Abstract
Purpose :
To compare the psychometric properties of a reduced (22 statements) to a long version (47 statements) of the Glaucoma Treatment Compliance Assessment Tool (GTCAT), which is based on the organizational structure of the Health Belief Model.
Methods :
We created a reduced GTCAT by removing statements that: 1) did not appear in either the final Principal Components Analysis (PCA) model and 2) did not have a univariate association with adherence; or 3) were highly correlated with a correlation coefficient of .60 or higher with one another. To verify we did not remove key statements, we used PCA on the 25 removed or remaining statements to verify a stable organizational structure and used univariate analysis to determine if we removed a potentially predictive statement. Last, we assessed the construct validity of the remaining statements using PCA and assessed their predictive validity using univariate and multivariate analysis.
Results :
Of the original 47 GTCAT statements, we removed 22 because they did not appear in either the PCA analysis or the univariate regression analysis. We removed an additional 3 statements due to high intercorrelation (r = .76-.80). PCA of the 25 excluded statements resulted in no components extracted, and univariate regression indicated no predictive statements. As a whole the removed statements were not predictive (R2=.035, p = .26) of adherence. Clinical acumen suggested re-inclusion of 4 statements with alternate wording. PCA of the remaining 22 items resulted in the extraction of 5 components (knowledge, susceptibility, cues to action, self-efficacy, barriers). Multiple regression showed 6 statements associated with higher adherence (adjusted R2 = 0.27, p<.001), and as a whole the GTCAT-S predicted adherence (R2= .167, p = .01).
Conclusions :
A reduced GTCAT maintained an organizational structure consistent with the Health Belief Model and high predictive ability. Future glaucoma medication adherence studies can use the reduced model to determine whether a change in health behavior as measured by the GTCAT is associated with a change in adherence.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.