Division of Hematologic Malignancies and Cellular Therapy, Department of Medicine, Duke University School of Medicine, Durham, NC.
Duke Cancer Institute, Durham, NC.
Blood. 2018 Aug 16;132(7):717-726. doi: 10.1182/blood-2018-03-842575. Epub 2018 May 30.
Hospice provides high-quality end-of-life care, but patients with leukemias use hospice services less frequently than those with solid tumors. Transfusion dependence (TD) may hinder or delay enrollment, because hospice organizations typically disallow transfusions. We examined the association between TD and end-of-life outcomes among Medicare beneficiaries with leukemia. From the Surveillance, Epidemiology, and End Results-Medicare database, we selected beneficiaries with acute and chronic leukemias who died in 2001-2011. We defined TD as ≥2 transfusions within 30 days before death or hospice enrollment. End points included hospice enrollment and length of stay, reporting relative risk (RR) adjusted for key covariates. Among 21 033 patients with a median age of 79 years, 20% were transfusion dependent before death/hospice enrollment. Use of hospice increased from 35% in 2001 to 49% in 2011. Median time on hospice was 9 days and was shorter for transfusion-dependent patients (6 vs 11 days; < .001). Adjusting for baseline characteristics, TD was associated with a higher use of hospice services (RR, 1.08; 95% confidence interval [CI], 1.04-1.12) but also with 51% shorter hospice length of stay (RR, 0.49; 95% CI, 0.44-0.54). Hospice enrollees had a lower likelihood of inpatient death and chemotherapy use and lower median Medicare spending at end-of-life, regardless of TD status. In conclusion, relatively increased hospice use combined with a markedly shorter length of stay among transfusion-dependent patients suggests that they have a high and incompletely met need for hospice services and that they experience a barrier to timely referral. Policy solutions supporting palliative transfusions may maximize the benefits of hospice for leukemia patients.
临终关怀提供高质量的临终关怀服务,但白血病患者使用临终关怀服务的频率低于实体瘤患者。输血依赖(TD)可能会阻碍或延迟登记,因为临终关怀组织通常不允许输血。我们研究了 TD 与 2001-2011 年期间死于白血病的 Medicare 受益人的临终结局之间的关联。从监测、流行病学和最终结果-Medicare 数据库中,我们选择了在 2001 年至 2011 年期间死亡或接受临终关怀的急性和慢性白血病患者。我们将 TD 定义为在死亡/临终关怀登记前 30 天内进行≥2 次输血。终点包括临终关怀登记和住院时间,报告经过关键协变量调整的相对风险(RR)。在 21033 名中位年龄为 79 岁的患者中,20%在死亡/临终关怀登记前存在输血依赖。临终关怀的使用率从 2001 年的 35%增加到 2011 年的 49%。中位临终关怀时间为 9 天,输血依赖患者更短(6 天比 11 天;<.001)。调整基线特征后,TD 与更高的临终关怀服务使用率相关(RR,1.08;95%置信区间[CI],1.04-1.12),但也与临终关怀时间缩短 51%相关(RR,0.49;95% CI,0.44-0.54)。无论 TD 状态如何,临终关怀登记患者住院死亡和化疗使用率较低,临终时 Medicare 支出中位数也较低。总之,相对增加的临终关怀使用,加上输血依赖患者明显缩短的临终关怀时间,表明他们对临终关怀服务的需求很高,但没有得到充分满足,并且他们面临及时转介的障碍。支持姑息性输血的政策解决方案可能会最大限度地提高临终关怀服务对白血病患者的益处。