Evaluating associations between trauma-related characteristics and functional recovery in individuals with spinal cord injury

Wednesday, November 9, 2022

3:56 PM – 4:01 PM

Location: Station 4

Presenting Author(s)

Shivayogi V. Hiremath, PhD

Assistant Professor
Temple University
Philadelphia, Pennsylvania, United States

Non-presenting Author(s)

RM

Ralph J. Marino, MD, MS

Professor of Rehabilitation Medicine
Sidney Kimmel Medical College at Thomas Jefferson University
Philadelphia, Pennsylvania, United States
DC

Donna L. Coffman

Associate Professor
Temple University
Philadelphia, Pennsylvania, United States
Amol M. Karmarkar, PhD

Professor and Director of Research
Virginia Commonwealth University (VCU) & Sheltering Arms Institute (SAI)
Richmond, Virginia, United States
CT

Carole A. Tucker, PhD

Associate Dean of Research
UTMB
Galveston, Texas, United States

Research Objectives: To utilize trauma variables, acute phase-related variables, and patient-level characteristics to estimate functional recovery at discharge from an inpatient rehabilitation facility (IRF) in individuals with spinal cord injury (SCI).

Design: A retrospective cohort analysis.

Setting: Two SCI Model Centers in Pennsylvania, United States.

Participants: We used a record linkage toolkit to connect 735 individuals (or cases) with traumatic SCI between the Pennsylvania Trauma Systems Outcomes Study (PTOS) and National Spinal Cord Injury Model Systems (SCIMS) databases.

Interventions: N/A

Main Outcome Measures: Functional recovery during inpatient rehabilitation determined by linear regression (LR) and gradient boosting regression (GBR) analyses.

Results: The optimized linking algorithm used five variables including injury date, discharge date, age, biological sex, and race and ethnicity to link the cases across PTOS and SCIMS databases. The algorithm development process aimed at maximizing the true-match rate (>90%) and minimizing the error rate ( < 1%). The final percentage for true-match and error rate were 92.0% and 0.1%, respectively. Post-linkage, an LR analysis in 604 individuals with SCI indicated that age, SCI level combined with American Spinal Injury Association impairment scale grade, length of stay at IRF, diabetes mellitus, and pulmonary disease complication during acute care were significant predictors of functional recovery. The results of GBR analysis (R2 = 0.58) was similar to LR analysis (R2 = 0.56) regarding key variables and their contribution to the functional recovery model.

Conclusions: Demographic variables and dates allowed us to link the PTOS and SCIMS databases. The trauma variables, acute phase-related variables, and patient-level characteristics were able to explain functional recovery at discharge from an IRF. Further research is necessary to evaluate long term functional recovery in individuals with SCI.

Author(s) Disclosures: None

Learning Objectives:

Describe the types of data collected by the Pennsylvania Trauma Systems Outcomes Study (PTOS) database.
Describe the types of trauma variables, acute phase-related variables, and patient-level characteristics to estimate functional recovery at discharge from an inpatient rehabilitation facility in individuals with spinal cord injury.
Describe the variables used to link individuals with spinal cord injury between the Pennsylvania Trauma Systems Outcomes Study and National Spinal Cord Injury Model Systems databases.