Hospital admission following an emergency department visit may increase health care spending without improving mortality among older patients with dementia.
In a study, investigators analyzed 872,085 emergency department visits among Medicare fee-for-service beneficiaries aged 66 years or older with dementia between 2017 and 2019. Overall, 55% (n = 482,208) of the visits resulted in hospital admission. To address differences in disease severity between the patients who were and were not admitted, the investigators used a quasi-experimental instrumental variable approach based on emergency physicians' propensity to admit patients.
The primary outcomes included mortality, inpatient days excluding the index admission, and health care spending within 30 and 90 days of the emergency department visit. In the instrumental variable analysis, the investigators found no evidence that hospital admission was associated with lower mortality or fewer subsequent inpatient days. However, hospital admission was associated with approximately $2,500 higher health care spending within 30 days and approximately $4,000 higher spending within 90 days.
The increase in spending was driven largely by greater use of postacute care services, leading to higher spending on home health care and skilled nursing facility care, while the investigators found no evidence of meaningful differences in inpatient, outpatient, physician service, or hospice spending.
Among long-term nursing home residents with dementia, the investigators found that hospital admission was not associated with differences in physical or cognitive function scores 90 days following the emergency department visit. They noted that estimates for these outcomes were imprecise because of the smaller subgroup sample.
The investigators also identified substantial variation in admission practices among emergency physicians caring for patients with dementia. After adjustment for patient characteristics, admission propensity ranged from 38% among physicians at the 10th percentile to 70% among physicians at the 90th percentile. Patient characteristics were similar among physicians with higher and lower admission propensities, supporting the study's quasi-experimental design.
The findings should be interpreted in the context of several limitations, according to the investigators. As with any instrumental variable analysis, key assumptions cannot be directly verified. The findings applied to patients whose admission status was influenced by physician admission tendencies and may not generalize to patients with more severe disease, Medicare Advantage enrollees, or populations outside the United States. The study also could not assess outcomes such as caregiver burden.
"Our findings indicate that some hospital admissions may provide little value among [patients with dementia]. Future research that explores the underlying mechanisms should be informative to reduce low-value hospital admissions," wrote lead study author Ryo Ikesu, MD, PhD, of the Department of Health Promotion and Behavioral Sciences at the Graduate School of Medicine at Kyoto University in Japan, and colleagues.
The study received no direct funding. The study authors reported institutional funding unrelated to the study. Additional disclosures are available in the study.
Source: Annals of Internal Medicine
