Aspirin prescribing in the secondary prevention of a stroke : a decision analysis approach.
Aspirin is the first line treatment for the majority of stroke patients. Despite national guidelines and
overwhelming evidence of the benefits of therapy, prescribing is sub-optimal: Research has
identified that general practitioners (GPs) are unsure whether to prescribe for patients because of
uncertainty about the balance of risks and benefits for individuals with unique characteristics.
This thesis aimed to reduce uncertainty by enhancing the evidence base available to practitioners.
Fresh data on the risks and benefits of treatment were produced specific to individuals with unique
combinations of characteristics such as their age, time since onset of the initial stroke and history
of diabetes and myocardial infarction.
To improve the likelihood of the new data being drawn upon in a time constrained primary care
consultation, an original computerised decision support system was also developed. This presents
the new evidence in a format more likely to be accessed by practitioners. The system and the data
were then evaluated among GPs.
Specifically, the first phase of research developed a decision analysis model to simulate the
benefits and risks of 'aspirin' and 'no treatment' (chapter 2). For each of 960 profiles the model
utilised the best available evidence from published literature (chapter 3) and primary research
(chapter 4) to reflect the risks of possible outcomes for individuals with each set of characteristics.
Patient preferences of potential outcomes were then combined with the probabilities of these
occurring to generate a prescribing recommendation and other risk data (chapter 5).
The second phase of research developed a user-friendly computer interface to enable GPs to draw
upon the data in a consultation (chapter 6). This provides GPs with easy access to a range of risk
data in several styles,including numerical, graphical and pictorial formats.
Fifteen GPs evaluated the system (chapter 7). Both quantitative and qualitative results indicated
that GP 'decision certainty' improved and 'decision conflict' decreased. GPs felt more organised
about their decision-making, were more able to consider the pros and cons and felt better
prepared. Practitioners also felt that the system would help them communicate the decision to
patients, would aid patient understanding and encourage them to involve patients in decision
This study therefore generated fresh evidence to support decision making in the secondary
prevention of stroke and developed a new decision support system to help GPs draw upon this in a
time constrained patient consultation. The positive evaluation results demonstrate the potential
value of this system and the wider potential of decision analysis to general practice