A Clinical Rules and Modifiers System in a Patient Decision Aid for Bipolar Disorder

Background In patient-centred healthcare, the core components of a decision should be personalised. The core knowledge components of a decision include the options and embedded procedures, the attributes of the options, and the probabilities associated with the attributes. Evidence about the knowledge components for an individual patient can be elicited from the patient herself, clinicians, caretakers, the electronic medical journal and from research-based sources such as textbooks and clinical guidelines. We name the functions that personalise general knowledge components for individual patients ‘clinical rules and modifiers’. In clinical practice, clinical rules often modify and personalise core components implicitly and unsystematically. The development of software functionality based on clinical rules and modifiers is a promising methodological approach for streamlining the personalisation of core decision components. We are developing a clinical rules and modifiers system as one component in a decision aid for patients with bipolar disorder. Objective We will evaluate the feasibility and validity of a clinical rules and modifiers system for patients with bipolar disorder type I. We will model the system in the Elicia and Annalisa software. Method We will examine the information related to the long-term treatment for individual patients with bipolar disorder type I, in high-quality textbooks and clinical guidelines. The purpose is to identify and extract clinicial rules and modifiers that personalise the options, the attributes of the options, the procedures that the options entail, and the probabilities of the attributes of the options. The content will be modelled using the survey and topic modifying functionality of the Elicia and Annalisa software. The clinical rules and modifiers system will include the following, but not restricted to: contraindications, measures taken to counter side effects, baseline risks, burdens of treatment, and possible interactions. We will determine the ratio of decision-relevant, personalising content in textbooks and guidelines that can be integrated into the software´s clinical rules and modifiers functionality. A heuristic, quantitative and qualitative evaluation of the system will be performed. We also compare the individualised core component output of the system with the output recommended by expert clinicians, using mock patient scenarios and a questionnaire. Results We will present the ratio of content in the textbooks and guidelines that can be included in the software, and the results from the heuristic, quantitative and qualitative evaluation of the system. We will evaluate the accuracy and reliability of the rules and modifiers system, and compare the system output with the recommendations made by the domain experts. Conclusion We will discuss the feasibility, reliability and validity of the clinical rules and modifiers system. []