Patients with inflammatory bowel diseases (IBD) and their doctors often struggle to decide on a treatment plan. There is no cure, only relief of sometimes debilitating symptoms, and both of the leading drug therapies carry long-term risks. Unchecked, severe disease can require surgery to remove inflamed areas of the bowel and, in rare circumstances, can cause fatal complications.
IBD affects nearly 1.5 million Americans. In this set of conditions, which include Crohn’s disease and ulcerative colitis, the body’s immune system attacks the intestines. The result: intermittent symptoms, including abdominal pain and diarrhea, that reduce productivity and quality of life.
One type of treatment, anti-TNF (tumor necrosis factor) drugs taken continuously suppresses the immune system, sparing the intestines. But over the long term, this treatment raises the risk of serious infections and may slightly increase the cancer risk.
Another group of drugs, corticosteroids, are used as needed to suppress flare-ups of the disease. Because they have been around longer, have other uses, and are not taken continuously, they strike some patients as a safer option. But they may not control IBD as well as anti-TNF drugs do, and they also increase the risk of infection and other adverse events.
“IBD in general is a scary set of illnesses,” says James Tarver, PhD, MBA, a chemist at the University of Pennsylvania and a patient partner in a PCORI-funded study comparing the two drug treatments for IBD. Tarver, who is 40, has had Crohn’s disease since he was 17. “A lot of the treatment decisions end up being really emotional,” he says. “You’re afraid of making the wrong decision; you’re afraid of feeling worse.”
Unfortunately, there has been little evidence to guide treatment choices. No previous clinical trials have directly compared the effectiveness of corticosteroids to that of anti-TNF drugs, says James Lewis, MD, MSCE, the study’s principal investigator. Randomized controlled trials of anti-TNF drugs have typically compared them to placebos, not to corticosteroids, he says.
A head-to-head randomized controlled trial would be too expensive and take too long, Lewis says. So he and his colleagues decided to carry out an observational study comparing the two treatments by looking at the records of patients who had been on one type of the medications or the other.
In addition, the researchers wanted to determine what outcomes—both good and bad—matter most to patients. Finally, they plan to figure out how to determine which treatment is best when patients’ priorities are taken into account. “Ignoring patients’ preferences could lead to incorrect conclusions since physicians and patients don’t always share preferences for treatments and outcomes, particularly in IBD,” Lewis says.
Scouring Medical Records
In one of the study’s three phases, the researchers combed through Medicare medical records from 2006 to 2013 and Medicaid records from 2001 to 2006 that, taken together, represent about 100,000 IBD patients. The team looked for new users of anti-TNF therapies and prolonged users of corticosteroids to treat IBD. Taking into account potential links between patient characteristics and the treatment received, the researchers compared the incidence of negative outcomes—such as severe infection, surgery removing a section of the bowels, and death. They also looked at cancer incidence, including lymphoma. Lewis’s team is currently analyzing their results.
Today, given the complexity of choices and worrisome potential side effects, it is increasingly important to base treatment decisions on systematically analyzed data rather than individual anecdotes, says Meenakshi Bewtra, MD, PhD, of the University of Pennsylvania, a co-investigator as well as an IBD patient. Findings drawn from thousands of patient records could serve as a reliable guide and influence doctors’ behavior, she says.
Ignoring patients’ preferences could lead to incorrect conclusions since physicians and patients don’t always share preferences for treatments and outcomes, particularly in IBD.
James Lewis, MD, MSCE University of Pennsylvania
Teasing Out Values
In another part of the study, the researchers set out to learn which treatment outcomes patients value most, and how much risk they will accept to achieve those outcomes. They used a novel statistical procedure called discrete choice analysis. This approach is common in marketing and economics, but little used, so far, in medicine.
Medical studies have usually relied on other methods to gauge patients’ preferences regarding treatment choices, Bewtra says. For example, some studies ask patients how many years they would be willing to subtract from their lives in order to live out their remaining years in a state of perfect health. “That’s not how we offer options to patients in the office,” Bewtra says.
In contrast, the PCORI-funded study described treatments as having specific risks and benefits, each with a certain likelihood. By responding to multiple hypothetical scenarios, participants indicate what trade-offs in risks and benefits they are willing to make.
“You can find out what aspect of a therapy was most important to driving its success or failure,” Bewtra says. “Was it the efficacy? Was it the infection risk? Was it the cancer risk?”
In collaboration with Tarver and the other patient partners, the team designed a 10-question online survey. It asks patients to imagine they were experiencing a flare-up of their disease and had to choose between two unnamed treatments. In one question after another, the survey plugs in different numbers and percentages. In one question, for example, the first choice might offer an 80 percent reduction in symptoms along with a 30 percent chance—within 10 years—of an infection so serious that the patient would land in the hospital. The other choice might offer only a 60 percent reduction in symptoms but with only a 10 percent chance of the severe infection.
The patient partners and other patients helped create the survey during in-depth, one-on-one interviews that resulted in many iterations. “These are really hard surveys to put together,” Lewis says. “You have to present probabilities in a way that people will understand them and, without overwhelming them, provide enough information about the different things that could happen.”
“We went over the wording, the color choice, the format,” says Bewtra, who conducted many of the patient interviews. “We challenged their answers to questions to make sure we were getting the highest or lowest amount of risk they would accept.”
The researchers then administered the survey to over 1,200 patients with IBD who are part of CCFA Partners, a patient-centered online research network organized by the Crohn’s and Colitis Foundation of America. As part of PCORnet, the National Patient-Centered Clinical Research Network, CCFA Partners receives PCORI funding.
The researchers are now analyzing their data. They are finding that patients have different preferences as they balance benefits of treatment and risks of infection, surgery, and cancer. That’s never been shown before in studies of IBD patients, Bewtra says. She adds, “There’s been this assumption that the average represents everyone. We found that no, that’s not true.”
Balancing Effectiveness and Preferences
The final part of the study, which will conclude later this year, synthesizes the results of the other two. The researchers will use the patient preferences they have documented to tailor the results from the study of treatment outcomes. To do this, they will apply recently developed methods to assess a treatment’s clinical benefits—weighted by patient preferences—and its risks. They will use two approaches to compute the net benefit of anti-TNF therapy relative to corticosteroid treatment.
This combination of approaches, they say, will lead to the clearest picture yet of which treatment choices best help patients achieve the outcomes that matter most to them. Lewis says, “The results will be highly informative for patients and caregivers who face this exact treatment decision every day.”
IBD in general is a scary set of illnesses. A lot of the treatment decisions end up being really emotional. You’re afraid of making the wrong decision; you’re afraid of feeling worse.
James Tarver, PhD, MBA Chemist, University of Pennsylvania and PCORI-funded study patient partner
Respecting Preferences
Recognizing differences in patient preferences could lead to a tool, perhaps similar to the survey, to facilitate conversations between doctors and patients about treatments. It could make a clinical appointment more like visiting a financial planner, Lewis says. “A financial planner tries to learn what your goals are, what your values are, how much tolerance for risk you have. If doctors had that sort of information, would it influence the conversation?” he asks.
Tarver says that helping create the survey led him to see his own values more clearly, and to understand how they’ve guided his care, for better and worse. Having such a survey would help facilitate discussions between patients and physician and perhaps lead to better treatment decisions.
For example, he says, “I was always reluctant to get on anti-TNF medication. I was afraid of increasing my risk of lymphoma and other side effects that turned out to be not very relevant.” In hindsight, Tarver says, he believes his reluctance to try anti-TNF medication was driven more by fear of perceived risks than by logic.
The researchers also hope that new insights into patients’ preferences will stimulate research in other fields. “Patient preferences get used in all kinds of things—from quality-of-life surveys to cost-effectiveness analysis,” Bewtra says. “Now, all of these studies could be redone and be more accurate by using better-measured preferences.”
She adds, “I think this research is a really big step toward better and better-personalized medicine.”
Posted: May 16, 2016; Updated: March 20, 2018
