Related Experiment Video
Updated: Aug 17, 2026
Using Visual and Narrative Methods to Achieve Fair Process in Clinical Care
Published on: February 16, 2011
Measuring medical complexity during inpatient rehabilitation after traumatic brain injury
David K Ryser1, Marlene J Egger, Susan D Horn
1Division of PM&R and Statistical Data Center, LDS Hospital, 8th Avenue & C, Salt Lake City, UT 84143, USA. lddryser@ihc.com
Objective:
To compare the performance of 3 severity of illness (SOI) indices--the Comprehensive Severity Index (CSI), All Patient Refined Diagnosis Related Groups Severity of Illness, case-mix group (CMG)--and 5 well-known neurologic parameters, as measures of medical complexity.
Design:
Retrospective chart review.
Setting:
Inpatient rehabilitation center within a level I trauma center.
Participants:
Consecutive traumatic brain injury (TBI) admissions (N=212).
Intervention:
Acute inpatient TBI rehabilitation. CSI and neurologic parameters were scored by chart extraction. SOI was based on diagnosis codes by using 3M PC Grouper software, version 15.
Main Outcome Measures:
Adjusted R 2 was used to predict rehabilitation charges as a proxy of medical complexity.
Results:
The highest adjusted R 2 values for single variables predicting charges were: CMG .349, CSI .293, duration of posttraumatic amnesia .260. Adjusted R 2 values for the CMG combined with the CSI, 5 neurologic parameters, and SOI to predict charges were .446, .431, and .365, respectively.
Conclusions:
The CMG was the best single predictor of rehabilitation charges for TBI. Predictive ability was better when the CMG was combined with the CSI or a combination of the 5 neurologic parameters. A severity index based on objective clinical findings rather than diagnostic codes may have distinct advantages for rehabilitation outcome studies and reimbursement methodology.
