Tonight the Healthcare Leadership (#HCLDR) Blog community will discuss the American Society of Clinical Oncology’s proposed Conceptual Framework on Value of Cancer Care. Preparing to participate in HCLDR tweet chats is great way to review areas of interested I’ve not thought about in while. I mostly think about healthcare IT workflow. So HCLDR usefully takes me outside my mental box, so to speak, though I did find some interesting healthcare and IT workflow wrinkles!
Value to patients and society of cancer treatments depend on preferences of various stakeholders. And closely tied to value and preference is the economic concept of utility. So I also recommend this white paper, What Are Health Utilities?
But first I read A Conceptual Framework to Assess the Value of Cancer Treatment Options.
While the title includes “Value”, the paper is motivated by the high cost of cancer treatment. Since I have an Accountancy undergraduate degree, I’m interested in the variety of ways healthcare measures (or fails to measure cost) and implications clinical and health policy decision-making. If cancer treatments cost a lot, then we need to know which treatments are most valuable to patients and society.
Coincidently, several weeks ago the HCLDR topic was trade-offs in healthcare. I did my research and wrote Healthcare Trade-offs, Shared Decision Making, Vulcan Mind-melds, and a Marriage Metaphor.
“A Conceptual Framework to Assess the Value of Cancer Treatment Options” is also about trade-offs. That’s why my title for this blog post starts of “Trade-offs, Preferences, Utilities, Part Two: ….” The very interesting Conceptual Framework paper is about value, and value cannot be understood without understanding patient preferences. Different cancer patients have different preferences about such things as quality of life versus length of life, and perhaps even financial consequences of treatment, “financial toxicity.”
In my previous post I explained how expected utility decision theory is the gold standard for combining probabilities (of whether a treatment will work or not) and patient utilities (preferences for possible consequences of treatment, pain, longevity, etc.). I also noted problems with formal methods such as expected utility approaches to shared medical decision making.
“Conceptual Framework…” is a high-level document. However, it mentions a variety of formal cost-benefit models in Table 1 summarizing how other countries measure costs and benefits of treatments. So I decided to drill down into the cancer patient preference topic from the point of view of seeing what has become the state of the art since I studied this subject in graduate and medical school. The best three papers (that I could find and read and for which full PDF text is available for free on the Web) are the following:
- Preferences For Cancer Treatments: An Overview Of Methods and Applications in Oncology
- Patient Preferences in Choosing Chemotherapy Regimens in Advanced Non-Small Cell Lung Cancer
- Deriving A Preference-Based Utility Measure For Cancer Patients From The European Organization For The Research and Treatment Of Cancer’s Quality Of Life Questionnaire C30
In addition to What Are Health Utilities, I also highly recommend Preferences For Cancer Treatments: An Overview Of Methods and Applications in Oncology. Published in reasonably recent 2012, Blinman et. al. include discussion of expected utility theory, similar to what I covered in my previous posts on trade-offs. A particular application of expected utility theory, the Standard Gamble, is “often considered to be the gold standard of [patient] preference assessment. They also cover Time Trade-Off, Discrete Choice Experiments, and Multi-Attribute Utility Instrument. As I noted in my previous blog post, I was sure there were many interesting developments in estimating patient preferences and utilities since when I studied the subject decades ago. And this paper is a great way to catch up on this topic.
The second paper, Patient Preferences in Choosing Chemotherapy Regimens for Advanced Non-Small Cell Lung Cancer, is a patient research survey study. The authors conclude “our study findings indicated that lung cancer patients have significant and varying concerns about the side-effects of chemotherapy and that these concerns are not being uniformly identified by physicians or integrated into decisions about treatment plans…. Further prospective studies of this issue should include exploration of both patient concerns and preferences and physician perceptions about tailoring chemotherapy regimens according to the adverse reactions they may cause.”
Deriving A Preference-Based Utility Measure For Cancer Patients… is a technical paper. It describes a method to derive a Health State Classification System, a collection of core domains of health-related quality of life (eg. fatigue, pain, nausea) and associated levels (eg. poor, moderate, good) and then a scoring algorithm of assigning a utility value to each possible health state. If you recall, from my previous post of healthcare trade-offs, I remarked how difficult it is to estimate the utilities to be plugged into expected utility models of shared decision making. This research addresses that problem.
Finally, since I’m a health IT guy, and always, always, looking for the workflow and usability angle, I’ll quote the following from the ASCO Conceptual Framework paper:
“The complexity of the value framework makes it clear that for it to eventually be used effectively in a practice setting, the information must be presented in a visually appealing, user-friendly way and acquired almost immediately. Thus, our vision entails preloading data for all regimens to be evaluated, and that of their comparators, into user-friendly software that can be used on a smart phone, tablet, or computer and integrated into the electronic medical record. The tool that is envisioned will include the key elements discussed here for clinical benefit and toxicity for the majority of commonly used cancer regimens in a variety of clinical scenarios and will permit incorporation of patient weighting preferences. For example, if, in the advanced disease setting, longevity is less important to a patient than freedom from toxicity, the tool should be able to adjust the clinical benefit and toxicity parameters to reduce the impact of clinical benefit and enhance the impact of toxicity, thereby producing a personalized NHB. The ability to modify the framework at the point of care would facilitate decision making by enabling patients to create a personalized NHB score that takes into account not only the specific clinical problem but also existing comorbidities, personal preferences, and values. In addition, access to the cost of the regimen in question and the patient’s out-of-pocket costs will provide additional context to the physician and patient in determining the relative value of treatment options.”
This topic fascinates me. It touches on some specific interests of mine. There’s an interesting connection between representing clinical protocols and clinical workflows. And clinical protocols, especially those with decision branch points, resemble decision trees such as I discussed in Part One of this series on trade-offs. So, naturally, I wonder if there might be a role for workflow technology at the point-of-care, to effect the kind of shared clinical decision-making decision support envisioned here…
That’s it! As usual, I learned lots preparing for a #HCLDR tweetchat. I look forward to it tonight, and every Tuesday night at 8:30EST. I hope I’ll see you there!