Abstract
While Stress Process Models of family caregiving have been examined extensively, little focus has been placed on caregiver’s actual management of care. We consider whether caregiving style classified previously through k-modes machine learning models and based on cognitive-behavioral approaches to care impact caregivers’ experiences of care-related stress and well-being. The three previously identified styles include: Adapters- strong dementia understanding and adaptability, encouraging behavioral approach; Managers- poor dementia understanding and adaptability, critical behavioral approach; and Avoiders- moderate dementia understanding and adaptability, passive behavioral approach. Participants included 100 primary family caregivers for PWDs who were on average 64 years old, 74% female, and 18% non-White. Utilizing linear regressions, each caregiving style was considered a key predictor (reference: Adapters) of the Zarit Burden Interview (ZBI), Caregiver Assessment of Function and Upset (CAFU) upset score, Neuropsychiatric Inventory (NPI-C) distress scale, and Positive and Negative Affect scale (PANAS) controlling for dementia severity, care duration, co-residency, and demographics. Relative to Adapters, Managers had more CAFU upset (β=0.4, p<.001), more NPI-C distress (β=0.4, p<.001), and greater burden (ZBI) (β=0.3, p<.001). Avoiders showed significantly greater CAFU upset than Adapters (β=0.2, p<.05). Positive affect was not associated with caregiving style. Caregiving styles associated with less understanding and adaptability and a more critical behavioral approach showed worse caregiving outcomes accounting for dementia severity. Results can inform a nuanced approach to tailoring and targeting interventions based on caregiver styles with the goal of reaching caregivers at risk for poor outcomes and ultimately leading to significant public health impact.