Related Experiment Videos
Development and description of a decision analysis based decision support tool for stroke prevention in atrial
Richard Thomson1, Angela Robinson, Jane Greenaway
1Department of Epidemiology and Public Health, School of Health Sciences, Newcastle Medical School, Framlington Place, Newcastle upon Tyne NE2 4HH, UK. richard.thomson@newcastle.ac.uk
Background:
There is an increasing move towards clinical decision making that engages the patient, which has led to the development and use of decision aids to support better decisions. The treatment of patients in atrial fibrillation (AF) with warfarin to prevent stroke is a decision that is sensitive to patient preferences as shown by a previous decision analysis.
Aim:
To develop a computerised decision support tool, building upon a previous decision analysis, which would engage individual patient preferences in reaching a shared decision on whether to take warfarin to prevent stroke.
Methods:
The development process had two main phases: (1) the development phase which employed focus groups and repeated interviews with GPs/practice nurses and patients alongside an iterative development of a computerised tool; (2) the training and testing phase in which GPs and practice nurses underwent training in the use of the tool, including the use of simulated patients. The tool was then used in a feasibility study in a small number of patients with AF to inform the design of a subsequent randomised controlled trial.
Results:
The prototype tool had three components: (1) derivation of an individual patient's values for relevant health states using a standard gamble; (2) presentation/discussion of a patient's risks of stroke using the Framingham equation and the benefits/risks of warfarin from a systematic literature review; and (3) decision making component incorporating the outcome of a Markov decision analysis model. Older patients could be taken through the decision analysis based computerised tool, and patients and clinicians welcomed information on risks and benefits of treatments. The tool required time and training to use. Patients' decisions in the feasibility phase did not necessarily coincide with the output of the decision analysis model, but decision conflict appeared to be reduced and both patients and GPs were satisfied with the process.
Conclusions:
It is feasible to develop a decision analysis based computer software package that is acceptable to elderly patients and clinicians, but it requires time and expertise to use. It is most likely that a tool of this type will best be used by a small number of clinicians who have developed experience of its use and can maintain their skills.