Advance care planning (ACP) is a process that helps people prepare to make decisions about their future medical care. We sought to understand who was received billed ACP visits and measure the association with health care utilization, cost, and mortality. We used a randomly sampled 20 % cohort of Medicare fee-for-service (FFS) beneficiaries' files to conduct a retrospective cohort study. Beneficiaries with a billed ACP visit were matched to controls using a 2-stage propensity score matching process that included assigning a pseudo-ACP visit date for controls. Outcomes included healthcare utilization, mortality, and total medical cost per month. We used descriptive statistics for univariate analysis and fit multilevel logistic regression, multilevel linear regression, or Cox regression models. We identified 183,513 beneficiaries who received any billed ACP visit and 550,539 matched controls. Of those who had a ACP visit, the mean age was 76.5 years and high-risk comorbidities were common: 16 % dementia, 10 % congestive heart failure, 10 % cancer. Beneficiaries who had an ACP visit had slightly more health care utilization than controls. Beneficiaries who had an ACP visit were more likely to die (3.1% vs. 1.0 %, p < 0.01, OR=3.0, 95 %CI 2.9–3.2) in the unadjusted and adjusted analyses compared to matched controls. Total monthly medical costs were 33 % higher among beneficiaries who had an ACP visit. Our results suggest that ACP visits may be preferentially utilized amongst individuals with higher risk of mortality. There may be an opportunity to increase ACP visits among older adults at lower risk for mortality. This article suggests that ACP visits are likely targeted to older adults with a higher risk of mortality than those at lower risk of mortality suggesting an opportunity to reach people before they are facing end-of-life decisions.