Assessment Procedures
A variety of barrier assessment procedures have been developed
for a spectrum of diverse cancer screening behaviors, other
conditions including diabetes, HIV, pain, and panic attacks,
and specific behaviors including medication adherence, healthy
eating, physical activity, smoking cessation, and weight management.
A PubMed search for "barriers to adherence" produced
1,002 articles and a search for "barriers to cancer screening"
produced 1,346. Various procedures have been used to assess
barriers including qualitative interviews, open-ended questionnaires,
and structured surveys. Barrier items are sometimes based
on theory, sometimes on experience, and occasionally on frequency
or strength of endorsement in prior research. Several research
programs have developed and revised barrier lists over time
as a result of data obtained in prior studies. For example,
Champion
et al. (2007)
xClose
Champion, V., Skinner, C. S., Hui, S., Monahan, P., Juliar,
B., Daggy, J. et al. (2007). The effect of telephone versus
print tailoring for mammography adherence. Patient Education
and Counseling, 65, 416-423. have evolved a measure
of barriers to mammography over time, as
have McCauley
and colleagues (1998)
xClose
McCauley, E. & Mihalko, S. L. (1998). Measuring exercise-related
self-efficacy. In J.L.Duda (Ed.), Advances in sport and
exercise psychology measurement (Morgantown, WV: Fitness
Information Technology). for physical activity,
and Glasgow
et al. (2001)
xClose
Glasgow, R. E., Gillette, C., & Toobert, D. (2001). Psychosocial
barriers to diabetes self-management and quality of life.
Diabetes Spectrum, 14, 33-41. for diabetes
self-management. Ideally, barrier items would be
culturally appropriate, based on both theory and prior research,
and pilot tested with one's target population.
Responses to barrier items have been assessed using scales
of frequency, probability, strength of barrier, level of concern,
how often the barrier has prevented the desired behavior,
and finally self-efficacy level. A few studies have
investigated use of a combination of frequency and strength
responses, but this strategy has not enhanced predictive validity
compared to that of a frequency scale alone (Glasgow
et al., 2001
xClose
Glasgow, R. E., Gillette, C., & Toobert, D. (2001). Psychosocial
barriers to diabetes self-management and quality of life.
Diabetes Spectrum, 14, 33-41.).
Validity
A number of analysis procedures have been used to validate barriers
measures, but the typical analysis relies on cross-sectional
correlations (Glasgow
et al., 2001
xClose
Glasgow, R. E., Gillette, C., & Toobert, D. (2001). Psychosocial
barriers to diabetes self-management and quality of life.
Diabetes Spectrum, 14, 33-41.). To justify
use as a practical method of predicting future behavior, a
better strategy would be to determine that a barrier instrument
explains incremental variance over and above that explained
by simple and more easily obtained variables, such as demographics
and past behavior. (For explanatory purposes and to understand
why, for example, demographic subgroups may have different
levels of cancer screening, a barrier instrument may still
be useful). In addition, stronger conclusions can be made
based upon prospective than cross-sectional analyses of predictive
validity. The issue of how respondents make judgments about
barriers is of particular concern. Especially problematic
are scales and analyses that a) ask the respondent to rate
how often a given barrier caused the person
to not be able to engage in behavior X (see Glasgow
et al., 2001
xClose
Glasgow, R. E., Gillette, C., & Toobert, D. (2001). Psychosocial
barriers to diabetes self-management and quality of life.
Diabetes Spectrum, 14, 33-41., for more details),
and then b) use this rating to predict concurrent self-reports
of behavior X. For example, part of my judgment of
"how often the weather prevented me from exercising" is based
upon my frequency of exercise. It then does not make sense
to use such a barrier measure to predict exercise frequency,
since the self-assessment of exercise frequency is part of
both barrier and exercise constructs.
|