The GRiST response to problems with risk prediction has been to develop GRiST in such a way that it provides a very structured and systematic approach to risk assessment, but one which is based on a consensual and holistic model of risk shared by multidisciplinary mental health experts (see Buckingham et al. 2008 for a full description). GRiST is unique amongst risk assessment tools in combining the best of both established approaches to mental health risk assessment: structured clinical judgment (i.e. how clinicians formulate risks using their own expertise and training) and actuarial approaches (i.e. those based on statistical analysis of population data). Each time GRiST is used, the risk profile of the person assessed and the risk judgments attached to it by the assessor are stored in the GRiST database in anonymous form. No other risk assessment tool has the capacity to link the psychological knowledge structures and reasoning processes of clinicians to their clinical judgements in such a precise way.
The GRiST model of clinical risk expertise was generated from in-depth interviews with 46 mental-health experts that were transcribed and analysed. The results were subjected to validation by a panel of over 100 mental health clinicians in the first instance through focus groups and on-line exercises. This expert clinical input has been massively expanding over the years as each new organisation takes on GRiST and has input to its ongoing development. The resulting hierarchical model of risk expertise contained by GRiST provides a precise index and structuring of risk information which makes it easy to store and analyse data as well as link it to other information-gathering systems.
GRiST is designed to support structured clinical judgment, but not to take away the clinicians prerogative in clinical decision making. It does this by leading clinicians through a detailed and systematic process for collecting risk information, by the end of which they will be in possession of a comprehensive patient risk profile on the basis of which to make their best judgment.
Clinicians are asked to assess each piece of risk information and overall risk judgments using an 11 point scale, to assist them in accurate decision making. In addition there is the capacity for people to add narrative information to any part of the hierarchy. For example, there are questions exploring links between physical and mental health. People can use the comment boxes attached to them to put in more precise details about how their mental health is affected by, for example, their diabetes.
GRiST therefore collects both quantitative and qualitative information about risk, which fits well with clinicians tradition of capturing risk information in text-based form. GRiST is certainly not a tick box exercise, which refers to ticking an item off a list to show that it has been completed. Clinicians using GRiST tell us that GRiST makes you think and anecdotal evidence from training sessions suggests that GRiST can improve clinical decision making. Clinicians are reported to take time and trouble making sure that their risk judgments properly reflect assessed patient profiles, and vice versa.