Latent Factor Structure of Outcome Measures Used in the HABIT® Mild Cognitive Impairment Intervention Programs

Article type: Research Article

Authors: DeFeis, Brittany^{a; 1} | Ying, Gelan^{a; 1} | Kurasz, Andrea M.^a | De Wit, Liselotte^a | Amofa Sr., Priscilla^a | Chandler, Melanie^b | Locke, Dona^c | Shandera-Ochsner, Anne^d | Phatak, Vaishali^e | Smith, Glenn^{a; *}

Affiliations: [a] Department of Clinical and Health Psychology, University of Florida, Gainesville, FL, USA | [b] Department of Psychiatry and Psychology, Mayo Clinic Florida, Jacksonville, FL, USA | [c] Department of Psychiatry and Psychology, Mayo Clinic Arizona, Scottsdale, AZ, USA | [d] Department of Psychiatry and Psychology, Mayo Clinic Midwest, LaCrosse, WI, USA | [e] Department of Neurological Sciences, University of Nebraska Medical Center, Omaha, NE, USA

Correspondence: [*] Correspondence to: Glenn Smith, PhD, Department of Clinical and Health Psychology, University of Florida, P.O. Box 100165, Gainesville, FL 32610-0165, USA. Tel.: +1 352 273 6155; E-mail: glennsmith@ufl.edu.

Note: [1] These authors contributed equally to this work.

Abstract: Background:In Alzheimer’s disease and related disorders (ADRD) research, common outcome measures include cognitive and functional impairment, as well as persons with mild cognitive impairment (pwMCI) and care partner self-reported mood and quality of life. Studies commonly analyze these measures separately, which potentially leads to issues of multiple comparisons and/or multicollinearity among measures while ignoring the latent constructs they may be measuring. Objective:This study sought to examine the latent factor structure of a battery of 12-13 measures of domains mentioned above, used in a multicomponent behavioral intervention (The HABIT® program) for pwMCI and their partners. Methods:Exploratory factor analysis (EFA) involved 214 pwMCI-partner pairs. Subsequent Confirmatory factor analyses (CFA) used 730 pairs in both pre- and post-intervention conditions. Results:EFA generated a three-factor model. Factors could be characterized as partner adjustment (29.9%), pwMCI adjustment (18.1%), and pwMCI impairment (12.8%). The subsequent CFA confirmed our findings, and the goodness-of-fit for this model was adequate in both the pre- (CFI = 0.937; RMSEA = 0.057, p = 0.089) and post-intervention (CFI = 0.942; RMSEA = 0.051, p = 0.430) groups. Conclusion:Results demonstrated a stable factor structure across cohorts and intervention conditions suggesting that three broad factors may provide a straightforward and meaningful model to assess intervention outcome, at least during the MCI phase of ADRD.

Keywords: Behavioral intervention, caregiver burden, factor analysis, functional status, mild cognitive impairment

DOI: 10.3233/JAD-210582

Journal: Journal of Alzheimer's Disease, vol. 84, no. 1, pp. 193-205, 2021