The Heart Hive: direct-to-participant recruitment provides an effective, scalable route to identify engaged participants for cardiomyopathy research.
Roberts, A. M.; Buchan, R. J.; Theotokia, P.; Powell, G.; Curran, L.; Hammersley, D. J.; Edwards, M.; Jennings, E.; Cavanagh, E.; Cartwright, R.; Tayal, P.; Halliday, B. P.; Prasad, S. K.; Ware, J. S.
Show abstract
BackgroundDirect-to-patient platforms can overcome geographic and capacity barriers to rare-disease research, supporting inclusive and scalable recruitment. The Heart Hive was developed to enable nationwide, genomics-integrated cardiomyopathy research and rapid patient-study matching through a digital, patient-centred platform. MethodsAdults self-reporting a diagnosis, or inherited risk, of cardiomyopathy enrolled online, completed electronic consent, and provided baseline and longitudinal data for study matching. Participants in the pilot Cardiomyopathy Study were invited to return saliva kits by post for central cardiomyopathy-panel sequencing. Engagement and platform development were supported through formal patient and public involvement and partnership with Cardiomyopathy UK. Feasibility of digital recruitment, retention, and genomic validation was assessed. Results338 participants met eligibility; 227 (67%) completed baseline surveys, 195 (86%) returned saliva kits, and 174 were sequenced (98 DCM, 76 HCM). Follow-up survey completion among sequenced participants was 83%. The cohort was 57% female, median age 57.5 years, and 98% White European; 78% had never previously participated in cardiovascular research. Genetic architecture was comparable to a clinic-recruited cohort (DCM 34.7% vs 30.6%; HCM 44.7% vs 38.2%), and actionable variants were identified in 17% of DCM and 29% of HCM participants opting for personal results. ConclusionsSelf-reported recruitment accurately identifies cardiomyopathy cohorts equivalent in genetic composition to clinic-ascertained populations. The Heart Hive demonstrates a decentralised, re-contactable, and registry-ready digital model for nationwide precision-cardiology research.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Evaluation of Women with Peripartum or Dilated Cardiomyopathy and Their First-Degree Relatives: The DCM Precision Medicine Study 94%
- A genotype-phenotype taxonomy of hypertrophic cardiomyopathy 94%
- Sex-Specific Clinical and Genetic Factors Associated with Adverse Outcomes in Hypertrophic Cardiomyopathy 93%
Similar papers in this journal
- ClinGen Hereditary Cardiovascular Disease Gene Curation Expert Panel: Reappraisal of Genes associated with Hypertrophic Cardiomyopathy 94%
- Comparative Effectiveness of Second-line Antihyperglycemic Agents for Cardiovascular Outcomes: A Large-scale, Multinational, Federated Analysis of the LEGEND-T2DM Study 92%
- Assessment of valvular function in over 47,000 people using deep learning-based flow measurements 92%
Similar papers in this journal
- Development and Multinational Validation of an Ensemble Deep Learning Algorithm for Detecting and Predicting Structural Heart Disease Using Noisy Single-lead Electrocardiograms 94%
- Clinical and genetic associations of asymmetric apical and septal left ventricular hypertrophy 93%
- Simple Models Versus Deep Learning in Detecting Low Ejection Fraction From The Electrocardiogram 93%
Similar papers in this journal
- Development and validation of imaging-free myocardial fibrosis prediction models, association with outcomes, and sample size estimation for phase 3 trials 93%
- Hypertrophic Cardiomyopathy -- The Impact of Age at Diagnosis of the Proband on Genetic Yield, Clinical Presentation, Outcomes, and Yield of Family Screening 93%
- Mineralocorticoid Receptor Antagonism Reduces Atrial Arrhythmias Post-Cardiac Surgery and Attenuates Atrial Stress Responses to Cardioplegic Arrest 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.