Public and private payers, policy advisors, and oncologists and nurses from the audience debated how to measure quality in cancer care during a panel discussion at the 2015 Association for Value-Based Cancer Care conference. John Fox, MD, MHA, Senior Medical Director and Associate Vice President of Medical Affairs, Priority Health moderated the discussion.
The panel members included Heidi Schumacher, MD, of the Patient Care Models Group, Center for Medicare & Medicaid Innovation; Michael A. Kolodziej, MD, National Medical Director, Oncology Solutions, Aetna; Bruce S. Pyenson, FSA, MAAA, Principal and Consulting Actuary, Milliman; and Gary M. Owens, MD, President, Gary Owens Associates, and Co-Chair of the conference.
In response to the question, “How do we measure quality using claims data?” Mr Pyenson said, “What is the easiest thing for a payer to measure? Money. It is audited, claims are adjudicated. If you are going to measure something, you start with what you can measure. You can measure money, and that is certainly an outcome.” He added that another basic measure was mortality. There can be arguments about what is progression-free survival, but usually you know if someone is dead or alive, as a starting measure for quality.
The panel elaborated on those 2 measures, with some concern expressed about the difficulty of tracking a patient’s death (times and days). Some practices monitor obituaries so as not to bother a patient’s family with details for tracking, and even payers find it a challenge to monitor death, for database management purposes.
Other quality measures cited include the amount of time spent in a hospice, the number of hospital admissions in the last 2 weeks of life, and chemotherapy use in the last 2 weeks of life, but even these measures were debated as to whether they are merely surrogate metrics to track in the absence of more detailed outcomes measures. To apply any metrics in oncology to track quality means that, to some degree, imperfect data sets will push analytics to probabilities.
With all the activity related to the development of health information exchanges, Mr Pyenson noted that probably 90% of what they are trying to track could be obtained from claims data systems at almost 5% of the cost.
An oncologist from the audience asked how practices could possibly be compared in any statistically significant manner because of the low numbers of patients with any given disease treated in an oncology practice. His 4-physician practice may see only 10 patients with metastatic colon cancer annually, a number that would yield no statistical significance for any analysis. Mr Pyenson amused the audience by responding, “you have to validate it with caution,” and then went on to note that with aggregation of large numbers of patients across many physicians, and with application of probabilities, we can get to some estimates, but even those will need some degree of confidence intervals for utility.
Dr Kolodziej commented that the discussion was really centered around 2 types of data metrics—measuring the clinical outcome and measuring the cost of care. Even small practices can yield useful cost-of-care data, but clinical outcomes and survival data are much more difficult to achieve. He cited a recent study showing that irrespective of what type of cancer patients had, 75% of the total cost of care was fairly tightly grouped, and approximately 25% of the patients had very expensive treatments. This meant that the challenge in dealing with variability in cost of cancer care is those high-cost outliers.
Even survival is a difficult clinical outcome to measure, he said, because there are so many factors that can affect survival but have nothing to do with the skill of the physicians. Dr Kolodziej’s caution was that we can measure outcomes for quality metrics in cancer care, but we also have to recognize their limitations, especially because we often accept process measures as surrogates for the desired outcome.
According to Dr Kolodziej, Aetna is building oncology medical home models with oncologists, and uses surrogate measures, such as whether a pain assessment was done, a survivorship plan was developed, or an informed consent was documented. At some point, the medical community will need to track not only whether these items were done, but also what was the patient response, the patient engagement, and the clinical differences that resulted from these actions. It may take a long time for us to get to those quality metrics, he said, because the data are in multiple silos and have to be aggregated and normalized.
We are now functioning in a unidimensional—or, at most, bidimensional—world, when we need to look at quality from a 3-dimensional direction. Some examples noted were whether giving the right antiemetic therapy mattered to patient experience; how many phone calls a patient made to a practice within 72 hours of getting chemotherapy, and why; how often rescue medications are being given, and when; and how frequently that same patient may be admitted to the hospital, and why.
The panel agreed that as we move into the next phase of measuring quality in oncology, patient-reported outcomes will become increasingly important. The Centers for Medicare & Medicaid Services has tied patient satisfaction scores to their quality measures.
Some payers are now funding patient care management to manage the patient experience and navigation outside of the clinical treatment, and providing feedback (at times) to the treating clinicians for quality improvement of the whole cancer treatment process. Even patient-reported outcomes and patient satisfaction may mask true quality of care; patients may be very satisfied, but not necessarily because they are receiving good or proper patient care.
An oncologist in the audience noted that there is great difficulty using checkboxes to track care quality. Giving a patient a survivorship plan does not mean that the patient will actually use it, share it with other care providers, or whether their other primary care physicians found it useful. Her point was that we need to make sure that the measures (surrogates or not) that cost practices resources to expend their compliance, and that payers use resources to review, are actually worth the expenditure of those limited resources.
There needs to be better literacy among cancer professionals about what meaningful metrics really are. It may be possible to improve that dialogue with more technological advances, but we have not yet reached that point.
The panel expressed concern that movement toward episode-based reimbursement could result in underutilization of appropriate cancer care, and that some of the metrics being developed are to ensure that appropriate care is provided when needed in a system that could reward giving less treatment—an area in which clinical pathways could be useful.
There are no easy solutions. Even when the panel was challenged to consider whether risk models or pathways programs could replace prior authorization programs, there was no agreement that prior authorization programs would go away any time soon. Ultimately, the panel expressed the view that the growth of attribution of patients in integrated delivery systems and accountable care organizations will continue to press the question of value and quality metrics in cancer care.
Looking at quality in cancer care will require a team transformation in the care of patients. To be able to address the full continuum of care from cancer screening through treatment and long-term survivorship will be the key to identifying quality and value in cancer care. That will require a team approach between patients, payers, and all the providers who touch the patient.