PROFIT: Prevention of Falls using IT

Falls are one of the most common and serious threats to older, especially vulnerable, persons. The falls’ high incidence comes with high costs, morbidity, reduced functionality, loss of independence, anxiety, and death. Vulnerable patients do not get the care pertaining to falls that they are entitled to. Due to the complex multidisciplinary character of the fall management process it is perceived that disease-management (DM) and innovative ICT are key to the solution. Current initiatives to improve fall management do not capitalize on employing innovative ICT opportunities available today to support DM. PROFIT’s main objective is to understand the role of innovative ICT in supporting DM by investigating a concrete fall management setting and developing applications for demonstrating this. Another objective is to consolidate an emerging multidisciplinary research team of experts in the fields of (medical) computer science, machine learning, and primary and secondary care to book advances not otherwise attainable. Concretely, we propose in PROFIT to investigate, develop and apply methods originating from Decision Support, Statistical Machine Learning, and Web-engineering. The proposal evolved from activities funded by a preparatory grant in which a multidisciplinary team committed to improving fall management along the lines of DM and ICT was established. The healthcare providers include the homecare organization Amsterdam Thuiszorg; the general practitioners (GPs) within the Stichting Gezondheidscentra Amsterdam ZuidOost; and the departments of Emergency, Geriatrics, and the Falls-outpatient clinic from the Academic Medical Center in Amsterdam. We applied a patient-centric disease-management approach to the analysis of the partners’ roles, the multidisciplinary care processes, and the bottlenecks and ICT opportunities to improve care provision. These activities culminated in three main ideas that we propose to pursue in PROFIT: The first idea concerns the DM concepts of prevention, pro-active use of evidence-based standards, and enhancing multidisciplinary work. It entails the use of decision support systems to: broaden the primary care safety net to timely “catch” vulnerable elderly patients based on automating the Identification of Seniors At Risk (ISAR) triage tool; to lower variability of the GPs’ care provision by applying quality indicators from the Assessing Care of the Vulnerable Elderly (ACOVE) initiative; and to alert partners about events happening elsewhere in the process. The second idea is linked to the DM concept of the “learning organization” for using outcomes to understand the process and refine it. It entails the use of machine learning algorithms to support managers at homecare by finding segments and useful patterns in care consumption; and to support medical scientists and care providers by building prognostic models through learning from joining the national fall-prevention database, the Amsterdam GPs’ academic population database, and the Amsterdam (hospital)pharmacy database to identify risk factors (such as medications) and to refine and validate risk assessments tools. Finally, the third idea is linked to the DM concept of self-management and patient empowerment. It entails employment of Web-engineering to develop a usable demand-driven website that guides patients and/or their informal caregivers in choosing and expecting the next steps in their process, and educating them on the quality of care provision they are entitled to. The website will also include risk-assessment tools that can be accessed from home, from the GP office and from the Emergency department. The latter two locations will allow observing users interacting with the website. We collaborate with the Clinical Informatics department at LUMC in Leiden on this idea. PROFIT addresses the following ICT scientific challenges: 1. How to integrate clinical decision support with workflow and documentation? How to exploit ACOVE, an audit instrument, for pro-active support? Does such a system improve the process? Is it acceptable to GPs? This will contribute to knowledge on designing decision support systems and understanding their effects. 2. Which algorithms are needed for providing patterns of care-consumption to homecare managers? Which algorithms are needed for analyzing risks of adverse events? This will contribute to the state of the art in machine learning and predictive modelling. 3. How to design a demand-driven process-based usable website to empower patients? This will contribute to our knowledge of designing websites for special groups and understand barriers and acceptance issues. The societal importance of PROFIT originates from its strive to promote an improved safer environment for the elderly. The methods will be developed and tested in the everyday clinical practice leading to demonstrators allowing others to share our experience.