Calibration of clinical and self-report measures
in blind rehabilitation
JL Babcock, Ph.D1., DN Head, Ed.D1, PE McKnight, Ph.D.2, R Massof, Ph.D.3
1Southern
AZ VA Health Care System, 2University of Arizona, Lions Vision Research
and Rehabilitation Center, 3Wilmer Ophthalmological Institute, Johns Hopkins
University
Objectives: The primary purpose of this project
is to calibrate two outcome instruments used to measure functional ability
in geriatric patients who are blind or severely visually impaired. Calibration
of the two measures permits easy transformations of values in order to
establish a "gold standard" to be used for benchmarking blind
rehabilitation outcomes across VA Blind Rehabilitation Centers and to
exam alternative service delivery models. The two measures calibrated
are the Blind Rehabilitation Service Follow-up Outcome Survey (BRSFOutSur)
and the Functional Assessment of Self-reliance on Tasks (FAST). The BRSFOutSur
is a self-report instrument developed and administered over the past three
years by researchers at the Atlanta RR&D Center. The FAST is a clinical
measure of functional ability and was developed during the same time period
by researchers at Southwestern Blind Rehabilitation Center (SWBRC).
Methods: The BRSFOutSur data were collected
3-months post-discharge (i.e., retrospective pretest and posttest) through
a phone survey and the FAST data were collected at admission and discharge.
Both measures sampled veterans who attended the inpatient blind rehabilitation
program at SWBRC. The two databases were merged yielding approximately
300 matching cases for the secondary analyses. In order to develop a common
measurement metric, Rasch analyses were conducted to estimate the difficulty
of the survey items. Subsequent to the Rasch analysis, a generalizability
study was used to identify sources of variation and under what circumstances
survey responses are meaningful. Following these procedures, both Rasch
score matching as well as classical raw score bivariate regression analyses
were performed to compute and algebraic formula for converting scores
from the BRSFOutSur to the FAST and vice versa.
Results: The Rasch analyses provided information
about the stability of items both within a single assessment period and
over two separate assessment periods (i.e., pre- and post). From these
results, only those items that performed well were kept for the calibration
procedures. An algebraic conversion algorithm was determined and the implications
and their use are discussed in detail. The analyses also identified the
personal characteristics that moderate change (e.g., eye disease, acuity,
and health comorbities) and item characteristics (e.g., item difficulty,
interval stability, assessing tasks versus goals). From these findings,
the development of a patient profiling system will lead to the establishment
of defensible benchmarks that will set standards of expectation for improvement
in veterans’ functional abilities following blind rehabilitation training.
Conclusions: The calibration methods employed
in this study have identified functional ability items that can be used
interchangeably. These items can also be used as benchmarks to identify
a program’s effectiveness. Further development of these calibration methods
should provide the basis for developing a common measurement metric with
which to determine functional improvement. These findings also provide
a foundation for understanding measurement theory in low vision rehabilitation.
Funding Acknowledgment: This study was funded
by the VA Rehabilitation Research and Development Service, project #C-2710I.