Abstract:The objective of this study was to estimate the structure and relationships between four h ypothesized frailty dimensions (physical, emotional, cognitive, and social) and the extent to which personal and HIV-related factors and comorbidity associate with these frailty dimensions. This is a secondary analysis of an existing dataset arising from Positive Brain Health Now study (n = 856) in people aging with HIV (mean age: 52.3 ± 8.1 years). Structural equation modeling (SEM) models were applied to two cross-sections of the data: one at study entry and one at second visit, 9-month apart. Multidimensional frailty was modeled based on the combined Wilson-Cleary and International Classification of Functioning, Disability and Health framework. Four dimensions were operationalized with patient-reported and self-report measures from standardized questionnaires. The SEM model from the first visit was replicated using data from the second visit, testing measurement invariance. The proposed model showed acceptable fit at both visits (including no violation of measurement invariance). The final model for the first visit showed that sex, body mass index, HIV diagnosis pre-1997, current or nadir CD4 counts, and comorbidity did not associate with any frailty dimension; however, age (β range: 0.12-0.25), symptoms (β range: -0.35 to -0.58), and measured cognition (β range: 0.10-0.24) directly associated with all frailty dimensions. The model remained stable across the two visits. This study contributes evidence for operationalizing multidimensional frailty. Evidence-based interventions are available for many of the measures considered here, offering opportunities to improve the lives of people with frailty in the context of HIV.

In Support of Multidimensional Frailty: A Structural Equation Model from the Canadian Positive Brain Health Now Cohort