How can any analysis of the quality of cancer care being delivered by providers that focuses on a survival outcome as the ultimate measure of that quality-rather than on an evaluation of the optimization of the care process-be considered objectively valid and clinically meaningful?
Maurie Markman, MD
During a provocative address at the 2015 ASCO Annual Meeting, the well-known and highly respected health economist Professor Michael E. Porter, PhD, challenged ASCO to move beyond its current focus on optimizing the process of cancer care in its quality efforts and develop systems that hold oncologists and their programs accountable for the outcomes of that care.1
In arguing for a value-based system, Porter exhorted the oncology community to move away from compliance with clinical practice guidelines as a means of measuring quality in cancer care and instead define quality by analyzing whether the patient achieves outcomes such as survival and quality-of-life measures.
Clearly, the initial or primary outcome of interest to the large majority of patients and their families is survival, with quality of life and other metrics likely being considered quite important but often secondary.
However, the intent of this commentary is not to debate this belief—even while recognizing this question itself is worthy of discussion—but rather to vigorously challenge the basic assertion suggested by some in the oncology community that the more appropriate measure of quality should be in demonstrating a favorable survival impact as the ultimate and most objectively valid measure of the outcome of care, as opposed to merely optimizing the process of that care.
What Optimizing the Process of Care Means
Indeed, the underlying question is whether survival outcomes can be considered the ultimate measure of the quality of cancer care when so many factors influencing survival are simply beyond the control of oncology care providers.First, it is relevant to define what the phrase optimizing the process of cancer care means. This expression conveys the message that the clinical team, led by a physician, has successfully undertaken and appropriately applied all measures such as diagnostic tests, treatments, and follow-ups that the best currently available evidence suggests would provide an individual patient in a specific clinical setting the greatest opportunity to achieve a favorable clinical outcome—that is, progression- free survival or overall survival.
Importantly, this definition fully recognizes that any or all of these appropriately applied process activities simply may not be able to successfully influence to a major or even minor extent the natural history of a given patient’s malignancy. In the popular parlance, such a natural history, characterized by a progressive or recurrent cancer following currently defined optimal therapy, is often labeled “bad biology.”
There is no intent here to suggest our existing anticancer strategies do not influence the natural history of cancer, but rather that the goals of our quality initiative efforts must be focused on what our clinical teams can directly influence, rather than on factors that are today simply beyond our control.
Nothing stated above should come as a surprise to any oncologist. For example, the entire cancer staging system recognizes that all “survival figures” only represent the statistical probability for a particular survival outcome.
When one declares that a patient with a localized colon cancer has a 90% chance of 5-year survival, this clearly identifies the objective fact that 10% of patients in this otherwise objectively favorable setting will not be alive 5 years after the diagnosis.2
But the question to be asked here is whether the failure of a patient with localized colon cancer to survive for 5 years in any way indicates or even suggests the delivery of an inferior quality of care, compared with a similar patient who survived 5 years.
Of course, this question must include the proviso that the management provided included all elements of the process of care at the time of delivery believed (or felt) to be necessary to optimize the opportunity for the most favorable survival outcome— for example, diagnostic testing, surgical evaluation, and administration of standard-of-care adjunctive and second-line strategies.
How Provider Activity May Impact Survival
Similarly, if one were discussing a considerably less favorable clinical setting such as localized esophageal cancer, which has a 5-year survival rate of 40%, would the failure of a patient to survive 5 years indicate or suggest inferior care compared with the unfortunately statistically less likely situation that the individual would be alive 5 years later?2In the example of localized esophageal cancer, provider activity might have an unequivocal impact on a survival outcome independent of whether the appropriate process was followed. In this situation, we must consider the skill levels and any incidence of medical error by the individual provider or particular organization.
For example, it is well recognized that surgical skill as well as institutional experience and essential support can quite meaningfully influence short-term mortality, including the performance of curative surgeries, for esophageal and other cancers.3
In this situation, because the survival outcome being measured will be quite early, and specifically before the potentially overwhelming impact of the biology of the cancer becomes a major component defining survival, it is possible to distinguish the role of the provider/ institution in the event of a negative outcome.
Many Other Factors Affect Outcomes
The accidental administration of the wrong dose of chemotherapy or other treatment-related errors also has the potential to result in serious harm including death or to prevent the subsequent timely delivery of optimal care. However, in such settings we are still discussing the process of care that unfortunately has been applied inappropriately, or potentially even tragically, possibly resulting in an inferior survival outcome.In this discussion, it is important to fully appreciate and objectively acknowledge the serious limitations of our understanding of the multiple factors that influence a survival outcome in the oncology arena, and in our ability to either directly influence those factors or even correct for their impact, in any examination of survival as a measure of the delivered quality of care.In fact, it was more than 60 years ago that David Karnofsky reported the major relevance of a patient’s baseline performance status (Karnofsky Performance Status Scale) in defining an individual cancer patient’s survival.4
Multiple other groups have confirmed the importance of a patient’s baseline performance status in defining survival in a number of tumor types, including cancers of the lung,5,6 pancreas,7 colon and lung cancer (combined),8 urothelial malignancies,9 and hepatocellular carcinoma.10
How can this observational, but largely subjective evaluation, as important as it is in the design and analysis of clinical trials, be used to correct for serious imbalances in patients seen in a given clinical practice, assuming that one desired to use a survival outcome as a measure of the quality of care delivered?In a most provocative report in a head and neck cancer patient population, it has even been shown that the distance an individual patient lives from the treatment center may measurably impact a survival outcome, possibly due to differences in the socioeconomic status of those living close to an academic medical center (eg, inner city) versus individuals having sufficient financial resources or family support to permit them to travel longer distances to receive their care.11 How would this factor be handled in an evaluation of the quality of provider care?And finally, we come to the major issue—perhaps a true giant in the room—of the clinically relevant comorbidities that are extremely common in the cancer patient population, due to the fact cancer is a condition seen with increasing frequency as one ages or where there are well-recognized disease associations such as cigarette smoking as a cause or predisposing factor in lung cancer, heart, or pulmonary disease.12-14
CCI indicates Charlson Comorbidity Index; CML, chronic myeloid leukemia; OS, overall survival.
The ultimate impact of comorbidities on the ability of oncologists to deliver optimal cancer treatment such as the risk of serious adverse events (eg, surgical and chemotherapy complications) or as an independent cause of death (eg, myocardial infarction or stroke) may be substantial, as evidenced by a recent chronic myeloid leukemia study (Infographic).14In summary, the critical question to be asked here is: How can any analysis of the quality of cancer care being delivered by providers that focuses on a survival outcome as the ultimate measure of that quality— rather than on an evaluation of the optimization of the care process—be considered objectively valid and clinically meaningful?
Maurie Markman, MD, OncologyLive’s editorin- chief , is president of Medicine & Science at Cancer Treatment Centers of America, and clinical professor of Medicine, Drexel University College of Medicine. firstname.lastname@example.org.