Factor Structure of Functional Assessment of Chronic Illness Therapy: Spiritual Well-Being Scale in Patients with Heart Failure Depends on Method Used

Dear Editor:

The Functional Assessment of Chronic Illness Therapy—Spiritual Well-Being Scale (FACIT-Sp) was developed to assess spiritual well-being in chronically ill patients and comprises 12 items, each rated on a 5-point scale by the respondent. An early study of cancer patients performed principal components analysis (PCA) and proposed that the FACIT-Sp contains two subscales, or factors. ¹ The first eight items constituted the meaning/peace factor, and the last four items constituted the faith factor. A later study, again in cancer patients, used maximum likelihood exploratory factor analysis (EFA) and proposed a 3-factor model for the FACIT-Sp, which was argued to be superior to the 2-factor model. ² In the 3-factor model, the meaning/peace factor was split into a meaning factor and a peace factor, each containing four items.

Consensus regarding the factor structure of the FACIT-Sp has not emerged. Some studies supported the 3-factor model, whereas others argued that the FACIT-Sp is best understood as having two factors.^3,4 Almost all existing studies focused exclusively on cancer patients, and none analyzed heart failure patients.

With the lack of factor structure consensus and lack of prior factor structure studies in heart failure, we examined the factor structure of the FACIT-Sp in patients with heart failure using PCA and EFA.

Baseline data collected from 217 participants in the CASA (Collaborative Care to Alleviate Symptoms and Adjust to Illness) trial (clinicaltrials.gov identifier: NCT01739686) were used. ⁵ We used two different methods of extraction: maximum likelihood EFA (three factors max) and PCA (eigenvalue ≥1), both followed by promax/oblique rotations. Both extraction methods have been used previously on the FACIT-Sp,^1,2 and using both allowed us to compare the results and determine whether the factor structure varied based on the method chosen.

The factor loadings obtained using both methods are shown in Table 1, broken down by method. EFA revealed the presence of three factors, which generally corresponded to the meaning, peace, and faith factors that have previously been reported, ² but item 4, a negatively worded item, exhibited weak loading onto the peace factor. PCA suggested a combined meaning/peace factor and a faith factor, similar to the original 2-factor model, ¹ but the two negatively worded items were excluded and formed a new but likely not meaningful factor.

Thus, two different factor structures seem possible, and the factor structure depended heavily on the method used. This was true not only in this study but also in previous studies. PCA yielded a combined meaning/peace factor and a faith factor in both our study and the study that proposed the 2-factor model. ¹ EFA yielded separate meaning, peace, and faith factors in both our study and the study that proposed the 3-factor model. Therefore, methodological differences appear to at least partially account for different findings about the factor structure of the FACIT-Sp. Confirmatory factor analysis of the FACIT-Sp in heart failure patients may be a valuable future direction.