 Computable Phenotyping of Pediatric Rehabilitation Population
 Biography Overview Pediatric rehabilitation medicine (PRM) stands to greatly benefit through Big Data research, as exemplified by the Model Systems’ success in expanding our understanding of spinal cord injury (SCI) and traumatic brain injury (TBI) in predominantly adult populations. The Model Systems excludes patients younger than 16 years of age, creating a gap in knowledge on how to best care for this vulnerable pediatric population. The Uniform Data System for Medical Rehabilitation (UDSMR) and eRehabData databases are the closest analogs; however, participation in these databases is voluntary, and the data is relatively limited and of variable quality. Use of a database that is neither disease nor setting specific, such as the PEDSnet database, could meet this gap; however, no method currently exists to efficiently identify pediatric patients that have undergone acute inpatient rehabilitation care in such databases. The clinical concept of “acute inpatient rehabilitation care” must be made machine-readable, and one method to accomplish this task is computational/computable phenotyping. In essence, identifying this pediatric cohort could be comparable to the Model Systems adult-focused cohort with potentially more patients and more data overall, albeit less rehabilitation-specific. In collaboration with PEDSnet – a national, pediatric, learning health system – and their database of several million pediatric patients with more than 200 data elements and 19 data domains, I propose developing a computable phenotype (CP) to efficiently and effectively identify these patients. The aims of this project include CP development, testing, and validation. We hypothesize that an algorithmic CP will be required given that patients undergoing acute inpatient rehabilitation care are not solely identified by diagnostic codes, rather, they are identified by the multidisciplinary care that they receive. This work will require collaboration amongst Children’s Hospital Colorado (CHCO), the Department of Physical Medicine & Rehabilitation at the University of Colorado (CU PM&R), PEDSnet, and two additional PEDSnet sites.
Time
|