups as follows: non-Hispanic Asian, 1169 children (6.5%); non-Hispanic Black, 4636 children (27.3%); Hispanic, 1786 children (10.2%); and non-Hispanic White, 9534 children (55.9%). Patients were further subdivided into groups aged 0 to 9 years (546 Asian children, 1675 Black children, 689 Hispanic children, and 3340 White children) and 10 to 18 years (623 Asian children, 2961 Black children, 1091 Hispanic children, and 6104 White children). Within 1 year of CA-AKI diagnosis, compared with White children, Black children experienced greater rates of ED visits (hazard ratio [HR], 1.53; 95% CI, 1.40-1.67), ICU admissions (HR, 1.31; 95% CI, 1.16-1.47), mechanical ventilation (HR, 1.33; 95% CI, 1.13-1.56), and all-cause mortality (HR, 1.27; 95% CI, 1.09-1.48), as well as the greatest risk for composite outcomes (HR, 1.43; 95% CI, 1.33-1.53). Hispanic children experienced greater rates of ED visits (HR, 1.40; 95% CI, 1.21-1.62) and the greatest risk of all-cause mortality (HR, 1.66; 95% CI, 1.31-2.09), whereas Asian children experienced greater rates of mechanical ventilation (HR, 1.69; 95% CI, 1.26-2.27), compared with White children. Black and Hispanic children aged 0 to 9 years were at greatest risk of experiencing poor clinical outcomes. Black children had a 11.41% lower survival probability and Hispanic children had a 7.14% lower survival probability compared with White children after an initial ED encounter. Among children with an identified episode of CA-AKI diagnosed in an ED, within 1 year, Black and Hispanic children had a poorer survival probability compared with White children. Future studies are needed to understand these disparities and improve awareness and follow-up after emergency care.

Electronic Health Record Phenotyping of Pediatric Suicide-Related Emergency Department Visits.

Suicide is a leading cause of death among young people. Accurate detection of self-injurious thoughts and behaviors (SITB) underpins equity in youth suicide prevention. To compare methods of detecting SITB using structured electronic health information and measure algorithmic performance across demographics. This cross-sectional study used medical records among youths aged 6 to 17 years with at least 1 mental health-related emergency department (ED) visit in 2017 to 2019 to an academic health system in Southern California serving 787 000 unique individuals each year. Analyses were conducted between January and September 2023. Multiexpert electronic health record review ascertained the presence of SITB using the Columbia Classification Algorithm of Suicide Assessment. Random forest classifiers with nested cross-validation were developed using (1) International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes for nonfatal suicide attempt and self-harm and chief concern and (2) all available structured data, including diagnoses, medications, and laboratory tests. Detection performance was assessed overall and stratified by age group, sex, and race and ethnicity. The sample comprised 2702 unique youths with an MH-related ED visit (1384 youths who identified as female [51.2%]; 131 Asian [4.8%], 266 Black [9.8%], 719 Hispanic [26.6%], 1319 White [48.8%], and 233 other race [8.6%]; median [IQR] age, 14 [12-16] years), including 898 children and 1804 adolescents. Approximately half of visits were related to SITB (1286 visits [47.6%]). Sensitivity of SITB detection using only codes and chief concern varied by age group and increased until age 15 years (6-9 years: 59.3% [95% CI, 48.5%-69.5%]; 10-12 years: 69.0% [95% CI, 63.8%-73.9%]; 13-15 years: 88.4% [95% CI, 85.1%-91.2%]; 16-17 years: 83.1% [95% CI, 79.1%-86.6%]), while specificity remained constant. The area under the receiver operating characteristic curve (AUROC) was lower among preadolescents (0.841 [95% CI, 0.815-0.867]) and male (0.869 [95% CI, 0.848-0.890]), Black (0.859 [95% CI, 0.813-0.905]), and Hispanic (0.861 [95% CI, 0.831-0.891]) youths compared with adolescents (0.925 [95% CI, 0.912-0.938]), female youths (0.923 [95% CI, 0.909-0.937]), and youths of other races and ethnicities (eg, White: 0.901 [95% CI, 0.884-0.918]). Augmented classification (ie, using all available structured data) outperformed classification with codes and chief concern alone (AUROC, 0.975 [95% CI, 0.968-0.980] vs 0.894 [95% CI, 0.882-0.905]; P < .001). In this study, diagnostic codes and chief concern underestimated SITB prevalence, particularly among minoritized youths. These results suggest that priority on algorithmic fairness in suicide prevention strategies must extend to accurate detection of youths with suicide-related emergencies.

The Investigation of the Application of RAG Technology in the Field of EHR

Electronic Health Records (EHR) play a crucial role in contemporary medical information systems, capturing extensive data in multiple scenarios. This review explores the impact of Retrieval Augmented Generation (RAG) on managing EHR. By integrating retrieval and generation processes, RAG significantly enhances data handling, clinical decision support, and patient management. EHRs contain extensive data like patient histories and diagnostic information, which are becoming increasingly complex. The RAG improves the accuracy and richness of this data processing by using a retrieval module to extract relevant text fragments and a generation module to create precise outputs. Despite its potential, RAG faces challenges such as data inconsistency, privacy concerns, and the need for efficient training. Recent studies highlight the RAG’s effectiveness in summarizing clinical notes and enhancing prediction accuracy, suggesting a promising future in healthcare. However, ongoing research is necessary to optimize the RAG’s application and address these challenges, aiming to transform healthcare delivery effectively.

Artificial Intelligence (AI)-Driven Frailty Prediction Using Electronic Health Records in Hospitalized Patients With Cardiovascular Disease

Artificial Intelligence (AI)-Driven Frailty Prediction Using Electronic Health Records in Hospitalized Patients With Cardiovascular Disease

Health Catch-UP!: a realist evaluation of an innovative multi-disease screening and vaccination tool in UK primary care for at-risk migrant patients

BackgroundMigrants to the UK face disproportionate risk of infections, non-communicable diseases, and under-immunisation compounded by healthcare access barriers. Current UK migrant screening strategies are unstandardised with poor implementation and low uptake. Health Catch-UP! is a collaboratively produced digital clinical decision support system that applies current guidelines (UKHSA and NICE) to provide primary care professionals with individualised multi-disease screening (7 infectious diseases/blood-borne viruses, 3 chronic parasitic infections, 3 non-communicable disease or risk factors) and catch-up vaccination prompts for migrant patients.MethodsWe carried out a mixed-methods process evaluation of Health Catch-UP! in two urban primary healthcare practices to integrate Health Catch-UP! into the electronic health record system of primary care, using the Medical Research Council framework for complex intervention evaluation. We collected quantitative data (demographics, patients screened, disease detection and catch-up vaccination rates) and qualitative participant interviews to explore acceptability and feasibility.ResultsNinety-nine migrants were assessed by Health Catch-UP! across two sites (S1, S2). 96.0% (n = 97) had complete demographics coding with Asia 31.3% (n = 31) and Africa 25.2% (n = 25), the most common continents of birth (S1 n = 92 [48.9% female (n = 44); mean age 60.6 years (SD 14.26)]; and S2 n = 7 [85.7% male (n = 6); mean age 39.4 years (SD16.97)]. 61.6% (n = 61) of participants were eligible for screening for at least one condition and uptake of screening was high 86.9% (n = 53). Twelve new conditions were identified (12.1% of study population) including hepatitis C (n = 1), hypercholesteraemia (n = 6), pre-diabetes (n = 4), and diabetes (n = 1). Health Catch-UP! identified that 100% (n = 99) of patients had no immunisations recorded; however, subsequent catch-up vaccination uptake was poor (2.0%, n = 1). Qualitative data supported acceptability and feasibility of Health Catch-UP! from staff and patient perspectives, and recommended Health Catch-UP! integration into routine care (e.g. NHS health checks) with an implementation package including staff and patient support materials, standardised care pathways (screening and catch-up vaccination, laboratory, and management), and financial incentivisation.ConclusionsClinical Decision Support Systems like Health Catch-UP! can improve disease detection and implementation of screening guidance for migrant patients but require robust testing, resourcing, and an effective implementation package to support both patients and staff.

What does routine healthcare data tell us about the diagnostic pathway for older people with heart failure in Wales: under-testing, problems with access, or poor coding?

Abstract Background United Kingdom guidelines (NICE, NG106) for suspected chronic heart failure (HF) recommend that NT pro-BNP is performed as a first-line test. The result is also used to triage individuals; when NT pro-BNP is &amp;lt;400ng/L, a diagnosis of HF is less likely, echocardiography (echo) and specialist assessment should be performed within 6 weeks when the result is 400-2000ng/L, and within 2 weeks if &amp;gt;2000ng/L. How accurately this guidance is followed in clinical practice in Wales is not known. Purpose To examine how the current diagnostic pathway is recorded in routinely-collected electronic health record (EHR) data for older people in Wales with a diagnosis of heart failure. Methods We conducted a retrospective, population-level, observational study using linked anonymised EHR data in Wales (2015-22). We included patients aged 65y+ with a HF diagnosis in their primary or secondary care records and with a minimum of 12m pre and post-diagnosis data (for assessment of diagnostic pathway). We looked back from the point of HF diagnosis for codes associated with prior cardiovascular conditions (any time prior), symptoms (within 3 years), and key diagnostic tests (within 90 days). Results The final cohort comprised 51,020 individuals 65y and over with a diagnosis of heart failure between 2015-2022. Incidence and prevalence of heart failure increased across the study period (1.4-1.5% and 8.5-9.1% respectively). HF diagnosis was first evident in secondary care records in 54.9%, in primary care 36.3%, and in both data sources on the same day in 8.8%. Preceding cardiovascular diagnoses included hypertension (69.5%), atrial fibrillation (36.7%), myocardial infarction (30.5%), coronary artery disease (25.7%), and valve disease (17.7%). The most common recorded symptoms preceding diagnosis were breathlessness (41.4%), ankle oedema (9.8%), and fatigue (8.8%). Searched for symptoms were not recorded in 43.8%. In the 90 days preceding the diagnosis, 7-30% had recorded NT pro-BNP with the number of tests increasing throughout the study period (Figure 1). In the 90 days prior to HF diagnosis, 28-31% had an echo recorded. NT pro-BNP was &amp;gt;2000ng/L in 51.8%, 400-2000ng/L in 38.6% and &amp;lt;400ng/L in 9.6%. Conclusions In Wales, a greater number of individuals with heart failure appear to be diagnosed during hospital admission, compared with primary care. Many individuals with a code for HF in their records, do not have recorded objective evidence of cardiac dysfunction. The proportion of first-line and definitive tests appears low. It is unclear whether this is because of under-referral, difficulty accessing tests in primary care, or inaccurate coding of test referrals and results. These results are important in evaluating health service delivery and workforce planning.

Experiences from the European Region: case studies and scenarios to inform ways forward in digital health

Abstract Member States in the WHO European Region face similar challenges when it comes to strengthening their HIS and implementing digital health solutions. From WHO/Europe’s work with its Member States, several important preconditions have become apparent for successful and sustainable digital transformation. Firstly, the importance of recognising that implementing digital tools is not a one-off investment, but that this takes a long-term vision including long-term budgeting. Having business continuity plans also is a key element of sustainable implementation of digital tools, that is often not adequately addressed in practice. Secondly, the importance of comprehensive strategic planning. Implementing digital tools is not only about software and hardware, but also about people and processes. Here it is also crucial for policymakers to understand that, to be able to harvest the benefits of data that are gathered by means of digital tools, resources for data processing and analysis also need to be made available. Last but not least, an important precondition for successful digital transformation is stakeholder involvement. Often, new software is developed without adequately involving the end-user, resulting in for example apps that do not meet patients’ needs. Or in electronic health record systems that are not user-friendly for the healthcare staff that need to enter the data, with incomplete and low quality information as a consequence. Stakeholder involvement is essential in strategic planning for digital health as well. Think for example about standardization and accreditation agencies, and institutes that will be secondary users of data for statistics and research. In this presentation, several case studies from across the European Region will be presented, illustrating how digital transformation in health can be successfully and sustainably implemented.

The association between the mental health disorders, substance abuse, and tobacco use with head & neck cancer stage at diagnosis.

Mental health disorders, substance abuse, and tobacco use are prevalent in the US population. However, the association between these conditions and head and neck cancer (HNC) stage is poorly understood. This research aims to uncover the relationship between pre-existing mental health disorders, substance abuse, and tobacco use and HNC stage at diagnosis in patients receiving care in an integrated, public safety-net healthcare system. This study was a secondary data analysis of linked hospital tumor registries and electronic health record (EHR) data. The study's primary independent variables were the comorbidities of mental health disorders, substance abuse, and tobacco use. The dependent variable was HNC stage at diagnosis, operationalized as early stage (i.e., stages I, II, and III) and advanced stage (stage IV, IVA, IVB, or IVC). The analysis included multivariable logistic regression adjusted for covariates of demographic variables, tumor anato RESULTS: The study population consisted of 357 patients with median age of 59years, and was primarily male (77%), diverse (Black or African American 41%; Hispanic 22%), and from neighborhoods with low income (median average annual household income $39,785). Patients with a history of mental health disorders with or without tobacco use had significantly lower odds of advanced stage HNC at diagnosis (adjusted OR = 0.35, 95% Confidence Interval [CI]: 0.17-0.72.) while patients with a history of substance abuse with or without tobacco use had significantly higher odds of advanced stage HNC at diagnosis (adjusted OR 1.41, 95% CI: 1.01-1.98) than patients with no history of mental health disorders, substance abuse, or tobacco use. The relationship between HNC stage at diagnosis and the comorbidities of mental health disorders, substance abuse, or tobacco differs depending on the type and co-occurrence of these comorbidities. These findings demonstrate the need for innovative care delivery models and education initiatives tailored to meet the needs of patients with mental health disorders, substance abuse, and tobacco use that facilitate early detection of HNC.

Excess mortality in individuals with and without COVID-19 vaccination in the Netherlands in 2021

Abstract Background The COVID-19 pandemic greatly affected public health, evident in higher mortality rates in 2020-2022. COVID-19 vaccinations reduced global mortality, though this may vary across subpopulations. We aimed to study the role of COVID-19 vaccination in excess mortality by examining individuals with and without COVID-19 vaccine while accounting for characteristics such as income and frailty. Methods In this observational study, electronic health record data from general practitioners participating in the Nivel Primary Care Database were linked to demographic data and COVID-19 vaccination data. Patients were categorized in age groups, based on vaccination invitation dates, with the index date set as vaccination date for vaccinated individuals and the expected vaccination date for nonvaccinated ones. Expected mortality rates were calculated using 2015-2018 data, accounting for season, sex, income, migration background, palliative care, and frailty. We compared expected mortality to observed mortality in 2021 to determine excess mortality. Results Preliminary results show that nonvaccinated individuals (n = 160.253) were younger, and less frail, but more often received palliative care than vaccinated individuals (n = 819.363). In nonvaccinated individuals, excess mortality in the three months after expected vaccination ranged from 22-61% for 36-60 year olds to 209-233% for the 81+ year olds. In vaccinated individuals, excess mortality ranged from -64- -61% in 36-60 year olds to 2-4% in 81+ year olds. When not correcting for palliative care, excess mortality estimations increased significantly in nonvaccinated individuals (e.g. for 81+ year olds: 504%-514%). Conclusions Excess mortality was hardly present in vaccinated subgroups, while it was high in all nonvaccinated subgroups. Our study shed light on characteristics of nonvaccinated individuals, who require special attention during future vaccination programs to prevent excess mortality. Key messages • Excess mortality in the three months after (expected) COVID-19 vaccination was high among nonvaccinated individuals and hardly observed amongst vaccinated individuals. • Routine care data could help identify terminally ill individuals, who are overrepresented amongst nonvaccinated individuals, and mitigate the so called healthy vaccinee bias in vaccination research.

Characterization of individuals with atherosclerotic cardiovascular disease, with and without chronic kidney disease: results from a UK electronic health record database

Abstract Introduction Elevated C-reactive protein (CRP) levels are often present in individuals with atherosclerotic cardiovascular disease (ASCVD), including in those who also have chronic kidney disease (CKD), and are associated with an increased risk of major adverse cardiovascular events (MACE). To understand unmet needs in the identification and management of at-risk individuals, we characterized patients with ASCVD in a UK database, according to CKD stage and CRP levels. Methods This study comprised separate retrospective cross-sectional and longitudinal cohort analyses of patients in the Clinical Practice Research Datalink (CPRD) Aurum database, a representative UK database of anonymized primary care records. In the cross-sectional analysis, CPRD Aurum was linked with Hospital Episode Statistics (HES) and Index of Multiple Deprivation data to assess patient characteristics on 1 Jan 2021 (defined as index date). In the cohort analysis, CPRD Aurum was linked with HES and Office for National Statistics mortality data to determine the incidence of 3-point MACE (3P-MACE) over approximately 5 years of follow-up (index date: 1 Jan 2016). Both analyses included individuals ≥ 18 years old with an ASCVD diagnosis in both primary and secondary care in the 5 years pre-index. Data were analysed using descriptive statistics. Patient characteristics were assessed by CKD stage; for individuals with CKD stage 3–4, characteristics were also assessed by CRP levels. Results In total, 225,767 and 230,055 individuals were included in the cross-sectional and cohort analyses, respectively. The prevalence of CKD stage 3–4 among individuals with ASCVD was 24.3% (Table 1). The proportion of individuals with elevated CRP (CRP ≥ 2 mg/L) and anaemia increased for successively higher CKD severity groups; in addition, glycated haemoglobin (HbA1c) ≥ 7% and elevated blood pressure were generally more frequently observed in individuals with CKD than in those without (Table 1). Individuals with ASCVD, CKD stage 3–4 and elevated CRP had higher HbA1c and total cholesterol levels and a greater proportion had obesity, anaemia and elevated blood pressure, compared with individuals with CRP &amp;lt; 2 mg/L (Table 2). Over 5 years’ follow-up, the incidence of 3P-MACE increased with CKD severity (Table 1) and with increasing CRP levels in individuals with ASCVD and CKD stage 3–4 (Table 2). However, only 28% of individuals with ASCVD and CKD stage 3–4 had CRP measurements. Conclusion Adults with ASCVD, CKD and elevated CRP identified from primary and secondary UK care records had a higher disease burden than individuals with CRP levels &amp;lt; 2 mg/L. In the UK, testing CRP levels is not routinely recommended, resulting in a high proportion of missing CRP measurements in this study. Routine testing of CRP may help in identifying patients with residual cardiovascular risk and has the potential to support risk assessment and management of patients with ASCVD and CKD.

Impact of heart failure diagnosis on recorded healthcare resource use by older people in Wales; is primary care shouldering a hidden burden?

Abstract Background Heart failure (HF) is associated with high healthcare resource use and mortality. Most of the evidence focuses on hospital admission data. The impact on primary care interactions and outpatient attendance is less well described, but this information is important in understanding resource utilisation, patient pathways, and in planning services. Purpose To examine how a diagnosis of HF impacts healthcare resource use and outcomes for older people in Wales. Methods We conducted a retrospective, population-level, observational study using linked anonymised electronic health record (EHR) data in Wales (2015-22). 737,230 individuals were aged 65y+ in the study period in primary and secondary care data. Of these, 65,010 had a code for HF in their primary or secondary care records. To evaluate changes in resource use, we excluded 13,990 individuals with fewer than 12 months health data pre- and post-diagnosis. We summarised general practice (GP) interactions, emergency department (ED) attendance, outpatient (OP) attendance, inpatient (IP) episodes and days, time between HF diagnosis and death, and place of death. Results The final cohort comprised 51,020 individuals 65y and over with a diagnosis of HF between 2015-2022. Incidence and prevalence of HF increased across the study period (1.4-1.5% and 8.5-9.1% respectively). Figure 1 shows the percentage change in healthcare resource use for the cohort when HF was diagnosed (12 months pre-and post-diagnosis comparison). Table 1 shows the average number of healthcare utilisation events per person. After a diagnosis of HF, there are fewer ED attendances (-28.2%) and IP admissions (-9.3%) but more OP appointments (31%) and GP interactions (20.3%). When admissions occur, hospital stays tend to be longer (55.4%). Median time (IQR) from HF diagnosis to death from any cause was 247(34-840) days, and with HF listed as a cause 186(25-780) days. In this cohort, the place of death was recorded as NHS establishment (20,379, 67.8%), community (5,450, 18.1%), non-NHS establishment (3,789,12.6%) and unknown (450, 1.5%). Conclusions In Wales, a diagnosis of HF in those 65y+ increases primary care interactions and OP attendances. The change in volume of primary care interactions is considerable (increase from 1,047,920 to 1,315,080). Longer hospital stays may suggest difficulty in managing patients or discharging back to community settings. Majority of those with HF die in NHS establishments. It is unclear whether this reflects deficits in end-of-life care. These results are important in planning effective health service delivery and in establishing baseline data from which to monitor quality improvements.

Lipid management among very-high-risk patients with atherosclerotic cardiovascular disease and type 2 diabetes

Abstract Background Patients with established atherosclerotic cardiovascular disease (ASCVD) and type 2 diabetes (DM2) often meet very-high-risk (VHR) criteria for recurrent ASCVD events and are recommended for aggressive low-density lipoprotein-cholesterol (LDL-C) lowering in guidelines. Purpose Our study aimed to evaluate the proportion of patients achieving guideline-based lipid goals and identify factors associated with achieving these goals among VHR patients with established ASCVD and DM2. Methods Electronic health record data were used to identify VHR patients with established ASCVD and DM2 across 14 large US healthcare systems between 1/1/2022-12/31/2022 with an available LDL-C measurement. We examined the percentage of patients achieving guideline-based lipid goals based on the 2018 AHA/ACC Multisociety Guideline, defined as either having an LDL-C &amp;lt;70 mg/dL or receiving maximal lipid-lowering therapy (i.e., on a PCSK9i monoclonal antibody (mAb)) at the time of the last LDL-C test, and performed multivariable logistic regression to evaluate factors associated with the achievement of those goals. Results Among 213,380 eligible VHR patients with ASCVD and DM2 (mean age 70.2 years, 42% women), 51.8% (110,452) patients achieved guideline-based lipid goals. This included 50.9% with an LDL-C &amp;lt;70 mg/dL and 1.7% on a PCSK9i mAb. Female sex (odds ratio [OR], 0.64; 95% confidence interval [CI], 0.61-0.66), Black race (OR, 0.67; 95% CI, 0.63-0.72), Medicaid insurance (OR, 0.92; 95% CI, 0.86-0.97), and non-coronary artery disease forms of ASCVD including peripheral artery disease (OR, 0.67; 95% CI, 0.65-0.69) and cerebrovascular disease (OR, 0.80; 95% CI, 0.77-0.83) were associated with a lower likelihood of achieving guideline-based lipid goals. Conversely, having undergone percutaneous coronary intervention (OR, 1.31; 95% CI, 1.19-1.43) or coronary artery bypass graft (OR, 1.24; 95% CI, 1.06-1.45) in the previous year, and comorbidities including dialysis (OR, 1.38; 95% CI, 1.23-1.55), atrial fibrillation or flutter (OR, 1.09; 95% CI, 1.06-1.12), and systolic heart failure (OR, 1.11; 95% CI, 1.08-1.14) were associated with a higher likelihood of achieving these goals. Overall, 24.0% of patients were on no statin, 27.2% were on a low- or moderate-intensity statin, and only 5.8% were on a statin in combination with ezetimibe or a PCSK9i mAb. Conclusion Large gaps in care exist for VHR patients with ASCVD and DM2, particularly among women, Black individuals, and those with Medicaid. Targeted interventions are needed to improve the uptake of evidence-based secondary prevention therapies in these patients.

Public trust in national electronic health record systems: A scoping review of qualitative studies

Abstract Background Having public trust in national electronic health record systems (EHRs) is essential for the successful implementation of electronic health records within a country’s healthcare system. Without trust, people are less likely to consent to sharing their health data through an EHR and government stakeholders will have difficulties promoting uptake and facilitation of EHRs. Research investigating public trust in EHRs is limited, however. In response, we conducted a scoping review in order to gain clarity on the influences and themes behind public trust in EHRs, which can support the implementation of EHRs. Methods We reviewed 27 qualitative studies focusing on public trust in EHRs. Studies were identified between January 2022 and June 2022 and were ascertained using an inclusive search method and guided by the PRISMA-ScR checklist. In an iterative process, conceptual themes were derived, and detail the promoters and outcomes of public trust in EHRs. Results Three major themes that describe the promoters of public trust in EHRs and two major themes that detail the outcomes once public trust in EHRs exists were ascertained from the literature. Additional findings bring forth the consequential role of healthcare actors in the public trust building process. Conclusions Results show comprehension, autonomy, and data protection are key concepts that help build and consolidate public trust in EHRs, while the public and the healthcare system benefit from public trust in EHRs. Furthermore, health system actors can be supportive or detrimental to the implementation of EHRs. Their effect is dependent on their activities within the health care system and how the public perceives those activities. Based on: Papadopoulos K, von Wyl V, Gille F. What is public trust in national electronic health record systems? A scoping review of qualitative research studies from 1995 to 2021. DIGITAL HEALTH. 2024;10. doi:10.1177/20552076241228024 Key messages • This research clarifies the conceptualizations that make up the trust relationship between the public and EHRs. • The findings can assist researchers, policymakers, and other health system actors in attaining a better understanding of the intricacies of public trust in EHRs.

Predicting Food Insecurity in a Pediatric Population using the Electronic Health Record

Predicting Food Insecurity in a Pediatric Population using the Electronic Health Record

Exploring Electronic Personal Health Record usage in Northern Italy: a cross-sectional study

Abstract Background The Electronic Personal Health Record (EPHR) is a patient-centered tool that enables citizens to manage their health information. The EPHR has been implemented worldwide, including in Italy, and holds immense potential, but it has shown a generalized underutilisation. Therefore, this paper aims to identify its fulfillment and the reasons behind its poor use, potentially uncovering the key to unlocking its benefits. Methods A survey was distributed on social media and in the waiting rooms of vaccination facilities in a Region of Italy. Multivariable regression models were conducted to detect a connection between the general population’s usage of the EPHR and specific independent variables. Results Of the 1634 respondents, 710 activated the EPHR, especially women, Italians, highly educated, unemployed, with good health but affected with chronic diseases. Amongst those who activated the EPHR, the multivariate analysis showed that the number of accesses to the platform was significantly associated with lower age (adjusted Odds Ratio (adjOR) 0.96 p = 0.015), higher level of education (adjOR 2.17 p = 0.029), unemployment (adjOR 5.61 p = 0.008) and considering the information on the EPHR complete (adjOR 3.83 p = 0.001). As age increases, the perceived ease of access to the EPHR decreases (adjOR=0.98 p = 0.049). Having a higher level of education (adjOR=0.41 p = 0.042), autonomy in reading health materials (adjOR=0.32 p = 0.043), and less self-confidence in filling out medical forms (adjOR=0.61 p = 0.033) were linked to a decrease in perceived ease of access. Conclusions The study showed that the EPHR is more utilised by specific categories of citizens, especially the young and those with higher education. More efforts should be made to increase access to the EPHR for the whole community by making it more intuitive and user-friendly. The present study provides valuable foundations to foster the implementation of a platform for sharing health data at a European level. Key messages • The Electronic Personal Health Record (EPHR) is underutilised despite its potential; it’s mainly accessed by young and highly educated individuals. • Efforts to increase EPHR usage must focus on improving accessibility and ease of access, as well as enhancing user confidence and autonomy in managing health information.

Comorbidity neural network model for personalised recommendations in cardiometabolic risk factor management: towards early precision medicine in routine clinical practice

Abstract Background and Aims Hypertension and hypercholesterolaemia often develop in tandem to confer a risk of cardiovascular disease that may be more than additive. The present study explored early comorbid therapeutic intervention, i.e., beyond individual disease silos, while minimising data bias to provide clinical decision support tools that are portable across clinics. Methods A neural network was created to reproduce UK national clinical guidelines with an accuracy higher than 95%. Relying on real-world evidence from the UK primary care Clinical Practice Research Datalink electronic health record resource, a transfer learning methodology was then applied to achieve recommended risk factor goal attainment after two years. Shapley values derived from game theory were used to define similar digital twin cohorts, and to evaluate individualised recommendations that deviated from national clinical guidelines by rejecting a no-benefit hypothesis retrospectively in a classical proportion test. The methodology was ported to an independent hold-out dataset with anonymized records from six primary care clinics in the South of England. Results The adult patient population (n=469,496, mean+SD age 61+11yr; M 56% vs F 44%; primary 84% vs secondary prevention 16%) included 101,255 diagnosed for hypertension and 112,180 for cholesterol management. While a majority of the 52,504 therapeutic decisions in a test set agree with single morbidity guidelines, for 4639 deviations from the guidance, a significant minority was identified for whom alternative therapy was recommended. Rejecting a no-clinical benefit hypothesis at the 95% confidence level in a retrospective analysis, 1367 decisions showed statistically significant better outcomes over guidelines, for example by escalating anti-hypertensive therapies for patients receiving statin therapy (1085) or by complementing anti-hypertensive therapy with a statin (158). Similar findings were obtained when tested in a hold-out dataset with 13,725 decisions for 6531 patients. Conclusions The results support the contention that machine learning can identify patient subgroups for whom comorbid intervention deviating from clinical guidelines may be appropriate. A novel aspect of the study is the demonstration of clinically portable and actionable application of machine learning in the context of comorbid cardiometabolic diseases. The results represent practical steps towards personalised precision medicine for cardiovascular comorbidity in routine clinical practice.