If the world had unlimited dollars for cancer research, investigators would compare new treatments to standard therapies using the most rigorous, painstaking method known to medical science: a randomized clinical trial, in which competing approaches are tested under specific experimental conditions.
In the real world, where budgets are tight and pressures for rapid results are intense, such costly, laborious studies are not always feasible or, in fact, necessary. For all their scientific precision and control, randomized phase 3 clinical trials provide a kind of "hothouse" view of how cancer treatments perform in patients.
For a more practical understanding of the risks and benefits of therapy, investigators often employ an approach known as comparative effectiveness research (CER).
"CER evaluates the effectiveness of care, and measures not how well treatments can work but how well they do work in normal conditions," says Dana-Farber's Deborah Schrag, MD, a CER investigator.
Whereas randomized clinical trials generate their own data for analysis, CER studies frequently draw on existing sources of information, collected by doctors, insurance companies, and others in the course of patient care. The result is a ground-level ability to explore how well treatments work relative to one another – information that can have a powerful impact on the therapies that patients ultimately receive.
Consider, for example, a screening technology known as three-dimensional computed tomography (3D CT), which produces images that can be rotated so doctors can see a tumor's precise contours. Introduced about 15 years ago, 3D CT has become the imaging technology of choice for locating and tracking non-small cell lung tumors.
Yet when it was first introduced, no one had examined whether the advantages of 3D CT were significant enough to justify its higher cost over conventional imaging techniques.
Dana-Farber's Aileen Chen, MD, MPP, was able to supply a scientifically grounded answer. She and her colleagues analyzed data collected on more than 15,000 Medicare patients who received radiation therapy for advanced non-small cell lung cancer between 2000 and 2005. By statistically controlling for differences in age, sex, and geographic region, Chen's team concluded that screening with 3D CT was associated with higher survival rates.
Although 3D CT has been widely accepted, some parts of the country have been slower to adopt it. With the help of Chen's research, physicians can make better-informed decisions about radiation therapy, leading to better outcomes for patients.
"Comparative effectiveness research starts with the fact that many decisions in cancer treatment, while based on expert opinion, are not backed by evidence of effectiveness," explains Jane Weeks, MD, MSc, director of Dana-Farber's Center for Population Sciences.
"Clinical trials, in which potential therapies are tested in patients, provide us with a lot of rich information, but there isn't enough money or time for them to tell us all we want to know."
CER also helps to reduce the cost of cancer care by revealing which treatments provide the most value for the money. The National Institutes of Health estimates cancer care in the United States cost nearly $264 billion in 2010. Being able to cut even a small percentage of these expenditures could result in billions of dollars in savings.
Recognizing this fact, the U.S. government dedicated $1.1 billion for CER projects in 2010 and 2011. Of these funds, $7 million was awarded to Schrag and Weeks' labs for CER-related studies at Dana-Farber.
Comparative effectiveness research fills in some of the knowledge gaps that clinical trials may leave behind. Patients who participate in trials are randomly assigned to receive either a new, potential therapy or a standard therapy (or, in some cases, an inactive agent known as a placebo).
Many such trials are also blinded, meaning neither the investigator nor the patient knows which treatment any individual is receiving. These guidelines ensure that the results of the study accurately reflect the workings of the novel treatment itself, not unforeseen factors (known as "confounders").
However, the guidelines also limit the number of people who can participate in certain trials: older patients, people with other health conditions, and those who live far from major cancer centers may not be able to take part, so it remains unknown which treatments work best for these groups.
In addition, clinical trials are not often used to examine the effectiveness of non-drug treatments such as surgery and alternative therapies, or of diagnostic techniques such as advanced imaging procedures.
The "raw material" for comparative effectiveness research is information collected on patients during their treatment and recovery. In that respect, the health-care databases maintained by health insurance companies, Medicare, Medicaid, hospitals, and clinics are a gold mine. These databases contain information on thousands (or even millions) of patients, ensuring that the results of CER studies are statistically sound and relevant to people outside a restricted study sample.
Weeks is currently leading the Cancer Care Outcomes Research and Surveillance Consortium (CanCORS), a 6-year, National Cancer Institute-funded study of patients newly diagnosed with lung or colorectal cancer. By following 10,000 patients across the country through their treatment, she hopes to address specific questions about why certain groups, such as the elderly and members of certain minorities, sometimes receive lower-quality care. The results promise new momentum for improving such care.
Another advantage of CER studies is that they can often be done faster than randomized phase 3 clinical trials. All that's needed is access to existing health-care databases and sufficient computing power to sift and sort the data. The rigor of analyzing the information – of making meaningful comparisons, finding relevant trends, and drawing cogent conclusions – calls for well-honed scientific skills.
In some cases, CER investigators don't rely on others' databases, but collect information themselves. While this can take longer, it allows researchers to work and talk directly with patients.
Understanding how patients feel is critical to understanding the benefits of treatment. And it eliminates some of the uncertainty involved in decoding billing information from insurers' records. CER investigators at Dana-Farber are currently connecting with patients for a variety of studies aimed at improving care.
"Sometimes the questions we ask in comparative effectiveness research are less cutting-edge than in randomized trials," says Schrag. "But they are highly important."
Weeks was first introduced to CER during her postgraduate training at Brigham and Women's Hospital. "I saw the power this kind of research had in cardiology and general medicine and thought, 'Why does this not exist in cancer?'" she says.
When Weeks joined Dana-Farber in 1989, she began to pursue CER in cancer at a time when few others were doing so. Since then, she and her team have trained members of the next generation of CER investigators, attracting top-tier trainees from around the world.
CER investigators, Weeks explains, combine the skills of a statistician with those of a clinician. Dana-Farber has a variety of faculty members who are proficient in the technical aspects of CER but also understand, from their clinical experience, the right questions to ask – and how to ask them.
The future of CER may dovetail with projects such as Profile, Dana-Farber and Brigham and Women's Hospital's new research effort to scan patients' tumor tissue for nearly 500 gene mutations known or suspected to be involved in cancer. The genetic data collected for Profile will be linked to data from patients' medical records, making it possible to track which diseases respond best to which treatments.
If that sounds like the exact type of question CER seeks to answer, it's because it is. Weeks hopes to focus her future research on why treatment outcomes for specific types of cancer often vary from one clinic or part of the country to another. The goal is to ensure that patients receive the best treatment, no matter where they are.
“People I talk to hear that there are different practice patterns and want to know who is right,” says Weeks. “Give us time. We are going to try to answer that question.”
Over the years, comparative effectiveness studies by Dana-Farber population scientists have addressed a variety of issues relating to cancer treatments and the cost of cancer care. Some examples of their findings are:
Paths of Progress Spring/Summer 2012 Table of Contents
Dana-Farber Cancer Institute, 450 Brookline Avenue, Boston, MA 02215 | Call us toll-free: