Rethinking comparative effectiveness research

Jan. 19, 2016

Efficacy-based randomized clinical trials are designed to determine how well a particular — and usually unproven — intervention might work under ideal circumstances compared with placebo. These trials, which use data-driven numeric outcomes and exclude critical patient groups, bear little resemblance to real-world practice, yet their findings may influence clinical decision-making.

Effectiveness trials, on the other hand, assess how well a particular intervention works compared with other options commonly used in clinical practice. Ideally, comparative effectiveness research (CER) also identifies the features that predict which intervention would be most effective in any one patient. This type of research is intended to foster evidence-based changes in practice that reduce costs and improve patient care.

The 2009 American Recovery and Reinvestment Act designated $400 million in discretionary funds to accelerate CER efforts, and among other areas of study, the Institute of Medicine prioritized research comparing different strategies of introducing biologics in the treatment algorithm for inflammatory bowel disease (IBD).

Arguably the most well-known IBD effectiveness study is the landmark SONIC trial, which compared early use of infliximab plus azathioprine to monotherapy with either drug in patients naive to corticosteroids, immunomodulators and anti-tumor necrosis factor (anti-TNF) agents. The findings, published in the New England Journal of Medicine in 2010, showed that patients receiving combination therapy had higher corticosteroid-free remission rates at 26 weeks — the primary endpoint — than did patients receiving infliximab or azathioprine alone. Infliximab alone was also shown to be more effective than azathioprine monotherapy.

Although the SONIC trial led to increased use of combination therapy in Crohn's disease (CD), the small sample size of this and other CER trials is a significant limitation. Effectiveness trials require substantially larger patient populations than placebo-controlled trials, which makes them far more costly and harder to recruit.

"It's becoming increasingly difficult to recruit and enroll patients, which has led to the globalization of clinical trials — sponsors are now going far afield, to eastern Europe, India and Russia in order to conduct trials more easily and cost effectively," explains Edward V. Loftus Jr., M.D., a gastroenterologist at Mayo Clinic's campus in Rochester, Minnesota.

Is there a better way?

One alternative that avoids some of the challenges of head-to-head prospective trials is network meta-analysis, which allows investigators to make inferences about the effectiveness of interventions that have not been directly compared with one another. For example, in a 2014 study published in Alimentary Pharmacology and Therapeutics, Mayo Clinic researchers performed a network meta-analysis of all approved biological agents for treatment of ulcerative colitis in first-time biologic users.

On indirect comparison of nine trials, no single agent proved clearly superior to others. Although patients taking infliximab were twice as likely to achieve remission as those on adalimumab, the difference was not statistically significant.

A similar Bayesian network meta-analysis of 17 randomized controlled trials comparing six biologic agents for Crohn's disease suggested that infliximab was the most effective for inducing clinical remission in first-time biologic patients. All other agents seemed to work equally well in patients who responded to the index biologic agent, although adalimumab, natalizumab and infliximab were favored.

The investigators note that their findings, which were published in Mayo Clinic Proceedings in 2014, have lower confidence because they are based solely on indirect treatment comparisons and do not take into account cost, safety and patient preference — all of which must be factored into treatment decisions.

In another approach, Mayo Clinic investigators took advantage of the unique partnership between Mayo Clinic's Center for the Science of Health Care Delivery and Optum Labs to compare the effectiveness of anti-TNF agents approved for treatment of CD using a large health care claims database. In a cohort of biologic-naive patients, they found that infliximab was superior to both adalimumab and certolizumab for clinically relevant outcomes, including reduced new steroid prescriptions and CD-related hospitalizations and surgeries. Study findings were presented at Digestive Disease Week 2015.

Dr. Loftus notes that although head-to-head prospective trials remain the holy grail, they present ongoing challenges.

"We need to do more CER, but who is going to pay? Pharmaceutical companies would prefer placebo-controlled trials because it is easier to demonstrate the efficacy of their therapeutic agents," he says. "Having said that, in the pipeline are several companies willing to perform comparative trials of anti-integrin drugs versus anti-TNF agents. We can also get some circumstantial evidence from other studies, but ultimately, we will need big, prospective trials. Team medicine and team science are essential for getting National Institutes of Health funding for multicenter trials. And as investigators, we have to be increasingly inventive and creative."

For more information

Colombel JF, et al. Infliximab, azathioprine, or combination therapy for Crohn's disease. New England Journal of Medicine. 2010;362:1383.

Singh S, et al. Letter: Comparative efficacy of biological therapy in patients with ulcerative colitis. Alimentary Pharmacology and Therapeutics. 2014;39:1432.

Singh S, et al. Comparative efficacy of biologic therapy in biologic-naive patients with Crohn disease: A systematic review and network meta-analysis. Mayo Clinic Proceedings. 2014;89:1621.