Framework-based qualitative analysis of free responses of Large Language Models: Algorithmic fidelity.

Today, with the advent of Large-scale generative Language Models (LLMs) it is now possible to simulate free responses to interview questions such as those traditionally analyzed using qualitative research methods. Qualitative methodology encompasses a broad family of techniques involving manual analysis of open-ended interviews or conversations conducted freely in natural language. Here we consider whether artificial "silicon participants" generated by LLMs may be productively studied using qualitative analysis methods in such a way as to generate insights that could generalize to real human populations. The key concept in our analysis is algorithmic fidelity, a validity concept capturing the degree to which LLM-generated outputs mirror human sub-populations' beliefs and attitudes. By definition, high algorithmic fidelity suggests that latent beliefs elicited from LLMs may generalize to real humans, whereas low algorithmic fidelity renders such research invalid. Here we used an LLM to generate interviews with "silicon participants" matching specific demographic characteristics one-for-one with a set of human participants. Using framework-based qualitative analysis, we showed the key themes obtained from both human and silicon participants were strikingly similar. However, when we analyzed the structure and tone of the interviews we found even more striking differences. We also found evidence of a hyper-accuracy distortion. We conclude that the LLM we tested (GPT-3.5) does not have sufficient algorithmic fidelity to expect in silico research on it to generalize to real human populations. However, rapid advances in artificial intelligence raise the possibility that algorithmic fidelity may improve in the future. Thus we stress the need to establish epistemic norms now around how to assess the validity of LLM-based qualitative research, especially concerning the need to ensure the representation of heterogeneous lived experiences.

Wearable Sensors to Monitor Physical Activity in Heart Failure Clinical Trials: State-of-the-Art Review.

BACKGROUND: Estimation of the effects that drugs or other interventions have on patients' symptoms and functions is crucial in heart failure trials. Traditional symptoms and functions clinical outcome assessments have important limitations. Actigraphy may help to overcome these limitations due to its objective nature and the potential for continuous recording of data. However, actigraphy is not currently accepted as clinically relevant by key stakeholders. METHODS AND RESULTS: In this state-of-the-art study, the key aspects to consider when implementing actigraphy in heart failure trials are discussed. They include which actigraphy-derived measures should be considered, how to build endpoints using them, how to measure and analyze them, and how to handle the patients' and sites' logistics of integrating devices into trials. A comprehensive recommendation based on the current evidence is provided. CONCLUSION: Actigraphy is technically feasible in clinical trials involving heart failure, but successful implementation and use to demonstrate clinically important differences in physical functioning with drug or other interventions require careful consideration of many design choices.

Estimating the clinical and budgetary impact of using angiotensin receptor neprilysin inhibitor as first line therapy in patients with HFrEF.

AIMS: Recent updates of international treatment guidelines for heart failure with reduced ejection fraction (HFrEF) differ regarding the use of angiotensin receptor neprilysin inhibitor (ARNI) as first-line treatment. The American Heart Association/American College of Cardiology/Heart Failure Society of America (AHA/ACC/HFSA) 2022 guidelines gives ARNI a Class IA recommendation for HFrEF patients while the European Society of Cardiology's guidelines are less clear when ARNI could be considered as first line treatment option in de novo patients. This study aimed to model the clinical and budgetary outcomes of implementing these guidelines, comparing conservative ARNI prescription patterns with less conservative in Sweden and in the United Kingdom. METHODS AND RESULTS: A health economic model was developed to compare different treatment patterns for HFrEF. Incident cohorts were included on an annual basis and followed over 10 years. The model included treatment specific all-cause mortality and hospitalization rates, as well as drug acquisition, monitoring, and hospitalization costs. Increasing the use of ARNI could lead to about 7000-12 300 life years gained and 2600-4600 hospitalizations prevented in Sweden. These health benefits come with an additional cost of 112-195 million euros. Similar results were estimated for the United Kingdom, albeit on a larger population. CONCLUSIONS: Increasing the proportion of patients receiving ARNI instead of angiotensin converting enzyme inhibitors as first-line treatment of HFrEF will lead to a considerable number of additional life years gained and prevented hospitalizations but with additional cost in terms of health care expenditure in Sweden and in the United Kingdom.

Integration of implantable device therapy in patients with heart failure. A clinical consensus statement from the Heart Failure Association (HFA) and European Heart Rhythm Association (EHRA) of the European Society of Cardiology (ESC).

Implantable devices form an integral part of the management of patients with heart failure (HF) and provide adjunctive therapies in addition to cornerstone drug treatment. Although the number of these devices is growing, only few are supported by robust evidence. Current devices aim to improve haemodynamics, improve reverse remodelling, or provide electrical therapy. A number of these devices have guideline recommendations and some have been shown to improve outcomes such as cardiac resynchronization therapy, implantable cardioverter-defibrillators and long-term mechanical support. For others, more evidence is still needed before large-scale implementation can be strongly advised. Of note, devices and drugs can work synergistically in HF as improved disease control with devices can allow for further optimization of drug therapy. Therefore, some devices might already be considered early in the disease trajectory of HF patients, while others might only be reserved for advanced HF. As such, device therapy should be integrated into HF care programmes. Unfortunately, implementation of devices, including those with the greatest evidence, in clinical care pathways is still suboptimal. This clinical consensus document of the Heart Failure Association (HFA) and European Heart Rhythm Association (EHRA) of the European Society of Cardiology (ESC) describes the physiological rationale behind device-provided therapy and also device-guided management, offers an overview of current implantable device options recommended by the guidelines and proposes a new integrated model of device therapy as a part of HF care.

Effect of implementing a heart failure admission care bundle on hospital readmission and mortality rates: interrupted time series study.

This study aimed to evaluate the impact of developing and implementing a care bundle intervention to improve care for patients with acute heart failure admitted to a large London hospital. The intervention comprised three elements, targeted within 24 hours of admission: N-terminal pro-B-type natriuretic peptide (NT-proBNP) test, transthoracic Doppler two-dimensional echocardiography and specialist review by cardiology team. The SHIFT-Evidence approach to quality improvement was used. During implementation, July 2015-July 2017, 1169 patients received the intervention. An interrupted time series design was used to evaluate impact on patient outcomes, including 15 618 admissions for 8951 patients. Mixed-effects multiple Poisson and log-linear regression models were fitted for count and continuous outcomes, respectively. Effect sizes are slope change ratios pre-intervention and post-intervention. The intervention was associated with reductions in emergency readmissions between 7 and 90 days (0.98, 95% CI 0.97 to 1.00), although not readmissions between 0 and 7 days post-discharge. Improvements were seen in in-hospital mortality (0.96, 95% CI 0.95 to 0.98), and there was no change in trend for hospital length of stay. Care process changes were also evaluated. Compliance with NT-proBNP testing was already high in 2014/2015 (162 of 163, 99.4%) and decreased slightly, with increased numbers audited, to 2016/2017 (1082 of 1101, 98.2%). Over this period, rates of echocardiography (84.7-98.9%) and specialist input (51.6-90.4%) improved. Care quality and outcomes can be improved for patients with acute heart failure using a care bundle approach. A systematic approach to quality improvement, and robust evaluation design, can be beneficial in supporting successful improvement and learning.

Circulating miR-133a-3p defines a low-risk subphenotype in patients with heart failure and central sleep apnea: a decision tree machine learning approach.

BACKGROUND: Patients with heart failure with reduced ejection fraction (HFrEF) and central sleep apnea (CSA) are at a very high risk of fatal outcomes. OBJECTIVE: To test whether the circulating miRNome provides additional information for risk stratification on top of clinical predictors in patients with HFrEF and CSA. METHODS: The study included patients with HFrEF and CSA from the SERVE-HF trial. A three-step protocol was applied: microRNA (miRNA) screening (n = 20), technical validation (n = 60), and biological validation (n = 587). The primary outcome was either death from any cause, lifesaving cardiovascular intervention, or unplanned hospitalization for worsening of heart failure, whatever occurred first. MiRNA quantification was performed in plasma samples using miRNA sequencing and RT-qPCR. RESULTS: Circulating miR-133a-3p levels were inversely associated with the primary study outcome. Nonetheless, miR-133a-3p did not improve a previously established clinical prognostic model in terms of discrimination or reclassification. A customized regression tree model constructed using the Classification and Regression Tree (CART) algorithm identified eight patient subphenotypes with specific risk patterns based on clinical and molecular characteristics. MiR-133a-3p entered the regression tree defining the group at the lowest risk; patients with log(NT-proBNP) ≤ 6 pg/mL (miR-133a-3p levels above 1.5 arbitrary units). The overall predictive capacity of suffering the event was highly stable over the follow-up (from 0.735 to 0.767). CONCLUSIONS: The combination of clinical information, circulating miRNAs, and decision tree learning allows the identification of specific risk subphenotypes in patients with HFrEF and CSA.

Predictors of outcome in a contemporary cardiac sarcoidosis population: Role of brain natriuretic peptide, left ventricular function and myocardial inflammation.

AIMS: Cardiac sarcoidosis (CS) is a potentially fatal condition that varies in its clinical presentation. Here, we describe baseline characteristics at presentation along with prognosis and predictors of outcome in a sizable and deeply phenotyped contemporary cohort of CS patients. METHODS AND RESULTS: Consecutive CS patients seen at one institution were retrospectively enrolled after undergoing laboratory testing, electrocardiogram, echocardiography, cardiac magnetic resonance (CMR) imaging and 18 F-flourodeoxyglucose positron emission tomography (FDG-PET) at baseline. The composite endpoint consisted of all-cause mortality, aborted sudden cardiac death, major ventricular arrhythmic events, heart failure hospitalization and heart transplantation. A total of 319 CS patients were studied (67% male, 55.4 ± 12 years). During a median follow-up of 2.2 years (range: 1 month-11 years), 8% of patients died, while 33% reached the composite endpoint. The annualized mortality rate was 2.7% and the 5- and 10-year mortality rates were 6.2% and 7.5%, respectively. Multivariate analysis showed serum brain natriuretic peptide (BNP) levels (hazard ratio [HR] 2.41, 95% confidence interval [CI] 1.34-4.31, p = 0.003), CMR left ventricular ejection fraction (LVEF) (HR 0.96, 95% CI 0.94-0.98, p < 0.0001) and maximum standardized uptake value of FDG-PET (HR 1.11, 95% CI 1.04-1.19, p = 0.001) to be independent predictors of outcome. These findings remained robust for different patient subgroups. CONCLUSION: Cardiac sarcoidosis is associated with significant morbidity and mortality, particularly in those with cardiac involvement as the first manifestation. Higher BNP levels, lower LVEF and more active myocardial inflammation were independent predictors of outcomes.

The cost-effectiveness of NT-proBNP for assessment of suspected acute heart failure in the emergency department.

AIMS: When relying on clinical assessment alone, an estimated 22% of acute heart failure (AHF) patients are missed, so clinical guidelines recommend the use of N-terminal pro-B-type natriuretic peptide (NT-proBNP) for AHF diagnosis. Since publication of these guidelines, there has been poor uptake of NT-proBNP testing in part due to concerns over excessive false positive referrals resulting from the low specificity of a single 'rule-out' threshold of <300 pg/mL. Low specificity can be mitigated by the addition of age-specific 'rule-in' NT-proBNP thresholds. METHODS AND RESULTS: A theoretical hybrid decision tree/semi-Markov model was developed, combining global trial and audit data to evaluate the cost-effectiveness of NT-proBNP testing using age-specific rule-in/rule-out (RI/RO) thresholds, compared with NT-proBNP RO only and with clinical decision alone (CDA). Cost-effectiveness was measured as the incremental cost per quality-adjusted life year (QALY) gained and incremental net health benefit. In the base case, using UK-specific inputs, NT-proBNP RI/RO was associated with both greater QALYs and lower costs than CDA. At a willingness-to-pay threshold of £20 000/QALY, NT-proBNP RO was also cost-effective compared with CDA [incremental cost-effectiveness ratio (ICER) of £8322/QALY], but not cost-effective vs. RI/RO (ICER of £64 518/QALY). Overall, NT-proBNP RI/RO was the most cost-effective strategy. Sensitivity and scenario analyses were undertaken; the conclusions were not impacted by plausible variations in parameters, and similar conclusions were obtained for the Netherlands and Spain. CONCLUSIONS: An NT-proBNP strategy that combines an RO threshold with age-specific RI thresholds provides a cost-effective alternative to the currently recommended NT-proBNP RO only strategy, achieving greater diagnostic specificity with minimal reduction in sensitivity and thus reducing unnecessary echocardiograms and hospital admissions.

Associations between baseline heart rate and blood pressure and time to events in heart failure with reduced ejection fraction patients: Data from the QUALIFY international registry.

AIMS: A high resting heart rate (RHR) and low systolic blood pressure (SBP) are a risk factor and a risk indicator, respectively, for poor heart failure (HF) outcomes. This analysis evaluated the associations between baseline RHR and SBP with outcomes and treatment patterns in patients with HF and reduced ejection fraction (HFrEF) in the QUALIFY (QUality of Adherence to guideline recommendations for LIFe-saving treatment in heart failure surveY) international registry. METHODS AND RESULTS: Between September 2013 and December 2014, 7317 HFrEF patients with a previous HF hospitalization within 1-15 months were enrolled in the QUALIFY registry. Complete follow-up data were available for 5138 patients. The relationships between RHR and SBP and outcomes were assessed using a Cox proportional hazards model and were analysed according to baseline values as high RHR (H-RHR) ≥75 bpm versus low RHR (L-RHR) <75 bpm and high SBP (H-SBP) ≥110 mmHg versus low SBP (L-SBP) <110 mmHg and analysed according to each of the following four phenotypes: H-RHR/L-SBP, L-RHR/L-SBP, H-RHR/H-SBP and L-RHR/H-SBP (reference group). Compared to the reference group, H-RHR/L-SBP was associated with the worst outcomes for the combined primary endpoint of cardiovascular death and HF hospitalization (hazard ratio [HR] 1.83, 95% confidence interval [CI] 1.51-2.21, p < 0.001), cardiovascular death (HR 2.70, 95% CI 1.69-4.33, p < 0.001), and HF hospitalization (HR 1.62, 95% CI 1.30-2.01, p < 0.001). Low-risk patients with L-RHR/H-SBP achieved more frequently ≥50% of target doses of angiotensin-converting enzyme inhibitors (ACEIs) and beta-blockers (BBs) than the other groups. However, 48% and 46% of low-risk patients were not well treated with ACEIs and BBs, respectively (≤50% of target dose or no treatment). CONCLUSION: In patients with HFrEF and recent hospitalization, elevated RHR and lower SBP identify patients at increased risk for cardiovascular endpoints. While SBP and RHR are often recognized as barriers that deter physicians from treating with high doses of recommended drugs, they are not the only reason leaving many patients suboptimally treated.

Real-time pulmonary artery pressure monitoring in heart failure patients: an updated cost-effectiveness analysis.

AIMS: Previous cost-effectiveness analysis suggests that CardioMEMS is cost-effective compared with usual care for patients with persistent New York Heart Association class III symptoms and at least one heart failure (HF) hospitalization within 12 months. The aim of the paper is to perform an update of the cost-effectiveness analysis of CardioMEMS using the most recent data from the published literature. METHODS AND RESULTS: A Microsoft Excel Markov model from a previous UK cost-effectiveness study of CardioMEMS was updated using the clinical effectiveness of pulmonary artery pressure (PAP)-guided treatment derived from the pivotal trials. The model included the device costs (and the implantation procedure and related complications), costs of remote monitoring, costs of HF-related hospitalizations, and costs of usual care. Quality-adjusted life years (QALYs) were estimated based on utilities from pivotal trials and published literature. Cost-effectiveness results were estimated as incremental cost per QALY gained of CardioMEMS compared with usual care. Scenario analyses were also performed using data from real-world studies that showed a significant decrease in HF-related hospitalizations. In the base case analysis over a time horizon of 10 years, PAP-guided HF therapy increased cost compared with usual care by £6337 (i.e. from £22 770 in usual care to £29 107 in PAP-guided HF therapy) and the QALYs per patient for usual care and PAP-guided patients were 2.62 and 2.94, respectively, reflecting an increase of 0.32 QALYs with PAP-guided treatment. The resultant incremental cost-effectiveness ratio (ICER), the ratio between incremental costs and the QALYs, is estimated at £19 761/QALY. Scenario analyses suggest that the ICER for CardioMEMS can range from being dominant to £27 910/QALY. Probabilistic sensitivity analyses suggested that PAP-guided HF therapy has 81.9% probability of being cost-effective at a threshold of £30 000/QALY. CONCLUSIONS: Our model suggests that CardioMEMS is likely to be cost-effective in the United Kingdom, at the currently considered thresholds of £20 000-30 000/QALY.

Placing patient-reported outcomes at the centre of cardiovascular clinical practice: implications for quality of care and management.

Patient-reported outcomes (PROs) provide important insights into patients' own perspectives about their health and medical condition, and there is evidence that their use can lead to improvements in the quality of care and to better-informed clinical decisions. Their application in cardiovascular populations has grown over the past decades. This statement describes what PROs are, and it provides an inventory of disease-specific and domain-specific PROs that have been developed for cardiovascular populations. International standards and quality indices have been published, which can guide the selection of PROs for clinical practice and in clinical trials and research; patients as well as experts in psychometrics should be involved in choosing which are most appropriate. Collaborations are needed to define criteria for using PROs to guide regulatory decisions, and the utility of PROs for comparing and monitoring the quality of care and for allocating resources should be evaluated. New sources for recording PROs include wearable digital health devices, medical registries, and electronic health record. Advice is given for the optimal use of PROs in shared clinical decision-making in cardiovascular medicine, and concerning future directions for their wider application.

Benefits and challenges of telemedicine for heart failure consultations: a qualitative study.

BACKGROUND: Prior to the Covid-19 pandemic, heart failure (HF) disease management programmes were predominantly delivered in-person, with telemedicine being uncommon. Covid-19 resulted in a rapid shift to "remote-by-default" clinic appointments in many organisations. We evaluated clinician and patient experiences of teleconsultations for HF. METHODS: From 16th March 2020, all HF appointments at a specialist centre in the UK were telemedicine-by-default through a mixture of telephone and video consultations, with rare in-person appointments. HF clinicians and patients with HF were invited to participate in semi-structured interviews about their experiences. A purposive sampling technique was used. Interviews were conducted using Microsoft Teams®, recorded and transcribed verbatim. Narrative data were explored by thematic analysis. Clinicians and patients were interviewed until themes saturated. RESULTS: Eight clinicians and eight patients with HF were interviewed before themes saturated. Five overarching themes emerged: 1) Time utilisation - telemedicine consultations saved patients time travelling to and waiting for appointments. Clinicians perceived them to be more efficient, but more administrative time was involved. 2) Clinical assessment - without physical examination, clinicians relied more on history, observations and test results; video calls were perceived as superior to telephone calls for remote assessment. Patients confident in self-monitoring tended to be more comfortable with telemedicine. 3) Communication and rapport - clinicians experienced difficulty establishing rapport with new patients by telephone, though video was better. Patients generally did not perceive that remote consultation affected their rapport with clinicians. 4) Technology - connection issues occasionally disrupted video consultations, but overall patients and clinicians found the technology easy to use. 5) Choice and flexibility - both patients and clinicians believed that the choice of modality should be situation-dependent. CONCLUSIONS: Telemedicine HF consultations were more convenient for patients, saved them time, and were generally acceptable to clinicians, but changed workflows, consultation dynamics, and how clinical assessment was performed. Telemedicine should be used alongside in-person appointments in a "hybrid" model tailored to individual patients and settings.

Personalized care of patients with heart failure: are we ready for a REWOLUTION? Insights from two international surveys on healthcare professionals' needs and patients' perceptions.

AIMS: Guidelines for the management of heart failure (HF) are evolving, and increasing emphasis is placed on patient-centred care. As part of the REWOLUTION HF (REal WOrLd EdUcaTION in HF) programme, we conducted two international surveys aimed at assessing healthcare professionals' (HCPs) educational needs and patients' perspectives on the care of HF. METHODS AND RESULTS: Anonymous online questionnaires co-developed by HF experts and patients assessed HCPs' educational needs (520 respondents, mostly cardiologists, in 67 countries) and patients' perceptions on HF impact and management (98 respondents in 18 countries). Among HCPs, 62.7% prioritized rapid initiation of all guideline-mandated medications over up-titration of some medications, and 87.7% always or frequently discussed treatment goals with patients. There was good agreement between HCPs and patients on key treatment goals, except for a greater emphasis on reducing hospitalizations among HCPs. The most frequently cited barriers to the provision of guideline-recommended pharmacological therapy were treatment side effects/intolerance, complex treatment regimens, low blood pressure, cost/reimbursement issues, and low estimated glomerular filtration rate. Most patients (81.6%) reported no difficulties taking medications as prescribed, although 21.4% felt they were taking too many pills. Patients wanted more information about HF and its consequences, prognosis, and treatments (70.4%, 74.5% and 76.6%, respectively). Cardiologists were the preferred source of information about HF, followed by general practitioners and HF nurses. CONCLUSIONS: These surveys provide valuable insights into HCPs' needs about personalized care for patients with HF, as well as patients' perceptions, expectations and preferences. These findings will be helpful to develop patient-centred, needs-driven quality improvement programmes.

Changes in the Inferior Vena Cava Are More Sensitive Than Venous Pressure During Fluid Removal: A Proof-of-Concept Study.

BACKGROUND: Congestion is central to the pathophysiology of heart failure (HF); thus, tracking congestion is crucial for the management of patients with HF. In this study we aimed to compare changes in inferior vena cava diameter (IVCD) with venous pressure following manipulation of volume status during ultrafiltration in patients with cardiac dysfunction. METHODS AND RESULTS: Patients with stable hemodialysis and with systolic or diastolic dysfunction were studied. Central venous pressure (CVP) and peripheral venous pressure (PVP) were measured before and after hemodialysis. IVCD and PVP were measured simultaneously just before dialysis, 3 times during dialysis and immediately after dialysis. Changes in IVCD and PVP were compared at each timepoint with ultrafiltration volumes. We analyzed 30 hemodialysis sessions from 20 patients. PVP was validated as a surrogate for CVP. Mean ultrafiltration volume was 2102 ± 667 mL. IVCD discriminated better ultrafiltration volumes ≤ 500 mL or ≤ 750 mL than PVP (AUC 0.80 vs 0.62, and 0.80 vs 0.56, respectively; both P< 0.01). IVCD appeared to track better ultrafiltration volume (P< 0.01) and hemoconcentration (P< 0.05) than PVP. Changes in IVCD were of greater magnitude than those of PVP (average change from predialysis: -58 ± 30% vs -28 ± 21%; P< 0.001). CONCLUSIONS: In patients undergoing ultrafiltration, changes in IVCD tracked changes in volume status better than venous pressure.

Risk prediction of mortality for patients with heart failure in England: observational study in primary care.

AIMS: Many risk prediction models have been proposed for heart failure (HF), but few studies have used only information available to general practitioners (GPs) in primary care electronic health records (EHRs). We describe the predictors and performance of models built from GP-based EHRs in two cohorts of patients 10 years apart. METHODS AND RESULTS: Linked primary and secondary care data for incident HF cases in England were extracted from the Clinical Practice Research Datalink for 2001-02 and 2011-12. Time-to-event models for all-cause mortality were developed using a long list of potential baseline predictors. Discrimination and calibration were calculated. A total of 5966 patients in 156 general practices were diagnosed in 2001-02, and 12 827 patients in 331 practices were diagnosed in 2011-12. The 5-year survival rate was 40.0% in 2001-02 and 40.2% in 2011-12, though the latter population were older, frailer, and more comorbid; for 2001-02, the 10-year survival was 20.8% and 15-year survival 11.1%. Consistent predictors included age, male sex, systolic blood pressure, body mass index, GP domiciliary visits before diagnosis, and some comorbidities. Model performance for both time windows was modest (c = 0.70), but calibration was generally excellent in both time periods. CONCLUSIONS: Information routinely available to UK GPs at the time of diagnosis of HF gives only modest predictive accuracy of all-cause mortality, making it hard to decide on the type, place, and urgency of follow-up. More consistent recording of data relevant to HF (such as echocardiography and natriuretic peptide results) in GP EHRs is needed to support accurate prediction of healthcare needs in individuals with HF.

CODE-EHR best practice framework for the use of structured electronic healthcare records in clinical research.

Big data is central to new developments in global clinical science aiming to improve the lives of patients. Technological advances have led to the routine use of structured electronic healthcare records with the potential to address key gaps in clinical evidence. The covid-19 pandemic has demonstrated the potential of big data and related analytics, but also important pitfalls. Verification, validation, and data privacy, as well as the social mandate to undertake research are key challenges. The European Society of Cardiology and the BigData@Heart consortium have brought together a range of international stakeholders, including patient representatives, clinicians, scientists, regulators, journal editors and industry. We propose the CODE-EHR Minimum Standards Framework as a means to improve the design of studies, enhance transparency and develop a roadmap towards more robust and effective utilisation of healthcare data for research purposes.

World Heart Federation Roadmap for Digital Health in Cardiology.

More than 500 million people worldwide live with cardiovascular disease (CVD). Health systems today face fundamental challenges in delivering optimal care due to ageing populations, healthcare workforce constraints, financing, availability and affordability of CVD medicine, and service delivery. Digital health technologies can help address these challenges. They may be a tool to reach Sustainable Development Goal 3.4 and reduce premature mortality from non-communicable diseases (NCDs) by a third by 2030. Yet, a range of fundamental barriers prevents implementation and access to such technologies. Health system governance, health provider, patient and technological factors can prevent or distort their implementation. World Heart Federation (WHF) roadmaps aim to identify essential roadblocks on the pathway to effective prevention, detection, and treatment of CVD. Further, they aim to provide actionable solutions and implementation frameworks for local adaptation. This WHF Roadmap for digital health in cardiology identifies barriers to implementing digital health technologies for CVD and provides recommendations for overcoming them.

Prioritising electrocardiograms for manual review to improve the efficiency of atrial fibrillation screening.

Screening for atrial fibrillation (AF) could reduce the incidence of stroke by identifying undiagnosed AF and prompting anticoagulation. However, screening may involve recording many electrocardiograms (ECGs) from each participant, several of which require manual review which is costly and time-consuming. The aim of this study was to investigate whether the number of ECG reviews could be reduced by using a model to prioritise ECGs for review, whilst still accurately diagnosing AF. A multiple logistic regression model was created to estimate the likelihood of an ECG exhibiting AF based on the mean RR-interval and variability in RR-intervals. It was trained on 1,428 manually labelled ECGs from 144 AF screening programme participants, and evaluated using 11,443 ECGs from 1,521 participants. When using the model to order ECGs for review, the number of reviews for AF participants was reduced by 74% since no further reviews are required after an AF ECG is identified; however, it did not impact the number of reviews in non-AF participants (the vast majority of participants), so the overall number of reviews was reduced by 3% with no missed AF diagnoses. When using the model to also exclude ECGs from review, the overall number of reviews was reduced by 28% with no missed AF diagnoses, and by 53% with only 4% of AF diagnoses missed. In conclusion, the workload can be reduced by using a model to prioritise ECGs for review. Ordering ECGs alone only provides only a moderate reduction in workload. The additional use of a threshold to exclude ECGs from review provides a much greater reduction in workload at the expense of some missed AF diagnoses. Clinical Relevance-This shows the potential benefit of using a model to prioritise electrocardiograms for review in order to reduce the manual workload of AF screening.