Evaluating Quality Metrics for the Care of Patients With Blood Cancer Who Are Near Death.

Abstract

Striking deficiencies in end-of-life (EOL) cancer care have been well documented. However, the vast majority of these studies have focused on overly aggressive care in patients with solid tumor cancers. More recently, research has suggested that patients with blood cancer receive higher rates of aggressive care at the EOL (eg, hospitalization near death) than do those with solid tumors, which has raised concerns over the treatment of patients with hematologic malignancies. Therefore, the question of what constitutes high-quality EOL care for patients with blood cancer is of substantial clinical importance. Research on EOL care in solid tumor cancers has greatly advanced our understanding of the medical and psychosocial issues encountered at the EOL. However, the generalizability of these findings to blood cancer is unclear. Patients with blood cancer may face unique adverse events that influence EOL care. For example, patients with leukemia with low platelet counts may be at risk of massive hemorrhages at EOL for which transfusions may significantly improve quality of life. Situations such as this may create a different medical and psychosocial context for EOL care than that experienced by patients with solid tumors. The article by Odejide et al that accompanies this editorial represents an important initial step toward addressing the question: What is quality EOL care for patients with blood cancer? Among other constructs, the authors assessed hematologic oncologists’ views of the acceptability of various quality measures of EOL care for patients with blood cancer. Data were collected using a survey mailed to hematologic oncologists listed in the clinical directory of the American Society of Hematology. The quality measures of EOL care included standard measures developed for solid tumor cancers and two new hematology-based measures. Contrary to the authors’ hypothesis, the standard EOL quality measures were rated as acceptable by the largest proportion of oncologists. Specifically, hospice enrollment for . 7 days (77.9%), no receipt of chemotherapy within 14 days of death (79.9%), and no cardiopulmonary resuscitation (85.1%) or intubation in the last 30 days of life (80.5%) were rated as acceptable by the highest proportions of the sample. The hematology-based indicators of no red blood cell transfusions (58.7%) or platelet transfusions (59.9%) in the last week of life were rated as acceptable by a smaller proportion of oncologists. These findings address important questions about the definition of quality EOL care in patients with blood cancer. However, this study also generates additional questions that need to be answered for the validation of quality care metrics, such as: Are the indicators of quality EOL care deemed acceptable by clinical experts empirically valid? More specifically: What is the definition of high-quality care at the EOL for dying patients with cancer and how might it differ for patients with blood cancer? The work of Odejide et al and others has focused on oncologists’ views of quality care. This is a logical starting point, given the specialized training and expert knowledge of oncologists regarding the nature and treatment of blood cancer. Nevertheless, it remains to be seen whether the opinions of clinical experts yield valid indicators of quality care. The study by Odejide et al represents an important first approach to the understudied topic of EOL care for patients with blood cancer. However, it also reveals the limitations of relying on expert opinion in the development of quality-of-care indicators. In this study, a large majority of the sample (77% to 85% across quality indicators) rated four standard EOL quality measures (developed in research with patients with solid tumors) as acceptable. In a prior study by these same authors, these measures were rated as acceptable by, 50% of the sample. This discrepancy suggests a lack of consensus among oncologists on the definition of quality EOL care in hematologic malignancies. The authors offer plausible explanations for the discrepant findings, including the possibility that differences in treatment setting may influence oncologists’ perception of what constitutes quality EOL care. However, the lack of consensus across oncologists and the potential influence of factors such as treatment setting suggest the unreliability of expert opinion. Furthermore, expert opinion offers insufficient evidence of empirical validation, which would be necessary to demonstrate the psychometric soundness of the proposed quality metrics. The Centers for Medicare & Medicaid Services (CMS) have formulated a potentially useful framework for the rigorous evaluation of quality-care metrics. Odejide et al note that the quality measures they propose are not intended for performance assessment or payment purposes. However, payment reform is not the only purpose of CMS quality measures. They are also designed to guide effective treatment by assessing health care processes and outcomes for their alignment with the CMS goals of “effective, safe, efficient, patient-centered, equitable, and timely care.” The Blueprint for the CMS Measures Management System describes a detailed multiphase process for developing measures to