Building a Better
Building a Better Cancer Screening Process
The cancer screening process is complex, and failures can occur at any step—from testing and diagnosis to referral for treatment. And although the reasons the process sometimes fails are not always clear, a better understanding of screening approaches could lead to significant improvements overall.
Some answers could emerge in the next few years from a project known as PROSPR (Population-Based Research Optimizing Screening through Personalized Regimens). This NCI-sponsored research initiative is designed to improve the cancer screening process for all who are eligible for particular screening tests, including underserved populations.
“It is important to understand which steps in the process may fail, because these failures are the kinds of things that lead to disparities in cancer screening,” said PROSPR investigator Dr. Jennifer Haas of Brigham and Women’s Hospital in Boston.
The researchers will focus on developing better ways to implement screening in community settings. At seven sites in the United States, investigators will study screening for the three cancers for which the evidence of potential benefits of screening is strongest—breast, cervical, and colorectal.
“We do a lot of cancer screening in the United States, and there is an accumulating body of evidence that says we don’t do screening as well as we should,” said Dr. Carrie Klabunde of NCI’s Division of Cancer Control and Population Sciences (DCCPS), a leader of the project.
Dr. Douglas Corley, a PROSPR investigator from the Kaiser Foundation Research Institute, added, “This project will look at the whole pathway of cancer screening to get a global understanding of how people go through this process and to identify where the problems are.”
Screening Vulnerable Populations
One of the main study sites for colorectal cancer screening is Dallas. Researchers there will follow more than 30,000 uninsured or underinsured residents served by a dozen health clinics for up to 5 years. The clinics provide primary health care, including cancer screening, regardless of a person’s ability to pay or immigration status.
“In several years, we’ll have a lot of knowledge about the best way to do screening in a large, vulnerable, low-insurance population,” predicted Dr. Celette Skinner of the University of Texas Southwestern Medical Center. “We will share this with other safety-net health care systems.”
In a study that is already under way, participants are using touchscreen computers to help them assess their own risks of developing colorectal cancer. Based on the user’s responses to questions about health and family history, a computer program recommends, for example, that a person be screened with a stool-based test or go straight to a colonoscopy.
“We’re trying, on the one hand, to prevent the underuse of screening and, on the other hand, to prevent the overuse of screening,” said Dr. Skinner. An example of overuse would be when someone who has had a colonoscopy that found no precancerous polyps and does not need to be screened again for a decade continues to get a colonoscopy every few years.
What Influences Screening Outcomes?
PROSPR researchers in California are studying a different aspect of colonoscopy screening: factors that influence outcomes. This is possible thanks to a database of health-related information on 6 million California residents who are insured through Kaiser Permanente.
“We have access to detailed information on these patients, such as which medications they were taking when they were screened,” noted Dr. Corley.
His team will look at factors that may influence outcomes from cancer screening, including race, sex, ethnicity, and the person’s overall health. The instructions a person received before screening and what time of day the person was screened will also be considered.
“The strength of PROSPR is that we can see the contributions of each step in the screening process,” explained Dr. Corley. “This will help us get the best bang for the buck in terms of trying to increase the effectiveness of cancer screening.”
All PROSPR investigators are contributing their results to a central repository, which will be a resource for the field.
Matching Screening Tools to Cancers
Meanwhile, researchers in Philadelphia are looking into whether women are getting the appropriate type of breast cancer screening. This PROSPR site serves a population that is 40 percent African American.
—Dr. Douglas Corley
“We’ve made some progress toward reducing breast cancer mortality, but we’re not close to being all the way there,” said co-leader Dr. Katrina Armstrong of the University of Pennsylvania Abramson Cancer Center. “And among African American women, it’s not clear that the tools we have for early detection are working the way they should be.”
The researchers are comparing the rates of false-positives associated with different types of digital mammography in the population they serve. Black women are more likely than white women to develop triple-negative breast cancer, which is difficult to treat, and some types of mammography may detect this form of the disease better than others, according to Dr. Armstrong.
“The million-dollar question is: Can we figure out how the screening modalities work for different subtypes of breast cancer?” she said.
This information, if available, could be used to improve the detection of breast cancer among all women and reduce false-positive rates.
Another PROSPR breast screening project will test an intervention designed to improve discussions of cancer risk between patients and doctors. Previous studies have suggested that white women who are not at high risk of breast cancer tend to overestimate their risk of the disease, whereas black women, who may be at high risk, tend to underestimate their risk.
The new project will test an intervention designed to foster more accurate self-assessments of risk. Dr. Haas and her colleagues will deliver information about the risk of breast cancer to some participants in a small randomized clinical trial before they see their doctors.
“If you don’t believe you are at risk for breast cancer, then you are probably not going to get recommended screening,” said Dr. Haas. “We hope that this intervention will move both groups toward the center and reduce the disparities in self-perceived risk.”
Realistic Screening Environments
Screening in the community is different from the carefully controlled randomized clinical trials in which new screening tests are usually evaluated. In such trials, participants are invited, guided, and supported throughout the process. But this support may not be present in the community.
“Just because a screening test works in an experimental setting doesn’t mean that it will also work when tried in the community setting,” noted Dr. Pamela Marcus of DCCPS, who is the NCI scientific coordinator for PROSPR. For example, if a screened person gets a positive test result but does not have health insurance, “that person may fall through the cracks of the system.”
This project will show us “what’s working and what is not working in cancer screening,” Dr. Marcus added. “Then we can address these issues and improve the process.”