Shelby Reed, PhD, RPh

Cancer patients participating in focus groups at the Duke Cancer Institute, conducted to elucidate factors associated with value in cancer care, consistently expressed ‘hope’ as an important feature, often described as living longer and doing things they enjoy. The value of hope concept is represented conceptually in both ASCO’s value framework in its ‘tail of the curve’ domain and as a novel element in ISPOR’s US value framework. To move from conceptualization to application, it is necessary to quantify the relative importance of hope in comparison to other aspects of value. Traditional utility theory defines value as how much people would give up of one desirable object to obtain another desirable object. Thus, to quantify the value of hope, it is necessary to determine what patients are willing to give up to obtain it. Traditional cost-effectiveness analysis multiplies probabilities and health-outcomes to measure the expected effectiveness of treatments. The expected-utility assumption posits that preferences are linear in probabilities and outcomes; thus, small probabilities of large health improvements will be valued the same as large probabilities of small improvements, given the same expected outcome. While this approach facilitates comparisons of alternative treatments, it precludes the possibility of finding differential value between the alternatives based on the certainty of their outcomes. Previous valuation studies have estimated willingness to pay or willingness to reduce expected survival for uncertain but significantly longer survival, and defined this value as the value of hope. One recent U.S. study elicited willingness-to-pay values using a single open-ended question and likely invoked cognitive biases due to poor risk communication. To obtain quantitative measures of the value of hope in cancer care, our project team, including experts in health economics, stated-preference research, decision psychology, oncology, and health policy will design a best-practice discrete-choice experiment (DCE) using novel graphical displays to more clearly depict survival periods and probabilities. We will design survey modules representing different stages of cancer, with expected survival of 18 and 48 months. Survey participants first will choose among treatment options or indicate their indifference to therapies that represent equal amounts of expected survival, but differ with regard to probabilities of shorter and longer survival to test the expected-utility assumption. Then, we will administer a DCE in which survival outcomes will be shown along with two opportunity ‘costs’, out-of-pocket expense and health status. The addition of these two costs will allow us to quantify the value of hope relative to changes in respondents’ income, while moderating survival gains with the health-related quality of life they would experience. The survey will be administered to 150 adults with a history of cancer and 150 adults without a history of cancer.