Assessing Measures of Atrial Fibrillation Clustering via Stochastic Models of Episode Recurrence and Disease Progression

TL;DR

This paper analyzes atrial fibrillation (AF) recurrence patterns using stochastic models to develop measures of disease progression, aiming to improve clinical decision-making despite current model limitations.

Contribution

It introduces a stochastic modeling approach to analyze AF episode recurrence and proposes developing robust measures of disease progression for clinical use.

Findings

Initial analysis of AF density using a simplified stochastic model

Highlights limitations of current high-resolution models for long-term data

Lays groundwork for future development of clinical progression measures

Abstract

Atrial fibrillation (AF) is a leading cause of morbidity and mortality. AF prevalence increases with age, which is attributed to pathophysiological changes that aid AF initiation and perpetuation. Current state-of-the-art models are only capable of simulating short periods of atrial activity at high spatial resolution, whilst the majority of clinical recordings are based on infrequent temporal datasets of limited spatial resolution. Being able to estimate disease progression informed by both modelling and clinical data would be of significant interest. In addition an analysis of the temporal distribution of recorded fibrillation episodes AF density can provide insights into recurrence patterns. We present an initial analysis of the AF density measure using a simplified idealised stochastic model of a binary time series representing AF episodes. The future aim of this work is to develop robust clinical measures of progression which will be tested on models that generate long-term synthetic data. These measures would then be of clinical interest in deciding treatment strategies.