Clinical Judgement: an investigation of clinical decision-making

Mental Health Professionals (MHPs) make many important decisions in their daily practice which direct the care pathway and type of treatment delivered. Despite access to patient outcomes, the evidence base, clinical guidelines and diagnostic criteria research suggests many decisions are likely to be influenced by a number of dynamic and context variables. Consequently, clinical judgements and decisions are often prone to inaccuracy. The manner in which heuristics and biases influence the decisions of MHPs has not been fully investigated and the methods previously used have been rudimentary. By conducting a systematic review (including meta-analysis where possible) and two empirical studies this thesis aimed to contribute to the evidence-base regarding clinical judgement and decision-making. The first part of this thesis describes the results of a systematic review and meta-analysis yielding 24 papers investigating the factors that have been shown to adversely influence accuracy of clinical judgement and decision-making. Results showed that several variables were implicated in this process. These include causal assumptions/causal theories, representativeness, contextual information, and race/culture. The decisions MHPs make are often inaccurate and evidence suggests that heuristics and biases are a probable cause for this. The second part of this thesis reports two studies. Their aim, to design and test a trial-based methodology to assess the influence of bias on decisions regarding treatment allocation and progression. The first study developed an innovative ‘real time’ scenario-based approach (referred to as a ‘dynamic measure’) to assess clinical judgement and reasoning traits of Psychological Wellbeing Practitioners (PWPs) working as part of the Improved Access to Psychological Therapy (IAPT) programme. A non-systematic review of the cognitive biases and heuristics literature was conducted to develop a preliminary draft of the dynamic measure. This included a case vignette of a fictional male client referred to step 2 of the IAPT programme. Ethnographic decision-tree modelling was employed in the final stage of development. This incorporated qualitative thematic analysis of a focus group and a pilot study to develop two final versions of the dynamic measure. In the second study 133 PWPs took part in an online survey completing two decision-making tasks (experimental and control). This was so that decisions when encountering a particularly challenging scenario during low intensity treatment could be compared with when treatment was relatively straightforward. Tasks included typical decisions a PWP is required to make during on-going clinical practice (e.g. assessing patient suitability for treatment, degree of alignment to treatment protocol and decisions when a client is not showing reliable improvement by session 4). The convergent validity of the dynamic measure as a test of heuristics and biases was not established but divergent validity was. Results suggest that the degree of treatment fidelity demonstrated by therapists and reasons they might sometimes prolong or conclude treatment may be due to an interaction between the therapist and the context. Given that the present study was explorative and convergent validity was not achieved further research is required.