11 Evaluating a Comprehensive Care Management Program for Dementia: Three Years into the Care Ecosystem Program.

Abstract

Objective:

Dementia prevalence and its costs to the health system continue to rise, highlighting the need for comprehensive care programs. This study evaluates the Care Ecosystem Program (CE) for dementia (memory.ucsf.edu/Care-Ecosystem) in New Orleans, LA and surrounding areas.

Participants and Methods:

The sample consisted of persons with dementia (PWD) and caregiver (CG) dyads enrolled in the CE from February-2019 to June-2022. Participants had a dementia diagnosis, lived in the community, and had at least one emergency department (ED) visit or hospitalization in the year prior. Healthcare utilization data was collected through self-report and electronic medical records. Dementia rating scales (QDRS, NPIQ) and caregiver wellbeing questionnaires (ZBI-12; PHQ-9; Self-Efficacy) were collected at baseline, 6-months, and 12-months. Dyads received monthly calls providing individualized care-management. One-way repeated measures Anovas were performed to identify change in utilization and caregiver wellbeing at 6-months and 12-months compared to baseline. Partial n2 effect sizes and post-hoc Bonferroni were calculated. Healthcare utilization extreme outliers were winsorized to the 95th percentile and a p-value of .05 was set.

Results:

A total of 150 dyads completed the program. PWD's age averaged 81 years (SD=8); they were mostly female (65%), White (63%), and had at least a High School education or higher (88%). CG's age averaged 65 years (SD=11.5); they were predominantly female (77%), White (63%), and had more than 12-years of education (70%). Half of the CGs were adult children (50%), followed by spouse/partners (41%). The QDRS indicated mild-moderate dementia severity, PWD had on average five neuropsychiatric symptoms, and Alzheimer's Disease was the most frequent diagnosis (35%).

A statistically significant decrease occurred in ED visits [F(1, 115)=14.970, p<.001, n2=.115] from baseline to 6-months (MD=1.043, p<.001) and 12-months (MD=.621, p<.001), while an increase was noted when comparing 12-month to 6-month data (MD=.422, p<.001). A similar pattern was observed for hospitalizations [F(1,115)=19.021, p<.001, n2=.142] were admissions were reduced significantly compared to baseline (6-month MD=.483, p<.001; 12-month MD=3.88, p<.001) and an increase was seen after the 6-month mark (MD=.095, p<.001). Caregiver self-efficacy significantly improved [F(1,115)=15.478, p<.001, n2=.119] from baseline to 6-months into the CE (MD=-1.457, p<.001) and was maintained a year after enrollment (MD=-1.474, p<.001). There were no differences in self-efficacy when comparing 6-month and 12-month data. Robust effect sizes were noted for all results previously reported. No other caregiver wellbeing measures showed significant changes over the three time points.

Conclusions:

CE successfully reduces healthcare utilization and improves caregiver self-efficacy for PWD-CG dyads 6-months and 12-months after enrollment. The utilization increase noted from the 6-month to the 12-month mark does not surpass baseline rates. This pattern is also consistent with literature reporting that healthcare utilization rises with the progression of dementia. More research is needed to identify potential moderating factors in the relationship between dementia progression and utilization. Future research will also benefit from including control groups to further understand the impact of comprehensive care programs for dementia.