Earlier and more frequent hospice use may be associated with a lower probability of inpatient death and a reduction in aggressive care at end of life (EOL), a study in Blood Advances suggests.
The study, led by Pamela Egan, MD, from Rhode Island Hospital, reviewed de-identified data of Medicare beneficiaries with acute or chronic leukemias, multiple myeloma, myeloproliferative neoplasm (MPN), myelodysplastic syndromes (MDS), and all subtypes of lymphoma who were diagnosed at age 66 or older. Only patients who had died between 2008 and 2015 with a documented death attributed to the hematologic malignancy were included. All patients had survived for at least 30 days from time of diagnosis.
Among the 34,088 Medicare beneficiaries included in the study, 8,859 had leukemia (26.0%), 6,750 had myeloma (19.8%), 4,941 had MDS/MPN (14.5%), and 13,538 had lymphoma (39.7%).
The following indicators of aggressive EOL care were identified from inpatient and outpatient Medicare claims:
- death in an acute care hospital
- intensive care unit (ICU) admission in the last 30 days of life
- chemotherapy administration in the last 14 days of life
More than half of patients (n=19,267; 56.5%) used hospice services prior to death, for a median of 9 days. Among the patients who used hospice services, approximately 33.0% died in the acute hospital setting, 36.8% had an ICU admission in the last 30 days of life (often before hospice enrollment), and 13.3% received intravenous or oral chemotherapy within the last 14 days of life. At the time of death for hospice enrollees, the median age was 80 years.
Hospice enrollment was more frequent for patients with poor performance status (71.9% vs. 28.1%), prior dementia (64.8% vs. 35.2%), and survival >6 months from diagnosis (p<0.0001 for all).
A significantly lower proportion of hospice enrollees had received chemotherapy within a 1-year period before death (61.0% vs. 68.5%; p<0.001). In addition, the use of hospice services was significantly associated with higher palliative needs, such as requirements for opioid prescriptions (adjusted risk ratio [RR]=1.40), blood transfusions (adjusted RR=1.32), and mean number of physician visits (adjusted means ratio = 1.21) within a 30-day period before hospice enrollment or death.
The investigators noted that transfusion dependence represents a significant barrier to hospice use in hematologic cancers, suggesting that earlier enrollment may be facilitated by more adequate coverage for palliative transfusions by a Medicare hospice.
In a multivariable model, hospice enrollment was associated with less aggressive EOL care and lower spending, according to the investigators. Patients in hospice had a 96% lower probability of inpatient death, a 44% lower probability of an ICU stay in the last 30 days of life, and a 62% reduction in chemotherapy use in the last 14 days of life.
On average, the use of hospice services was associated with 41% fewer days spent as an inpatient during the last month of life, as well as 38% lower mean Medicare spending in the last month of life, the authors reported.
A limitation of the study was its reliance on administrative claims, which did not provide details on the clinical reasons for the use of hospice services, ICU admissions, or hospitalizations.
According to the study investigators, “the association between hospice use and EOL outcomes supports the idea that inferior EOL care quality outcomes in patients with blood cancers may be more attributable to biases and preferences of hematologists and patients with blood cancers or to more global differences in the overall responsiveness of hematologic malignancies to treatment as compared with solid tumors, rather than any fundamentally different disease-specific needs.”
The authors report no relevant conflicts of interest.
Egan PC, LeBlanc TW, Olszewski AJ. End-of-life care quality outcomes among Medicare beneficiaries with hematologic malignancies. Blood Adv. 2020;4:3606-3614.