This recently published report from the European project PREPARE represents a significant advancement in personalised rehabilitation care. D2.2 Pilot cases planning, Definition of Scenarios, Validation Plan, is a comprehensive document outlining the methodology used to develop prediction and stratification (classification) tools, aimed at tailoring treatments to individual patients. To provide the best care, rehabilitation must be a deeply person-centered process, which is why the PREPARE project, involving 19 different partners, seeks to revolutionise rehabilitation practices through robust, data-driven computational tools.
The report details how scenarios were defined across nine clinical cases, focusing on musculoskeletal, neurological, and vascular health conditions that are common causes for rehabilitation in Europe and worldwide. Using existing databases, the project involves the collection and analysis of clinical data, as well as a comparison of different rehabilitation processes. The main objectives of the report are the definition of clinical scenarios and the creation of a pilot case validation plan, which will guide the development of a versatile Clinical Decision Support System (CDSS) designed to enhance patient outcomes and reduce treatment costs.
Nine Pilot Cases
Described in detail are the following nine clinical pilot cases – each focusing on a health condition, a population and setting, and a rehabilitation process that is different from the others:
- Idiopathic scoliosis in children, ISICO (Italian Scientific Spine Institute), Italy
- Intermittent claudication, Radboud University Medical Centre, Netherlands
- Lower limb loss, University Rehabilitation Institute, Slovenia
- Parkinson’s disease and parkinsonism, Universita Politecnica delle Marche, Italy
- Hip and knee replacements due to osteoarthritis, Ospedale Galeazzi Sant’Ambrogio, Italy
- Temporomandibular joint disorders, Università degli Studi di Milano, Italy
- Spine disorders, Eurospine (The Spine Society of Europe), Switzerland
- Hypertension, Aristotle University, Greece
Through these pilot cases, researchers aim to achieve several key objectives: verify the models’ clinical applicability, evaluate how these tools can influence decision-making, and gather data for future clinical trials. This approach seeks to improve patient outcomes while minimizing the social, psychological, and economic burdens associated with rehabilitation.
Methodology: Developing Clinical Decision Support Systems (CDSSs)
To develop the Clinical Decision Support Systems (CDSSs) for each pilot case, the clinical partners conducted comprehensive literature reviews of existing CDSSs pertinent to each health condition. They outlined one to four scenarios in which a CDSS could optimise rehabilitation care through patient stratification or outcome prediction. To determine the feasibility of these scenarios, they identified relevant predictors documented in the literature and compared them with variables available in their existing databases. This process allowed them to identify additional variables that could further refine the models.
The outcome of this collaborative effort was a comprehensive description of the nine pilot cases, detailing each health condition, population, rehabilitation process, and the range of variables involved, with a special focus on socio-behavioral factors. The information gathered serves as a comparative analysis, highlighting similarities and differences among the cases. This analysis will inform the technical partners developing the CDSSs, providing both a common foundation and guidance for tailoring tools to each specific clinical context.
The Diversity of Rehabilitation Processes
Each pilot case focuses on different aspects of rehabilitation, illustrating the vast diversity within this field. For instance, some cases explore optimising ongoing treatments, while others focus on predicting the need for rehabilitation. This diversity not only enriches the report’s findings but also ensures that the resulting CDSS tools will be adaptable to various rehabilitation scenarios.
The clinical partners gathered an array of variables, paying close attention to socio-behavioral data, which is crucial for tailoring rehabilitation interventions. This includes details like patient demographics, and disease and impairments focused classical clinical measures, quality of life measures, behavioral factors, and other socio-economic indicators. By focusing on these elements, the report acknowledges that successful rehabilitation goes beyond managing the health condition itself—it involves enhancing patients’ functional abilities and quality of life in their daily environments.
Applying Data-Driven Tools in Clinical Practice
Integrating predictive and stratification models into CDSSs can significantly enhance personalised rehabilitation care. These tools support rehabilitation professionals by offering data-driven insights, enabling them to make informed treatment decisions in collaboration with patients. The inherently variable nature of rehabilitation often creates uncertainty about the best treatment options. CDSSs help mitigate this uncertainty by providing timely, relevant information about expected outcomes, allowing professionals to adjust treatment plans to meet each patient’s unique needs.
Research indicates that CDSSs can improve clinical efficiency, patient engagement, and overall outcomes. However, conventional models have limitations; they often lack the external validity required for broader application in clinical practice. Advanced, AI-driven CDSSs, based on large-scale rehabilitation data, present a solution to this challenge. By analysing vast amounts of patient and treatment data, AI can uncover complex patterns, leading to more effective and personalised treatment plans.
The Significance of Socio-Behavioral Data
Each of the pilot cases differs substantially in terms of the amount and types of socio-behavioral data they collect. This diversity in data collection underlines the importance of a flexible CDSS capable of accommodating varying information needs across different rehabilitation contexts.
Building a Unified yet Flexible Platform
The report’s comprehensive comparison of the pilot cases identified overlapping domains, such as pain management, gait, and balance, common across different health conditions. These commonalities provide a foundation for developing unified tools within the CDSS platform. Simultaneously, the differences among the cases highlight where customisation is necessary, guiding the development of tailored components for each specific rehabilitation setting.
One of the key challenges was creating a consistent terminology for the predictors, which stressed the need for a standardised vocabulary to facilitate the development of the CDSS. Despite this challenge, the deliverable found areas of clear overlap among different partners’ outcome domains, indicating that shared questions about treatment efficacy exist even in different rehabilitation contexts. This insight will be crucial for building a versatile, internationally applicable CDSS that can be adopted across different settings.
Future Directions
The main objectives of the report include defining clinical scenarios and establishing a validation plan for the pilot cases. These steps will guide future efforts in building and validating the CDSS within the upcoming work packages. The goal is to develop a platform that not only serves current rehabilitation needs but can evolve to include additional health conditions as new data and insights emerge. By focusing on outcomes, interventions, and the functional capacity of the rehabilitation population, this approach emphasises optimising patients’ overall functioning and quality of life.
Authors: Stefano Negrini, University of Milan, and Carlotte Kiekens, ISICO