Frailty and Social and Behavioral Determinants of Health: Algorithm Refinement and Pattern Validation

Existing frailty and social and behavioral determinants of health (SBDH) algorithms were refined and used to examine SBDH and frailty groups, revealing patterns in interventions and outcomes of older adults in a large community-based care dataset. The dataset was randomly split into training (n = 2,881) and testing (n = 1,441) sets. The training set was used to visually identify patterns in associations among SBDH, frailty, intervention doses, and outcomes, and the testing set was used to validate the patterns. Seven valid patterns were identified, showing increases in SBDH and frailty were associated with poorer health outcomes and more interventions (all p < 0.01). Findings suggest that the refined SBDH and frailty algorithms facilitate the identification of older adults with SBDH and frailty for intervention tailoring.