To investigate the measurement properties of the Orientation Log (O-Log) and a hybrid scale using Rasch analysis techniques and to explore the relations between the items of the 2 scales and the latent linear construct modeled in the Rasch analysis.
Calibration of data collected weekly during inpatient rehabilitation.
Ninety patients admitted for rehabilitation after traumatic brain injury.
Measure reliability/construct validity of the O-Log was too low to justify restructuring of disordered rating scales, but the broader set of items in the hybrid scale demonstrated good measure reliability/construct validity prior to and following rating scale restructuring. Both O-Log and hybrid measures demonstrated statistical fit with the linear latent construct, suggesting that orientation and memory are only a subset of symptoms in a broader syndrome.
Posttraumatic amnesia is by definition a proxy measure of a broader syndrome during early recovery and not a measure of the syndrome itself. The results suggest that the O-Log cannot reliably measure progress during early recovery. Furthermore, the analysis suggests that the construct of posttraumatic amnesia is too narrow and should be revisited to improve monitoring of recovery and prognostic estimation after brain injury.
Department of Physical Medicine and Rehabilitation, Indiana University School of Medicine and Rehabilitation Hospital of Indiana, Indianapolis (Drs Kean and Malec); Department of Psychological and Brain Sciences, Indiana University, Bloomington (Ms Abell); Eli Lilly and Company, Indianapolis (Dr Trzepacz).
Corresponding Author: Jacob Kean, PhD, Department of Physical Medicine and Rehabilitation, Indiana University School of Medicine, Indianapolis, IN 46202 (firstname.lastname@example.org).
The authors thank Bradley Hufford, Mark Williams, Sherry Pryor, and Cindy Campbell for assistance with data collection and helpful discussion of study results.
The authors declare no conflict of interest