Section 10 BRAT value tree

Value tree from BRAT framework
Short name Description
Name/rubric Value Tree
Created in BRAT Lotus notes program plus PowerPoint, but can also be created in PowerPoint alone
Message To display qualitative listing of available key benefits and risks criteria for the decision model and the description of their measurements in a hierarchical way. Colours are used to indicate level of criteria.
Intended audience Regulator and Physician. Also for patients if background is provided on the medical terms in lay language.
Knowledge required Needs Medical knowledge (terminology of the medical terms; for judging whether selected benefits and risks make sense), no statistical knowledge needed. The tree could be amended for the patients as well using lay terminology to be easily readable.
Unintentional message As usually benefits are predetermined, and on the risk side, potential risks are also included so there is a tendency to show more risks than benefits. For the inexperienced reader this could be perceived as more risks if no further qualification of the risks is provided. This could be done by more explicit description of certainties, numbers are not weighted but weights could be added alongside. As presented, each criterion may be seen as being equally weighted.
Message not communicated The context in which the value tree is created is needed (underlying database, indication, what decision). There is no quantitative data and no information on the treatment alternatives.
Proposed improvement Some background information, such as definitions for the indications should be included via text annotation. Explanation of whether the order of criteria reflects weighting would be useful. The naming of criteria must intuitive reflect their role as a benefit or a risk of treatment, for example write relapse as reduction in relapse rate to reflect it is a benefit of treatment. Different colours should be used to distinguish benefit and risks. Risks should be characterised in terms of certainty, and lay terms to be added to make the value tree more useful for patients.