Dear OHDSI colleagues,
If you are interested in working in rehabilitation1 related research in the OHDSI community, please follow up here. We are interested in building a community working on data driven research in rehabilitation care. You can find a description of the working group and it’s aims below. So please join us and let’s work together!
Background
An increasing wealth of data on rehabilitation healthcare and a person’s functioning is becoming available. These data are largely still unstructured and unavailable for large-scale analysis. Leveraging these data for international research can further personalize evidence-based rehabilitation care, for instance, by developing person-centered prediction and/or stratification models. However, there are major barriers to overcome before rehabilitation data can be used for big data analyses. The challenges stem from considerable heterogeneity within rehabilitation datasets, comprising a diverse array of intervention types, and outcome measures. Furthermore, the notable variations within the population receiving rehabilitation care at various stages in their healthcare journey contribute to this complexity. This complex interplay in the data and relationships between variables presents a daunting challenge to the application of these data in comprehensive large-scale analyses. The substantial variability in datasets within the rehabilitation domain underscores the need for harmonizing rehabilitation data. The OHDSI CDM can create a breakthrough in this.
The OHDSI approach to learning from a large distributed dataset is very promising for rehabilitation medicine. However, until today, little work has been done to map rehabilitation-specific datasets on OHDSI, and little work has been done to answer rehabilitation-specific questions with existing datasets within the OHDSI and EHDEN data networks.
The current common data model and standard OHDSI analytics have been predominantly developed from a pharmaceutical and medical devices perspective. Implementing the OMOP-CDM on rehabilitation-specific data, and answering rehabilitation-specific research questions, has several challenges.
Goal
The goal of the OHDSI Rehabilitation workgroup is to establish a network of OHDSI enthusiasts interested in:
• Sharing and developing knowledge on rehabilitation-specific OMOP-CDM mapping:
• Discussing rehabilitation-specific data analytic challenges
• Stimulating the addition of new OMOP-ed data sets
• Expand the ICF vocabulary for representing rehabilitation measurement and treatment to enable capturing in OMOP-CDM
What makes rehabilitation different? What are the challenges?
In the OHDSI community, there is little experience in the OMOP process with mapping rehabilitation-specific:
• Outcome domains: Typical outcome domains of rehabilitation include PROMS, PREMS, clinical tests, physical examinations, functioning data (cfr. ICF), etc.
• Outcome timing: Some outcomes in rehabilitation take a long time to reach and can be influenced by growing up or aging.
• Rehabilitation-relevant environmental or personal factors: e.g., home situation, relational status, profession.
• Rehabilitation-specific Treatments:
o Rehabilitation interventions are usually complex (e.g., botulinum toxin treatment together with physical therapy), long-term (e.g., three months of physical therapy) and collaborative (requiring active participation by or interaction with the patient (or caregiver).
o Patients are likely to receive multiple treatments that are not a single complex treatment.
o Some treatments may be difficult to classify in a common data model (e.g., there are many different ‘physical therapy’ treatments).
In the OHDSI community, there is little experience with performing clinical rehabilitation studies:
• What rehabilitation-relevant data is already available with current OHDSI members?
• How do we analyze the available data when data are mapped?
o Is it possible to define the relevant phenotypes for analysis? E.g., some treatments may target patients with specific symptoms more than specific diagnoses.
o Is it possible to define treatments clearly in the data analysis?
On behalf of Prof. Dr. Selles, Prof. Dr. Negrini, Dr. Hoogeboom, Dr. Kiekens, Dr. Hoogendam and Dr. Janssen
- “In a health care context,” rehabilitation is defined as a “multimodal, person-centered, collaborative process” (Intervention-general), including interventions targeting a person’s “capacity (by addressing body structures, functions, and activities/participation) and/or contextual factors related to performance” (Intervention-specific) with the goal of “optimizing” the “functioning” (Outcome) of “persons with health conditions currently experiencing disability or likely to experience disability, or persons with disability” (Population).