| About StressTest
Historical Development of the StressTest
StressTest is based on the Treatment Outcome Profile (TOP). TOP was designed
to be a self-report measure of each of the key dimensions of
Quality of Life, Level of Functioning, and Symptomatology. Originally, TOP was designed to help counselors
asses how well they were meeting the needs of their clients.
Each of these sub-sections was scientifically developed with different populations,
including clients being seen in outpatient settings, inpatient settings,
residential care facilities and people living and working in the community.
The Quality of Life section of StressTest was developed out of research with
clients in a hospital and in community residential
treatment facilities (Holcomb, 1983).
An extensive review of the literature on Quality of Life was done, and a large
pool of items was generated to measure both general life satisfaction and
domains of quality of life. Results of this initial research indicated that
four major areas of quality of life accounted for the majority of the variance
in quality of life ratings. These four sub-dimensions of quality of life were
autonomy, self-esteem, social support, and physical health. Items that contributed to
higher internal reliability coefficients, and higher factor loadings (using factor analysis)
on these four sub-scales were then adapted and used for measurement of Quality
of Life in StressTest.
It should be added that these four subscales correlated very highly
with ratings of common symptoms using the Brief Rating Psychiatric
Scale (BPRS) and the Global Assessment Scale of Functioning (GAF). In
addition, they correlated with other self-report and interviewer ratings of
quality of life and significantly discriminated among people needing different
levels of serivces.
The second section of StressTest measures Symptomatology. The original research on this
section was completed with 451 patients admitted to an acute hospital treatment unit.
Originally, the SCL-90 was used to measure client symptomatology
as the patients were admitted and discharged from the hospital.
Factor analysis of this data resulted in nine stable factors, with three factors
accounting for the majority of variance in total symptom scores. These three
factors were depression, somatization and paranoia. Examination of the factor
analysis clearly showed that there was some overlap between a factor labeled
hostility and another factor of paranoia. Based upon factor analytic results, and internal
reliability coefficients, we grouped three symptoms with the highest
reliabilities for each of the sub-scales in StressTest. We then modified the wording of
the items to fit the grammatical style of StressTest (Holcomb, 1983).
The third section of StressTest measures self-reported Level of Functioning. The
initial pilot sample for this section consisted of 300 patients in
two hospitals and one nursing home in the state of
Alabama. Patients were chosen at random from the
hospitals and nursing home. Again, the
research literature on measuring Level of Functioning with individuals
was thoroughly reviewed, and a large pool of items was put
together.
Again, using factor analysis and internal reliability coefficients, we selected
two major dimensions of level of functioning, and then we selected the most
reliable items to be incorporated into these two sub-scales of the level of
functioning section of StressTest (Holcomb, et al., 1994).
Reliability
Since StressTest was constructed using the internal reliability coefficient as a criteria
for item selection, in addition to factor loadings from factor analysis, it is
expected that even the scales with few items will have good reliability. In a
recent outcome evaluation of three chemical dependency outpatient centers,
reliability for all three major scales was clearly above
minimal standards.
Validity
The three scales of StressTest were developed separately with preliminary concurrent
and criterion validity established for each of the sub-sections. In order to
test the construct validity of StressTest, multi-dimensional scaling was done with a
sample of outpatients receiving counseling. The objective of multi-dimensional
scaling was to ascertain if the three major scales and 9 sub-scales could be
meaningfully plotted in a two-dimensional space and this was demonstrated.
In our studies, it is clear that other instruments, such as the Beck Depression
Inventory and the Zung Anxiety Inventory, are highly related to appropriate
sub-scales of StressTest. In addition, a consistent finding
is emerging that the paranoia/hostility scale is negatively related to outcome.
For example, multiple regression has indicated that high scores in paranoia/hostility
are negatively related to days of sobriety after chemical dependency treatment.
Studies also are consistently showing that satisfaction with treatment is predictive
of better treatment outcome. Another finding is that people who complete treatment
have higher scores on Quality of Life, and lower scores on Symptomatology, higher
scores on Level of Functioning, and higher treatment satisfaction, as
would be expected. These results argue for both construct and predictive validity of StressTest.
Scoring
StressTest contains three overall factors, and 9 sub-factors. The following list details
the items that are added together to form the subscales and the overall major
scales of StressTest, as well as the reliability scores and original research information.
OVERALL STRESSTEST SCORE
Overall Reliability: .91 (Coefficient Alpha)
Original Research: Holcomb, W. R., Beitman, B. D., Hemme, C. A., Josylin, A., & Prindiville, S. (1998). Use of
a new outcome scale to determine best practices. Psychiatric Services, 49, 583-585.
QUALITY OF LIFE
Self esteem Items 1-3
Social Support Items 4-5
Health Items 6-8
Activity Items 9-10
Overall Reliability: .82 (Coefficient Alpha)
Original Research: Holcomb, W. R., Morgan, P., Ponder, H. , Farrel, M. (1993). Development of a structured interview scale for measuring quality of life of the severely mentally ill. Journal of Clinical Psychology, 49, 830-840.
SYMPTOMATOLOGY
Depression Items 11-13
Anxiety Items 14-16
Paranoia/Hostility Item 17-19
Overall Reliability: .84 (Coefficient Alpha)
Original Research: Holcomb, W. R., Adams, N. A., & Ponder, H. M. (1983). Factor structure of the Symptom Checklist-90 with acute psychiatric inpatients. Journal of Consulting and Clinical Psychology, 51,535-538.
LEVEL OF FUNCTIONING
Disruptive Behavior Items 20-23
Living Skills Items 24-27
Overall Reliability: .64 (Coefficient Alpha)
Original Research: Holcomb, W. R., Mirilli, E., & Ahr, P. R. (1994). Reliability and concurrent validity of the Level of Care and Utlization Survey. Psychological Reports, 75, 779-786.
For More Information
For more information, e-mail us at
info@bhcinfo.com
or call 573-446-0405.
© Behavioral Health Concepts, Inc.
2716 Forum Blvd., Suite 4 Columbia, MO 65203
Phone: (573) 446-0405
Fax: (573) 446-1816
|
