A Learning Health System Infrastructure for Precision Rehabilitation After Stroke. (February 2023)

• Record Type: Journal Article
• Authors: French, Margaret A.; Daley, Kelly; Lavezza, Annette; Roemmich, Ryan T.; Wegener, Stephen T.; Raghavan, Preeti; Celnik, Pablo
• Is Part Of: American journal of physical medicine & rehabilitation. Volume 1021(2023)Supplement 1
• Page Start: S56
• Page End: S60
• Publication Date: 2023-02
• Subjects: Stroke -- Electronic Health Record -- Rehabilitation -- Database -- Learning Health System; Rehabilitation -- Periodicals; Medicine, Physical -- Periodicals; 617.062
• DOI: 10.1097/PHM.0000000000002138
• Languages: English
• ISSNs: 0894-9115

Abstract

Abstract: Functional recovery and the response to rehabilitation interventions after stroke are highly variable. Understanding this variability will promote precision rehabilitation for stroke, allowing us to deliver targeted interventions to the right person at the right time. Capitalizing on large, heterogeneous data sets, such as those generated through clinical care and housed within the electronic health record, can lead to understanding of poststroke variability. However, accessing data from the electronic health record can be challenging because of data quality, privacy concerns, and the resources required for data extraction. Therefore, creating infrastructure that overcomes these challenges and contributes to a learning health system is needed to achieve precision rehabilitation after stroke. We describe the creation of a Precision Rehabilitation Data Repository that facilitates access to systematically collected data from the electronic health record as part of a learning health system to drive precision rehabilitation. Specifically, we describe the process of (1) standardizing the documentation of functional assessments, (2) obtaining regulatory approval, (3) defining the patient cohort, and (4) extracting data for the Precision Rehabilitation Data Repository. The development of similar infrastructures at other institutions can help generate large, heterogeneous data sets to drive poststroke care toward precision rehabilitation, thereby maximizing poststroke function within an efficient healthcare system.