Development of a Personalised, Multi-Criteria Decision Analytic Web-Based Patient Decision Aid for Long-Term Treatment in Bipolar Disorder Type I

Background Many patients with bipolar disorder experience poor decision quality. Common deficits include suboptimal assessments of the positive and negative effects of former medication, a lack of monitoring of the effects of ongoing medication, low adherence, and a lack of opportunity to bring their own preferences to bear in the evaluation of their medication. The increasing number of medication options, a relative lack of easily-accessible head-to-head comparisons of pharmacotherapeutic alternatives on patient-important outcomes, and the insufficient personalisation of research findings to individual patients, add up to a formidable challenge for both patients and clinicians in selecting and monitoring optimal treatment. Objective To develop a web- and multi-criteria decision analysis-based patient decision aid that assists the evaluation, selection, monitoring and follow-up of pharmacotherapy in the long-term clinical management of bipolar disorder type I. The decision aid generates a one-screen, interactive, personalised summary of all the core components of long-term medication decisions in bipolar disorder. The system enables the inclusion and integration of evidence about the current state and future probabilities of the outcomes of the options from the patient, from the clinician, from the EMR, and from research findings. The decision aid will be used repeatedly for the improvement of decisions over time, monitoring the quality of the personalised decisions longitudinally, and generate visual images of trends in the crucial decision components. The aid will display the degree of – and long-term alterations in – concordance levels between the beliefs and preferences of the patient and the clinician, and the differences in opinions expressed by the patient him/herself at different stages of the disease. In contrast to patient decision aids that support one-off decisions, the decision aid supports ongoing decisions. It aims to provide continuous process support and a loop of forecasting, action, assessment, and new forecasting, thereby facilitating substantially augmented discussions and shared decision-making. Methods A manual for the production of the decision aid has been developed. Several small-scale literature reviews and summaries are being conducted in accordance with the manual, as well as a supplementary feasibility analysis, along with content and requirement specifications. A formal web-based elicitation process involving domain experts will be used to derive the baseline performance ratings of the medication options on the patient-important outcomes, where these are not available from network meta-analyses. The content will be modelled in a pre-existing template. Patient representatives and clinicians will perform simple, iterative usability testing. Results The implementation of the manual, our associated experiences, and preliminary findings will be described. The incremental gain of using elaborate vs. simple procedures to produce the basic input to the system will described. Conclusion In the next phase, a clinical study applying a single-subjects research design will establish feasibility of the system, and the effects of the decision aid on personalised decision quality in individual patients, adherence rates and patient-clinician communication. The manual developed for this project will be applied in the construction of other patient decision aids with the aim of arriving at a generic manual. []