1 Elderly Services, General Infirmary at Leeds, Leeds LS1 3EX, UK
2 St James's University Hospital, Leeds LS9 7TF, UK
3 Centre for Decision Research, Leeds University Business School, Leeds LS2 9JT,
UK
4 Department of Statistical Sciences, University of Cape Town, 7701 Rondebosch,
South Africa Correspondence to: Dr R Fuller E-mail:
ftworf{at}aol.com
Dr Shah and Dr Dawson's dilemma (March 2004 JRSM1) of managing intracerebral haemorrhage (ICH) in a patient previously anticoagulated for a prosthetic heart valve will be familiar to many clinicians. Whilst re-anticoagulation would be current practice, this is based on perceptions that the benefits of thromboembolism prevention surpass the risks of re-haemorrhage.2 Such decisions are prone to recall or availability bias, given the varying risks of valve thromboembolism and the fragmentary evidence base on risks of rebleeding with further antithrombotics/anticoagulants.
Clinical trials are unlikely to resolve this dilemma, not least because of the difficulty of recruiting sufficient participants. Decision analysis presents an alternative solutiondeveloping a framework in which to combine variations in multiple probabilities within a mathematical model. Such techniques allow examination of current practices where no consensus exists, using a range of values/scenarios to test the feasibility of the model. These techniques have already seen successful use in re-examining evidence-based guidelines and in assessing decision-making in atrial fibrillation and intracerebral haemorrhage.3,4
We are currently constructing a decision model to explore the dilemma faced by Shah and Dawson, testing the effects of using anticoagulants and antiplatelet agents in patients with mechanical valves and ICH. A key test of decision modelling outcomes is to review them against current practices, and it is noteworthy that recent case reports have highlighted the potential short-term use of antiplatelet agents such as aspirin and clopidogrel in place of immediate anticoagulation following cerebral bleeding.5
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||