E-Thesis 61 views
Developing Population-Level Methods for Cardiovascular Risk Management in Atrial Fibrillation / FATEMEH TORABI
Swansea University Author: FATEMEH TORABI
DOI (Published version): 10.23889/SUThesis.71746
Abstract
Background: Stroke is a critical medical emergency associated with substantial lifelong complications and economic burden. Atrial Fibrillation (AF) escalates stroke risk by five-fold. This thesis used comprehensive population-level data to develop methods for assessment of cardiovascular risk manage...
| Published: |
Swansea
2026
|
|---|---|
| Institution: | Swansea University |
| Degree level: | Doctoral |
| Degree name: | Ph.D |
| Supervisor: | Gravenor, M., and Halcox, J. |
| URI: | https://cronfa.swan.ac.uk/Record/cronfa71746 |
| Abstract: |
Background: Stroke is a critical medical emergency associated with substantial lifelong complications and economic burden. Atrial Fibrillation (AF) escalates stroke risk by five-fold. This thesis used comprehensive population-level data to develop methods for assessment of cardiovascular risk management in a large AF population. Methods: A range of advance methods were developed using a comprehensive national AF cohort achieved through linkage of major national datasets including primary care, secondary care and drug dispensing dataset from SAIL Databank, enabling retrospective observational cohort studies.Results: Assessment of anticoagulation (AC) therapy between 2012 and 2018 revealed that achieving 90% AC coverage significantly reduced risk of stroke and systemic embolism (SSE). However, a notable treatment gap persisted, representing a missed opportunity for preventive care over the study period. Mathematical modelling of future scenarios demonstrated the opportunity for substantial (9%) reductions in hospitalisation rates if guideline targets could be met.Further investigations focused on international normalised ration (INR) control in warfarin-treated AF patients. Time-varying analysis of the cohort, revealed 34% of time was spent under poor INR control levels (according to NICE guidelines). Statistical analysis revealed poor INR control was significantly associated with increased risk of both SSE (Hazard Ratio=1.69, 95% CI: 1.54=, 1.86) and bleeding (HR=1.40, 95% CI: 1.33,1.48), providing additional prognostic information inaddition to recognised risk factors. The tracking of cardiovascular pharmaceutical dispensing was extended during the pandemic to all medications to investigate the impact of COVID-19 treatment patterns across over 70 million drug items, using the all-Wales population cohort. A 1% declinein overall dispensing was noted in 2020 compared to 2019, with notable disruption of cardiovascular medication across all age-groups. Finally, a validation study of two conventional bleeding risk scores, HAS-BLED and ORBIT, was conducted in a cohort of over 100,000 AF patients. Predicted bleeding rates were compared to those observed in the cohort, and stratified by bleed type and anticoagulation regime. Both scores had limited predictive performance (61% and 62%). Although performance was good for low-risk and aspirin treated patients, HAS-BLED tended to overestimate, and ORBIT underestimate, the bleed risk for DOAC and vitamin K treated populations. Conclusion: This thesis strengthens the case for optimising clinical outcomes and cardiovascular risk management in AF by applying pharmacoepidemiologic approaches to improve adherence to evidence-based therapies and enhanced risk stratification, harnessing longitudinal population-level patient data. |
|---|---|
| Keywords: |
Cardiovascular risk management, Atrial Fibrillation, Advance modelling, Predictive modelling |
| College: |
Faculty of Medicine, Health and Life Sciences |