Care partner burden and emergency service use: what is the link?
Allison Lindauer, Sarah Gothard, Hannah Bernard, Joel S. Steele- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Persons with dementia (PWD) are significantly more likely to use emergency services (defined here as emergency department or urgent care visits) than age‐matched controls (1). The bi‐directional relationship between care partner (CP) burden and their PWD’s emergency service use is not well‐understood (2). A possible limiting factor is the infrequent assessment of emergency service use by the PWD and corresponding CP burden. Here, we used high frequency longitudinal data from the telehealth‐based Tele‐STELLA (Support via TEchnology: Living and Learning with Advancing ADRD) study (NIA R01AG067546) to examine the relationship between CP burden and PWD emergency service use.
Method
CPs (n = 80; age = 65.7±11.6 years) completed weekly online “Orbit” surveys (3). The survey measures burden (ZBI‐4 [9±3.48 points]), cost‐of‐care, medication use, and emergency service use (4). Using R software (5), descriptive statistics and generalized mixed effect models were developed to examine the effects of CP burden on PWD emergency service use.
Result
A median of 27 (min = 1;max = 61) weeks of CP responses were captured. Simple generalized mixed effects models showed a significant positive relationship between ZBI‐4 scores and emergency service use the week of event occurrence (ß = 0.0853; p = 0.04). As ZBI‐4 scores increased, the greater the likelihood of an event. Additionally, a mixed effects model that examined the 5 weeks prior to emergency service use showed the absolute magnitude of change of ZBI‐4 scores had a negative relationship to the events (ß = ‐0.393;p = 0.00;OR = 0.675). Taking the results of the two models together, as the range of burden narrowed towards higher ZBI‐4 scores, the chance of emergency service use increased.
Conclusion
High frequency electronic self‐reported data in this telehealth‐based study allowed for identification of emergent events closer to real time than the standard practice. As the range of CP burden scores decreased, trending toward higher burden, there was an increased likelihood of emergency service use occurring for the PWD. This information can be used to anticipate the likelihood of future events and create targeted interventions both for the CP and PWD prior to the need of emergency service use. Limitations in this analysis and gaps in research on this topic warrant further investigation.