Lookup NU author(s): Emeritus Professor Richard Thomson,
Professor Martin Eccles,
Dr David Chinn
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Background. Stroke risk in nonvalvular atrial fibrillation can be reduced by warfarin or aspirin; the choice of therapy requires the assessment of risks and benefits. The authors compared methods of risk assessment and their implications for risk communication and treatment. Methods. Stroke risk was compared in 193 patients with atrial fibrillation using the Framingham equation; an atrial fibrillation-specific Framingham equation; the Congestive heart failure, Hypertension, Age, Diabetes and Stroke (CHADS2) score; the Stroke Prevention and Atrial Fibrillation (SPAF) scheme; and the Scottish Intercollegiate Guidelines Network (SIGN) guidelines. Treatment guidance from SIGN, a simple prediction rule, and a decision analytical approach was compared. In the latter, patients were classified as risk too low to benefit from warfarin if the risk of cerebral bleeding on warfarin approximated to, or exceeded, thromboembolic stroke risk reduction. Results. Framingham equations gave lower stroke risks overall than SIGN or SPAF. CHADS2 was intermediate. Using SIGN, warfarin would be given to all 103 patients without a history of stroke/transient ischemic attack and for whom warfarin was not contraindicated but only to 73 patients using the simple prediction rule and 48 patients using the decision analysis. Conclusion. Community-based cohorts give lower stroke risk estimates than CHADS2; both give lower estimates than schemes from control groups from randomized controlled trials. Using community-derived risks would lead to fewer patients being treated with warfarin than guidance derived from randomized controlled trial controls, which may lead to many low-risk patients being treated with high-risk therapy. This raises the debate about appropriate sources of data for risk assessment to support risk communication and effective clinical decisions.
Author(s): Thomson R, Eccles M, Wood R, Chinn DJ
Publication type: Article
Publication status: Published
Journal: Medical Decision Making
ISSN (print): 0272-989X
ISSN (electronic): 1552-681X
Publisher: Sage Publications, Inc.
PubMed id: 17641140
Altmetrics provided by Altmetric