AUTHORS
Hari S. Raman, MD, MBA, Elizabeth Packnett, MPH2, Edo Banach, JD3, Ryan Ross, MS4, Avina Rami, MD5, Oreofe Odejide, MD, MPH6, Gregory A. Abel, MD, MPH7
AFFILIATIONS
1 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
2 Merative Corporation, Ann Arbor, Michigan, United States
3 Foaly Hoag, Washington D.C., District of Columbia, United States
4 Merative Corporation, Ann Arbor, Michigan, United States
5 Brigham and Women's Hospital, Boston, Massachusetts, United States
6 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
7 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
ABSTRACT
INTRODUCTION: Patients with hematologic malignancies (HM) are at high risk of intensive healthcare utilization and decreased supportive care at the end of life (EOL; Earle, JCO 2008). While there are robust data for older adults (≥65 years) and pediatric patients (<18 years), there is a need to evaluate and improve EOL care for the middle population of adults with blood cancers aged 18-64 years, who, given the prevalence of private insurance, are often left out of administrative EOL quality care analyses. Moreover, a composite measure of “ideal” EOL care is needed, as many patients experience some aspects of poor EOL care (e.g., repeat inpatient admission or ICU care in the last month of life) alongside better ones (e.g., dying in hospice).
METHODS: We conducted a retrospective cohort analysis of adults aged 18-64 years who died from blood cancers and examined receipt of an “ideal” EOL care endpoint consisting of key healthcare utilization metrics and hospice enrollment. We utilized data from the Merative Marketscan Multi-State Medicaid and Commercial Databases, which contain de-identified demographic and longitudinal patient-level claims for approximately 20 million Medicaid and 62 million commercially insured patients. We linked eligible patients from the Commercial database with the Centers for Disease Control and Prevention’s National Death Index to obtain survival; survival status in the Medicaid database was determined via claims and enrollment records. The study cohort included adults aged 18-64 years who had an HM diagnosis (lymphoma, myeloma, leukemia, myelodysplastic syndrome [MDS], or myeloproliferative neoplasm [MPN]), survived at least 30 days following their diagnosis, and died between January 1, 2016 and December 31, 2021. We defined “ideal” EOL care to include hospice enrollment and avoidance of three high-intensity healthcare utilization outcomes in the last month of life: 1) ≥2 inpatient admissions, 2) ≥1 ICU admission, and 3) in-hospital death. This endpoint was tailored to the stated preferences and clinical needs of HM patients and their caregivers (Wright, JAMA 2008 and JAMA 2016 ). Univariable analyses examined the association between the receipt of “ideal” EOL care and the clinical and demographic covariates of interest using Chi-square and t-tests/ANOVA analysis for categorical and continuous variables; multivariable analyses were conducted with specific covariates selected through backward selection using Bayesian information criterion (BIC).
RESULTS: Among 2,669 patients, most were male (60.7%), and the median age at death was 57 years with 10.9% of the cohort < 40 years. The most common diagnosis was lymphoma (43.5%). Approximately 13.6% of the population underwent a hematopoietic stem cell transplant (HCT) in the 12 months prior to death and 17.0% received ≥2 transfusions in the last month of life. Most patients were commercially insured (68.4%) and lived in urban areas (82.8%).
Overall, most patients (86.0%) did not receive ideal EOL care as we defined it, with 74.0% of patients not enrolling in hospice, 50.5% and 16.4% experiencing at least 1 ICU and ≥2 inpatient admissions in the last 30 days of life respectively, and 48.7% dying in the hospital. In multivariable analysis, receipt of ≥2 transfusions in the last month of life was associated with decreased odds of receiving ideal EOL care (Odds Ratio 0.29, 95% Confidence Interval 0.15-0.53). Other covariates (sex, age, insurance type, comorbidity, prior HCT) were not associated with ideal EOL care.
CONCLUSION: In this cohort of adult decedents aged 18 to 64 years with blood cancers, most did not receive ideal EOL care as we defined it, with transfusion-dependent patients particularly at risk. This highlights the need for effective care models to support the EOL experience for patients navigating advanced illness alongside the unique psychosocial demands of dying at neither age extreme such as post-secondary education, raising families, establishing and growing careers, and caring for elders.
DISCLOSURES
Hari S. Raman, MD, MBA: No relevant conflicts of interest to declare.