Context

The choice of therapy in individual atrial fibrillation (AF) patients is largely dependent on AF subtype and present and predicted future symptom burden and morbidity, while the indication for anticoagulation treatment is closely related to stroke risk factors. Treatment options include rate controlling pharmacological agents, electrical and pharmacological cardioversion (CV), prophylactic antiarrhythmic medication (AAM) and catheter ablation procedures (Kirchhof 2016).

A central challenge in managing paroxysmal atrial fibrillation (PAF) and persistent atrial fibrillation (PeAF) is predicting future risks of morbidity and mortality. Accurate prediction models are crucial for personalizing treatment plans and improving patient outcomes.

Objectives

Develop a survival prediction model for morbi-mortality risk in patients with PAF and PeAF using statistical modeling techniques.

Data

Data from the “Atrial Fibrillation Survey–Copenhagen (ATLAS-CPH)” collected between January 1st, 2008 and December 1st, 2012 from both in- and outpatient clinics at the Department of Cardiology, University Hospital Copenhagen, Hvidovre, Denmark.

Inclusion Criteria:

PAF was defined by spontaneous conversion to sinus rhythm. PeAF was defined by episodes > 7 days or requiring medical/electrical cardioversion.

Out of 189 patients meeting the inclusion criteria, 15 were excluded (13 due to invasive ablation, 2 due to estimated survival <1 year). A total of 174 patients were included in the study, all of Caucasian ethnicity. The mean follow-up duration was 1279 days, contributing 222,459 person-days.

Dataset Access: All data are available here (Schroder et al., 2019).

Methodology

Requirements

In your analysis, justify the following choices:

  • The choice of endpoint(s): Why were certain endpoints selected for survival analysis?
  • The model(s): What survival models will you use, and why?
  • Model evaluation and validation: How will you validate the model ?
  • Handling missing data and covariates: What methods will you use to address missing data, and why?
  • Model assumptions: What assumptions are made about the data and how do you plan to check for them?
  • …

Deliverables

Please submit the following:

  • A slide deck focusing on methodology, experiments, and results. Do not include data description in your presentation; focus on the rationales behind your methodological choices.
  • 10-minute oral presentation (+5 minutes for questions). The focus should be on the rationale for your methodological approach, the experimental process, and key findings.
  • Code (Rmd or Jupyter notebook format) to be submitted by the morning of the exam date.

References

Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Europace. 2016. pp. 1609–1678. pmid:27567465

Geraets DR. Atrial fibrillation and atrial flutter. In: Clinical Pharmacy [Internet]. 1993 p. 721. Available: https://www.nbv.cardio.dk/af

Schroder, J., Bouaziz, O., Agner, B. R., Martinussen, T., Madsen, P. L., Li, D., & Dixen, U. (2019). Recurrent event survival analysis predicts future risk of hospitalization in patients with paroxysmal and persistent atrial fibrillation. Plos one, 14(6), e0217983.