Skip to Navigation
University of Pittsburgh
Print This Page Print this pages

October 14, 2010

Science 2010: Nontraditional clinical trials

Two Graduate School of Public Health (GSPH) researchers are advocating a greater national emphasis on nontraditional clinical trials that focus on comparative effectiveness research.Morton

“What is comparative effectiveness research?” asked Sally C. Morton, professor of biostatistics and chair of GSPH’s Department of Biostatistics. “A simple definition is: It’s research to find out which treatments work best, for whom and under what circumstances.”

Nationally, medical research is moving toward comparative effectiveness research, or CER, as it seeks alternatives to evidence-gathering in standard clinical trials, according to Morton and fellow presenter KelseySheryl Kelsey, professor of epidemiology and co-director of GSPH’s Epidemiology Data Center.

Their reports were part of the Oct. 8 presentation, “Pragmatic Clinical Trials and Comparative Effectiveness Research,” which was one of the sessions at last week’s Science 2010 symposium.

Morton was a member of the Institute of Medicine (IOM) committee that prepared the report titled “Initial National Priorities for Comparative Effectiveness Research,” which was released in June 2009.

She spoke on “Comparative Effectiveness and You: What, When, How and Why?” which provided an overview of CER. Morton also presented some of the issues that her committee reviewed for the report, which recommended 100 study topics related to a range of diseases, research methods and health care models deemed integral to a healthy U.S. population.

Kelsey spoke on “Does It Really Work? Making Clinical Trials Practical,” by presenting an overview of a five-year CER-based study that published findings in 2009. Kelsey was the principal investigator on the study.

A broader definition, Morton suggested, is that CER addresses the challenge faced by physicians who must choose between treatments when there is not adequate evidence to decide which treatment will work better.

The difference between standard clinical trials and those that measure comparative effectiveness is that the latter, as the term suggests, shift the focus from efficacy of a particular drug or medical therapy to comparing effectiveness of two or more known quantities.

Thus, CER aims to demonstrate the effectiveness of one strategy over another for a certain condition and population, helping doctors and patients make smarter health decisions founded in sound scientific evidence.

“These are not patient trials on one drug to determine efficacy in an academic clinical setting,” as we would see in a traditional clinical trial, Morton said. “These are trials performed in real-world circumstances with medications or therapies that are known. Instead, we’re looking for effectiveness,” she said.

“Why is there a new focus on CER now? There are studies that show that many medical decisions that are made in clinical practice today are actually not based on evidence. One study estimates that up to 50 percent of decisions are not based on evidence,” Morton said.

The characteristics of CER include measuring outcomes and using broad methods and data sources in settings similar to clinical intervention settings.

“We have to define what we mean by evidence,” Morton said. “Do we mean what we’ve learned from clinical trials, do we mean observation studies? Lots of physicians base decisions on experience. Or is it something else?”

There’s a lot of variation in practice patterns, and in which treatments are given to patients even when there’s no difference between them. Sometimes, she said, decisions are based on what the health care insurance will cover, rather than effectiveness.

CER also has its goal to reduce overall health care costs. Morton noted that a recent U.S. Government Accountability Office report states that the cost of the health care system in this country is unsustainable under current conditions.

“We spend a lot more than other countries, but we don’t enjoy as good outcomes, in morbidity and a number of other measures,” Morton said. “The hope is that this approach will summarize existing knowledge, identify gaps in knowing what we should do and filling in those gaps, all while decreasing costs,” she said.

The federal government is backing CER projects and funding them through health care legislation and the American Recovery and Reinvestment Act (ARRA) of 2009, Morton noted. “In fiscal year 2007, the government put about $20 million into CER projects. That increased to $50 million in FY08 and now is up to $88 million in the current fiscal year,” she said.

An additional $1.1 billion from ARRA has been earmarked for CER projects, including $1.1 million that funded the IOM committee’s work.

The committee began with 1,268 CER topics that were nominated by stakeholders and the public via a web site and narrowed them to 82 topics; the other 18 topics were recommended by the committee to fill gaps in the underrepresented health care areas.

The area with the highest number of recommended priorities in the report involves health delivery concerns. To show the range of priorities, Morton cited the examples of cardiovascular comparative research, which she said is a medical priority; comparing the translation testing techniques for CER, which she said reflects a need to develop a method to present that information as a way to inform medical decisions, and school-based health research, which she said is a broad societal issue.

“We also wanted to recommend actions to make sure there is an adequate infrastructure and also a workforce as we go forward,” Morton said. “Thinking about where these studies and data will come from, I want to stress the fact that it’s not about just trials. It’s about experimental studies, occupational studies and a variety of other things, and so we’re going to have to train a workforce. Also, there will be a lot of different data sources coming into play here, so we’re going to have to think about what evidence is allowed — electronic medical records, registries and so on.”

The committee’s report stressed that the 100 recommendations were initial priorities, subject to change. “We wanted to make it clear that this isn’t static, it’s only setting the stage for where we’re going to direct funding for CER,” she said.

That task now falls to the Patient-Centered Outcomes Research Institute, an independent non-profit institute. “It is responsible for establishing the CER research agenda and funding CER,” Morton said. Most of the money will be channeled through the Agency for Healthcare Research and Quality or the National Institutes of Health, she added.

“Is CER really different from what we’ve been doing before? I used to waffle on this question, but I’ve decided it is different. It’s a different philosophy, a different perspective and it’s patient-centered,” Morton said.

“We’ve been moving toward CER for a while, but I’m worried. I’m a statistician. I’m going to have all different types of patients, sample sizes, numbers of variables. Time in the field is increasing, costs are going up. How are we going to do it smarter and cheaper? We’re really re-thinking the paradigm of decision-making. We can’t do this research without changing the way we do things. The big question is, How can we best determine what is the appropriate evidence for the decisions we’re trying to make?”


Sheryl Kelsey’s presentation focused on the Bypass Angioplasty Revascularization Investigation in Diabetic Patients (BARI 2D), the first major trial to focus exclusively on patients with both diabetes and heart disease.

Kelsey cited the study as a successful example of a pragmatic clinical trial, using comparative effectiveness research.

The multicenter, randomized clinical trial tested two treatment strategies for diabetes and tested the effectiveness of maximal medical therapy compared to revascularization for patients with existing diabetes and heart disease. In the study’s design terms, it was a two-by-two factorial trial, Kelsey said.

BARI 2D enrolled 2,368 participants randomized at 49 sites in six countries. Thirty percent were women, 36 percent were minorities and the participants’ age range was 34-89.

“First, all subjects were given intensive medical therapy to control cholesterol and blood pressure and were given counseling, where needed, to quit smoking and to lose weight,” Kelsey said. “Beyond that, we compared whether prompt revascularization, either bypass surgery or angioplasty, was more effective than medical therapy alone.”

At the same time, the study examined which of two standard diabetes treatment strategies resulted in better outcomes. The two strategies were  insulin-provision and insulin-sensitizing, that is, comparing increasing the amount of insulin with making the insulin work better in the body, she said.

“The goals of clinical trials, as well as all medicine, include allowing people to live longer, preventing chronic diseases and making people feel better. We wanted to ask also, Is it practical? And is it cost-effective?” Kelsey said.

Standard clinical trials are characterized as testing an intervention, very broadly defined to include surgical procedures, medical devices, health services delivery and drug trials.

“In drug trials, the classic control is a placebo, whereas a practical control is alternative treatment, that is, looking for comparative effectiveness based on outcomes,” she explained.

“There are variations on this theme, but my definition of comparative effectiveness is to take a diverse patient population and compare active treatments with outcomes where we assess an array of measures. We also look at specific patient groups, defined by a variety of factors,” Kelsey said.

In the United States 24 million people have diagnosed diabetes; the heart disease death rate among those with diabetes is 2-4 times higher than for those without the disease, so there is an inter-relationship. About 65 percent of patients with diabetes die of heart disease or stroke, she noted.

“For BARI 2D, we compared people who had Type 2 diabetes and coronary heart disease who were suitable for either angioplasty, that is, inserting stents, or coronary artery bypass graft surgery — known as CABG — but they didn’t need those procedures urgently. We looked at those who had symptoms,” she said. “The goal in our practical clinical trial was to compare prompt revascularization versus delayed or no revascularization. Remember, revascularization could be either angioplasty or CABG.”

Kelsey noted that physicians were making all the medical decisions in the study.

“One of the practical factors we faced is that before the patients were assigned to treatment, the cardiologist decided which of the vascular treatments, CABG or angioplasty, would be more appropriate for the individual. There was a tendency for people who have more severe disease to get CABG — about 200 participants, versus the angioplasty group, which had about 400 research subjects,” she said.

“It’s also true that the doctors chose which drugs and which dosages, not in the old-fashioned way to pick just one drug and that’s what one gets. There were various adjustments made during the study for dosage and for changing drugs to fit the needs of the patients,” Kelsey said.

“How do you judge success in this trial? Primarily, we wanted to see if people live longer. We wanted to prevent heart attacks. We wanted to avoid further medical procedures and we wanted to relieve angina pain and improve the quality of life.”

The results of the study showed that neither angioplasty nor CABG affected mortality rates compared to intensive medical therapy alone. “This is a pretty important finding,” she said. “With heart attacks, we showed a lower heart attack rate with angioplasty, but there was no reduction using CABG.”

Over the course of five years following the patients, 42 percent of those who originally were just taking medication went on to have either CABG or angioplasty when their symptoms worsened.

Regarding angina pain, both   procedures lowered pain levels, so both were more effective than medical therapy alone. CABG did a little better than angioplasty in that the relief from pain lasted longer, Kelsey said.

“In terms of quality of life: All patients improved in activity level, energy levels and self-rated health. They also had fewer worries about their health over the course of the study,” she said. “That’s not unusual in clinical trials. Being in a trial and getting a lot of tender loving care and attention [alleviates] worry. People felt better after being in the trial. The prompt revascularization did a little better in terms of participants feeling better with less worries.”

A worry for the researchers, she said, are the costs of the study. Per-patient costs averaged $70,750, although that included the procedures, medications, hospital visits and doctor’s office visits. Costs for those who were selected for bypass surgery averaged $80,900 per patient, while those who underwent angioplasty averaged just over $60,000 per patient.

The usefulness of the study’s findings outweighs the costs, Kelsey said. “I think this is a wonderful example of a successful practical clinical trial, with diverse groups of patients, that compared the strategy of choosing which treatment is better for the patient,” she said.

“An important point is that we evaluated existing treatments. There was nothing experimental about the diabetes drugs, or about the medical procedures, which have been ongoing for more than 30  years, and none of the medicines we used was experimental. Certainly, we documented a number of important outcomes, including morality rates and quality of life measures,” Kelsey said.

“For many patients, we found that continuing with medication was as effective in living longer or not having a heart attack. Because of BARI 2D, the cardiologists became convinced that maybe it’s not always the best idea to rush off to the catherization laboratory or to surgery, but that a patient may do be just as well off with good medical therapy.”

—Peter Hart

Filed under: Feature,Volume 43 Issue 4

Leave a Reply