The Actionable Intime Insights (AI2) Study: Implementing a digital model for timely and needs-based interventions for individuals on anti-psychotic medication in mental health services by applying algorithms to health care data

The AI2 application: Essentially, AI2 is a digital cloud based application service for healthcare professionals and patients that is centralised around the use of My Health Record data. This digital health system contains information such as: medicines (prescription and dispense records, PBS), and Medicare Benefit Schedule (MBS) claims. The methodological premise behind the AI2 study intervention is to use this data as a way to monitor and manage patients, providing real-time visualisation algorithms that identify patients at-risk of relapse and re-hospitalisation. The AI2application is designed so that each MBS and PBS claim made by the patient is visually displayed as an event in a vertical axis, and placed against a time scale (with the day/month and year) on a horizontal axis. In addition to the event (e.g. GP visit), an accompanying simple notes entry box is also included which allows for clinicians to annotate certain insights or predicted outcomes. The additional benefit of adding the notes initiative into the AI2 application is that it adds a contextual backdrop to the documented events, which, in turn, could facilitate the clinicians to identify and develop effective strategies and post-treatment plans. When a clinician logs into the AI2 system, they will have access to the following: 1. their individual profile page; 2. patient list (which when clicked on, will take the clinician to a separate page for each individual patient showing their historic timeline of MBS and PBS events on a horizontal platform); 3. an insights alerts tab, which will show particular events flagged by the clinician as potential actionable insights requiring follow-up, whether such insights were followed-up or ignored, and whether the patient has received a phone call or SMS; 4. a list of patients in dashboard view who are at high risk, which will be updated on a weekly basis by the clinician Monitor; and 5. An SMS message tab, which will host a bank of SMS created by the two psychiatrists on the AI2 team, to send out to patients at sensitively appropriate time points. Clinicians will have access to the Ai2 system for the duration of the study (12 months), however, the involved local health networks and Flinders university may discuss an agreement for continual access, depending on the success of the study, and utility of the system. Healthcare professionals who consent to participation in the study will have their patients enrolled in AI2 in the same method that patients are entered in to their community based information systems by the clinicians who have consented to use the dashboard in clinical care. That is, this information the tool accesses for clinicians is the information that clinicians already have access to through other means, and patients have already provided consent for them to access as part of usual care. As such, the use of this tool will not replace or change usual care. Clinician participants will be asked to meet with the research team to undergo training on how to use the AI2 system, how to correctly identify and interpret the algorithms, and specific appropriate intervention pathways when patients are identified at risk of relapse and hospitalisation. The appropriate intervention pathways that clinicians choose will be based on individual patient-trajectories. For example, if the system triggers a high risk alert of a patient who has missed concurrent prescription refills, the clinician will be advised to immediately make contact with the patient. To help train for this, in-house psychiatrists and the research team will present a series of case studies of individuals with different diagnoses, care plans, and prescription/appointment types to workshop with clinicians involved in the study. These training sessions will be a mix of group and individual face-to-face sessions that will happen during the month before, and the two weeks after the initial start date. The group training sessions will be for an hour weekly, and the individual ones will run for half an hour per clinician once a fortnight. Additionally, a member of the research team will be stations at the trial site for the first two weeks of the intervention in order to provide as-needed training for clinicians. These data patterns will be presented as explained in the patient signup use case. Patients coded as red, identified as high risk, will be communicated to their case managers by the clinician Monitor who will then make contact with their patient. After the case manager has communicated with their patient, they will record an alert in the AI2 system of the outcome. These system analytics will be used to assess the clinician up-take and adherence to use of the Ai2 system in routine everyday patient monitoring. The nurse responsible for monitoring of the AI2 dashboard will create a list of all at risk patients (sorted form high risk to moderate risk) that will be presented to all healthcare professional staff during at least one team meeting per week where they will delineate how to contact and treat patients. The AI2 dashboard will also be used when reviewing patients ready for potential discharge from community services to aid as a decision making tool. At the end of the study, healthcare professionals will be asked to fill out an exit questionnaire about the usefulness of the AI2 dashboard in their daily practice.