Towards Personalised Clinical Management of Suicide Risk through Data-Driven Clinical Decision Support using Transnational Registry Data

Effective suicide risk management in clinical practice is a challenging task due to the lack of tools that can assist accurate assessment of suicide risk. This leads to ineffective clinical decision-making and unmet need of adequate mental health treatment. In this project we aim to develop a personalised Clinical Decision Support System (CDSS), i.e., a medical software program that can assist in the personalised clinical evaluation and management of suicide risk. This CDSS will enhance current clinical practice by increasing the prediction accuracy of suicide risk assessment, and by enabling the personalised matching of the identified risk profiles with effective treatment. We will use data from population registries from Ireland, Norway, Sweden, and Catalonia (Spain), and apply machine learning techniques to develop accurate and clinically useful prediction models for suicide and unmet healthcare need, including identification of the most important risk factors. This will form the basis for our further work to develop a software program of CDSS that can predict risk of suicide an individual patient may have. We believe, the CDSS developed from this project will have the potential to be implemented in regional or national healthcare system, to enable personalised and timely delivery of effective treatment for suicidal behaviour at large scale, and finally to contribute to suicide prevention in the population.

Two longstanding limitations hamper effective suicide risk management in clinical practice. First, unassisted clinical judgement is not sufficient to accurately assess suicide risk, leading to ineffective clinical decision-making and poor patient experience; and second, the need for adequate mental health treatment is often unmet among patients with suicide risk, and evidence indicates that adverse healthcare trajectories are associated with high morbidity, premature mortality and societal costs related to self-harm. The PERMANENS project aims to improve suicide prevention by developing a personalised Clinical Decision Support System (CDSS), i.e., a medical software programme that assists in the personalised clinical evaluation and management of suicide risk. The CDSS will innovate current clinical practice by (1) increasing the prediction accuracy of suicide risk assessment; (2) enabling risk assessment for inadequate treatment delivery among patients with suicide risk; (3) enabling fine-grained clinical risk stratification; and (4) enabling the personalised matching of the identified risk profiles with effective treatment in order to improve indicated and tailored treatment trajectories among patients with suicide risk. Data for the project will be obtained from population-representative electronic registries from Ireland, Norway, Sweden, and Catalonia (Spain). Machine learning techniques will be used to develop accurate and clinically useful prediction models for suicide and adverse healthcare trajectories, including the identification of most important risk factors. Through co-creation and user-oriented qualitative implementation research with patients and clinicians, a user-friendly personalised CDSS prototype will be developed. The CDSS will be provided with a computerised clinical knowledge base on effective suicide prevention interventions and with a transferable personal healthcare record, fully acknowledging patients as end-users of their data.