JMIR Publications

ABSTRACT

Background:

Most efforts to identify caregivers for research use passive approaches like self-nomination. We describe an approach where the EHR can help identify, recruit, and increase diverse representation of caregivers.

Objective:

Few health systems have implemented systematic processes to identify caregivers. We aimed to evaluate an electronic health record (EHR) algorithm for identifying Veterans with caregivers.

Methods:

We identified initial cohorts of Veterans likely to need supportive care from friends or family based with pre-defined EHR referrals for home and community care. Veterans were contacted assess whether the Veteran had an unpaid caregivers; unpaid caregivers were then contacted and offered enrollment in a caregiver survey. We compared Veteran characteristics from the EHR across these referral, screening, and recruitment groups using descriptive statistics and logistic regression models.

Results:

Of 12,212 Veterans identified through EHR referrals, 2,134 (17.4%) were selected for screening and 1,367 (11.2%) answered phone screening; 813 (60%) of those screened had a caregiver, and 435 (53%) caregivers participated in a survey. Married veterans had increased odds of having a caregiver (adjusted OR 2.63 [95%CI 1.65-4.24]) or had an adult day health care referral (adjusted OR 3.06 [95%CI 1.38 – 7.76]) or a respite care referral (adjusted OR 2.21 [95%CI 1.45-3.44].) Caregivers of Veterans with dementia had increased odds of participating in the survey (adjusted OR 1.78 [95%CI 1.20-2.65]).

Conclusions:

The EHR algorithm process is systematic, resource intensive, and imperfect. Sixty percent of successfully screened Veterans had an unpaid caregiver. Implementing discrete caregiver fields in the EHR would support more efficient systematic identification of caregivers. Clinical Trial: ClincalTrials.gov Identifier: NCT03474380.