Collaborative care management reduces disparities in dementia care quality for caregivers with less education

Abstract

Objectives: To examine educational gradients in dementia care and whether the effect of a dementia collaborative care management intervention varied according to the educational attainment of the informal caregiver.

Design: Analysis of data from a cluster-randomized controlled trial.

Setting: Eighteen clinics in three healthcare organizations in southern California.

Participants: Dyads of Medicare recipients aged 65 and older with a diagnosis of dementia and an eligible caregiver.

Intervention: Collaborative care management for dementia.

Measurements: Caregiver educational attainment, adherence to four dimensions of guideline-recommended processes of dementia care (assessment, treatment, education and support, and safety) before and after the intervention, and the adjusted intervention effect (IE) for each dimension stratified according to caregiver education. Each IE was estimated by subtracting the difference between pre- and postintervention scores for the usual care participants from the difference between pre- and postintervention scores in the intervention participants.

Results: At baseline, caregivers with lower educational attainment provided poorer quality of dementia care for the Treatment and Education dimensions than those with more education, but less-educated caregivers had significantly more improvement after the intervention on the assessment, treatment, and safety dimensions. The IEs for those who had not graduated from high school were 44.4 for the assessment dimension, 36.9 for the treatment dimension, and 52.7 for the safety dimension, versus 29.5, 15.7, and 40.9 respectively, for college graduates (P < .001 for all three).

Conclusions: Collaborative care management was associated with smaller disparities in dementia care quality between caregivers with lower educational attainment and those with more education.

Figure. Adjusted Mean Care Processes for Four Quality of Care Dimensions

LEGEND: REF = reference group, INT = Intervention, UC = Usual Care. Analytic sample is 396. Adjusted values at baseline and follow-up across the four caregiver education categories were estimated using fully adjusted models which included: patient age, sex, education, comorbidity, race/ethnicity, dementia severity; caregiver age, sex, race/ethnicity, social support, mastery, satisfaction with care; whether patient and caregiver live in same residence; whether the patient visited a primary care physician, geriatrician, or neurologist in the prior 6 months; study site; and baseline process score. P-values are for pairwise comparisons of intervention effects for each caregiver education category compared with the less than high school category (=reference group; REF). Intervention effects were estimated by subtracting the difference between adjusted baseline and follow up scores for the usual care participants from those of the intervention participants for the fully adjusted model. Standard errors for each Intervention Effect (IE) was estimated using bootstrapping. Intervention effects are significant for all four care dimensions across all caregiver education categories (i.e. the IE is different from zero).