Dr. Reed is a Professor in Population Health Sciences. She has also been affiliated with the Duke Clinical Research Institute since 2000, where she currently leads the Preference Evaluation Research Group and is the Population Health Therapeutic Area Leader. Dr. Reed’s research spans health services and health preference research and health economics.
Which of your cost-effective analyses has had the greatest economic impact?
I think it would be more meaningful to discuss a study that had the greatest net impact rather than the greatest economic impact since net impact studies account for the effect on survival and health-related quality of life as much as they do cost. An example is a study I worked on that assessed the cost-effectiveness of Gleevec (imatinib)—the first tyrosine kinase inhibitor for chronic myeloid leukemia and a real breakthrough in cancer treatment. In 2001, the drug was thought to be extraordinarily expensive at $26,000 per year. But, our research found that from a traditional value perspective, the projected impact of the drug on the patient’s survival and quality of life was so great we could support Novartis' initial pricing. Today, that same drug costs over $120,000 annually, and unlike in most situations, the generic versions are only a little less expensive.
Is it true that early interventions lead to cost-effective healthcare?
Generally, people believe that preventive care must be cost-effective, but in reality, sometimes it is and sometimes it isn't. It comes down to how many people must be screened or treated for a particular disease, how much the screenings or early interventions cost, and whether it's possible to prevent or simply delay disease onset. We also have to consider the cost and effectiveness of treatments to manage a condition. Recently, I collaborated with colleagues at Oxford University to determine whether delaying the onset of type 2 diabetes can be cost-effective. We estimated that interventions costing from about $600 to almost $3,000 per year can be cost-effective if they delay the onset of diabetes from one to nine years, respectively. Studies like this can provide information to help drug developers design interventions that deliver good value.
A similar situation may be playing out as we wait for a COVID-19 vaccine. But here we have to ask: How effective will the vaccine be? Will it provide continued immunity? Who will receive it? And how much will it cost? There's no question a vaccine will provide tremendous value to society. Economists believe we should consider the societal and economic benefits beyond the vaccine's immediate health effects—and many others agree. But again, there’s an important question to consider when pricing the vaccine, and that’s how to split the value it will generate between the drug developer and society. Its price should be sufficient to provide a significant financial reward to the company, but not so large that it’s unaffordable, drawing massive resources away from other highly valuable health and non-health services.
You've spoken on personalized medicine, with this in mind, what do you think will be its greatest financial benefit to the healthcare systems and personal benefit to patients?
There are benefits to knowing an individual's genetic profile; however, testing everyone to find a rare mutation can be costly. Here’s a good example using cancer care—we can test many patients to identify a small percentage with a given mutation, then we can give these cancer patients with the mutation a drug that may be cost-effective. To calculate the strategy’s true value we must account for both screening and treatment, because together we may find it’s not cost-effective at all. Perhaps we look at multigene panels that direct treatments for several mutations as a more cost-effective approach, but this too is a package deal because genetic testing isn’t cost-effective unless the results lead to appropriate treatments that are a good use of health care resources.
To answer the second part of your question, from a humanistic perspective, if we're to fully account for the value of personalized medicine, we have to ensure treatment decisions are aligned with a patient's preferences. Clinicians can then discern if patients are more concerned with maximizing their quality of life or increasing their duration of survival. To illustrate, I recently published a study with Dr. Laura Havrilesky where we examined the preferences of women who took a drug for maintenance therapy after treatment of an ovarian cancer recurrence. We found women, on average, were willing to accept less time with progression-free survival to avoid the drug's side effects of nausea and fatigue. Essentially, they were trading off improvements in quality of life for reductions in expected progression-free survival. The tipping point for accepting these treatment side effects was at least a three- to four-month gain in progression-free survival. Absent that, the women felt they would be better off forgoing treatment.
Although studies like this can be helpful, we recognize that preferences can vary dramatically between two individuals. My colleagues and I, at the Preference Evaluation Research (PrefER) Group, are designing approaches that elicit a patient's preferences so clinicians will have a profile to help guide shared decision making. Like in the study above, we may find that when patients' preferences are considered in treatment decisions, they receive less aggressive regimens—leading to lower healthcare costs. And from a societal perspective, that's a win-win proposition.