Systemic anticancer therapy near the end of life and acute health care utilization: A target trial emulation study using SEER–Medicare.

Abstract

612 Background: Use of Systemic anticancer therapy (SACT) at end-of-life (EOL) has been associated with higher costs, increased hospitalizations (HL), emergency department (ED), and intensive care unit (ICU) utilization and lower hospice (HP) use all of which may adversely affect quality of life. Prior studies have looked at the association SACT use and all acute care use in the last 30 days of life, regardless of timing. Using target trial emulation to reduce biases common in observational studies, we sought to examine the temporal relationship between SACT and acute care utilization in the last 30 days of life, while adjusting for time-varying confounders and patient clinical status changes in SEER-Medicare patient population. Methods: We identified Medicare beneficiaries diagnosed with breast, colorectal, lung, prostate, bladder, cervical, kidney, liver, ovarian, pancreatic, melanoma, or uterine cancer from 2005-2019 who died between 2015 and 2020. This TTE compared two treatment strategies: initiation of a new line of SACT versus no SACT during a treatment window 30 to 15 days before death. The primary outcomes were acute health care utilization within 14 days of death for HL, ED, ICU and HP. Propensity scores for SACT initiation were estimated using logistic regression incorporating age, sex, race, cancer type, SEER stage, Charlson comorbidity index, number of ICU and inpatient admissions occurring 60 to 31 days before death. Inverse probability of treatment weighting (IPTW) was applied to balance baseline characteristics between SACT and non-SACT groups. Results: The cohort included 315,089 patients, of whom 14,022 (4.45%) received SACT between 30 and 15 days before death. After IPTW adjustment, SACT use was associated with significantly higher odds of downstream acute care utilization at EOL, including: ED visits: odds ratio (OR), 2.56 (95% CI, 2.47–2.66; P < .0001); ICU admissions: OR, 1.93 (95% CI, 1.86–2.01; P < .0001); and HL: OR, 2.60 (95% CI, 2.51–2.70; P < .0001) and lower odds of HP: OR, 0.69 (95% CI, 0.66–0.71; P < .0001) compared with no SACT. Conclusions: By controlling for temporal factors and other co-variates, we found that among Medicare beneficiaries with solid tumors who received SACT between 30 and 15 days before death had significantly higher downstream acute care use. These findings provide evidence for the downstream impact of late-stage SACT use and can offer incentive to promote earlier goals-of-care discussions as a strategy to improve the quality of cancer care at EOL. Additionally, while this analysis cannot definitively state that EOL SACT causes acute care use, the fact that patients who receive SACT within the last 30 days of life are more likely to experience downstream acute care use, suggests that SACT use may contribute to the likelihood of downstream acute care events.