Healthcare utilization and morbidity preceding a dementia classification: the Atherosclerosis Risk in Communities (ARIC) study.

BACKGROUND

Dementia is frequently underdiagnosed in its early stages, but changes in healthcare utilization may precede diagnosis by several years. We examined healthcare patterns and chronic conditions before cognitive impairment among the Atherosclerosis Risk in Communities (ARIC) study participants to better understand this pre-diagnostic period.

METHODS

We included ARIC visit 6 (2016-2017) participants with continuous 5-year prior Medicare coverage (2011-2018) and used negative binomial regression to compare longitudinal trends in hospitalization in participants with Part A coverage, and emergency department (ED) and ambulatory care use in participants with Part A&B coverage, in the 5 years prior to and 1 year following an incident dementia according to ARIC syndromic classification. Inverse probability weights were used to balance participants with and without incident dementia on baseline age, race, sex, education, and prevalent comorbidities. Principal diagnosis discharge codes were used to assess chronic conditions associated with hospitalization and ED visits, whereas all available diagnosis codes were used for ambulatory care visits.

RESULTS

Among 3,923 eligible participants (mean age 79.6 ± 4.8 years; 59% women; 24% Black), 325 (8.3%) were classified as incident dementia at ARIC visit 6 or 7. During 5 years prior to visit 6, participants with incident dementia compared with non-dementia participants, had on average higher use of inpatient (0.12 vs. 0.08 hospitalizations per 6-month) and ED services (0.28 vs. 0.18 visits per 6-month) but lower ambulatory care use (3.53 vs. 3.76 visits per 6-month). Participants with incident dementia experienced more frequent hospitalizations for acute myocardial infarction (3.8% vs. 1.7%), chronic kidney disease (4.1% vs. 2.6%), congestive heart failure (4.7% vs. 1.6%), and stroke/TIA (8.5% vs. 2.6%), and more frequent ambulatory care visits for atrial fibrillation (23% vs. 15%), than non-dementia participants within 5 years prior to incident dementia.

CONCLUSIONS

Compared with non-dementia participants, those with incident dementia had higher average hospitalization and ED utilization but lower ambulatory care visits, with differences observable 3 years before incident dementia. Cardiovascular disease and concomitant cerebrovascular disease are important contributors to dementia. Close monitoring of healthcare-seeking patterns may help identify individuals at high risk of dementia earlier in the disease course.

TRIAL REGISTRATION: Not applicable.