Healthcare Resources Utilization throughout the Last Year of Life after Acute Myocardial Infarction

Healthcare resource utilization (HRU) peaks in the last year-of-life, and accounts for a substantial share of healthcare expenditure. We evaluated changes in HRU and costs throughout the last year-of-life among AMI survivors and investigated whether such changes can predict imminent mortality. This retrospective analysis included patients who survived at least one year following an AMI. Mortality and HRU data during the 10-year follow-up period were collected. Analyses were performed according to follow-up years that were classified into mortality years (one year prior to death) and survival years. Overall, 10,992 patients (44,099 patients-years) were investigated. Throughout the follow-up period, 2,885 (26.3%) patients died. The HRU parameters and total costs were strong independent predictors of mortality during a subsequent year. While a direct association between mortality and hospital services (length of in-hospital stay and emergency department visits) was observed, the association with ambulatory services utilization was reversed. The discriminative ability (c-statistics) of a multivariable model including the HRU parameters for predicting the mortality in the subsequent year, was 0.88. In conclusion, throughout the last year of life, hospital-centered HRU and costs of AMI survivors increase while utilization of ambulatory services decrease. HRUs are strong and independent predictors of an imminent mortality year among these patients.

Automated summary

Healthcare resource utilization (HRU) is highest in the last year-of-life and accounts for a significant portion of healthcare expenditure. We examined changes in HRU and costs during the last year-of-life among survivors of acute myocardial infarction (AMI) and assessed whether these changes can predict imminent mortality. Our retrospective analysis included AMI patients who survived at least one year. Mortality and HRU data were collected over a 10-year follow-up period. We found that HRU parameters and total costs were strong predictors of mortality during the following year. In addition, the utilization of hospital services was directly associated with mortality, while the association with ambulatory services utilization was reversed. The discriminative ability of a multivariable model including HRU parameters to predict mortality in the subsequent year was 0.88.