Physician characteristics associated with clinical trial enrollment


  • Mary Kay Barton MD

  • Note: The name of this section has been changed from “News & Views” to “Perspectives: Research in Context.” It continues to provide the context for major developments in cancer prevention, detection, and treatment.

A recent study found that specialty type, involvement in teaching, and affiliation with a Community Clinical Oncology Program (CCOP) or a National Cancer Institute (NCI)-designated cancer center were factors associated with the level of physician participation and patient enrollment in clinical trials (J Natl Cancer Inst. 2011;103:384-397).

Carrie Klabunde, MHS, MBA, PhD, an epidemiologist in the Health Services and Economics Branch of the Applied Research Program at the NCI in Bethesda, Maryland, and colleagues performed a survey-based study to examine associations between clinical trial participation and characteristics of physicians and their practice setting. Specialty physicians who care for patients with lung and colorectal cancer were identified by patients enrolled in the Cancer Care Outcomes Research and Surveillance Consortium (CanCORS), a national effort that is examining the care and outcomes of more than 10,000 patients with lung or colorectal cancer. Of the 6871 physicians identified, 4188 responded, and after excluding very recent medical school graduates and those who were not medical oncologists, radiation oncologists, or surgeons, 1533 physicians remained in the final study population.

Physician Characteristics

Of those physicians who responded, about 53% were surgeons, 32% were medical oncologists, and 15% were radiation oncologists. Most surgeons and radiation oncologists worked in hospital-based practices, and most medical oncologists worked in office-based practices.

A total of 869 physicians (about 57%) responded that they referred or enrolled at least one patient in cancer clinical trials within the previous year. The percentage differed by specialty, with about 88% of medical oncologists, 66% of radiation oncologists, and 35% of surgeons responding positively. Among the physicians referring patients to trials, the mean number of patients referred or enrolled in trials over the previous year was about 17 for medical oncologists, 12 for surgeons, and 9 for radiation oncologists.

On multivariate analysis of the nonsurgical physicians, the characteristics associated with referring or enrolling patients in clinical trials were:

  • Being a medical oncologist (22% higher than radiation oncologists).

  • Being in a hospital-based practice (19% higher).

  • Teaching medical students or residents (11% higher).

  • Being affiliated with an NCI-designated cancer center (10% higher).

  • Being affiliated with a CCOP (8% higher).

Among the surgeons, factors associated with referring or enrolling patients in clinical trials included:

  • Being a surgical oncologist (26% higher than general surgeons).

  • Seeing 5 or more patients with lung or colorectal cancer per month (13% higher).

  • Attending tumor boards on a monthly or weekly basis (12% and 17% higher, respectively).

  • Being affiliated with an NCI-designated cancer center (10% higher).

  • Teaching medical students or residents (8% higher).

A secondary analysis of the 760 physicians who were affiliated with a CCOP or NCI-designated cancer center showed similar findings on multivariate analysis.

Illustration 1.

Characteristics associated with referring or enrolling patients in clinical trials included being a medical oncologist, being affiliated with a National Cancer Institute-designated cancer center or Community Clinical Oncology Program, teaching medical students or residents, and being in a hospital-based practice.

Among the 869 physicians who reported enrolling patients within the previous year, about one-half participated in trials sponsored by cooperative groups and pharmaceutical companies. Compared with medical oncologists, radiation oncologists and surgeons more often participated only in trials sponsored by cooperative groups. Fewer than 10% reported they only enrolled patients in pharmaceutical trials.

Characteristics associated with enrolling a greater number of patients among the nonsurgical specialists on multivariate analysis included:

  • Affiliations with an NCI-designated cancer center (30% higher than radiation oncologists).

  • Being a medical oncologist (27% higher).

  • Seeing 20 or more patients with lung or colorectal cancer per month compared with 5 or fewer (22% higher).

  • Income incentives for enrolling patients (20% higher).

  • Attending weekly tumor boards as opposed to quarterly or less often (15% higher).

  • Being in a practice of 11 to 20 physicians as opposed to one with 5 or fewer (13% higher).

  • Teaching students or residents (12% higher).

  • Spending 60 minutes or more with new patients (8% higher).

  • Being in an office-based practice versus a hospital-based practice (8% higher).

The results for the surgeons were similar, but also included higher enrollment numbers for specialized surgeons versus general surgeons. In addition, surgeons aged younger than 60 years and female surgeons more often referred or enrolled patients in clinical trials.

Patricia Ganz, MD, the senior author of this study and director of cancer prevention and control research at the Jonsson Comprehensive Cancer Center at the University of California at Los Angeles, says she believes the major barriers to physician referral to trials are lack of time and infrastructure as well as belief in the importance of research and randomization. In addition, she says many physicians would rather treat patients using personal preference and reimbursement criteria. “There are believers and nonbelievers,” she says. “The believers will try to have lots of trials open and will approach every eligible patient. Among other physicians, even when trials are available at their institution, they will not take the time unless they are the study coordinator or the patient inquires. This is the root of the problem from my standpoint.”

Study Limitations

The authors noted several study limitations. The participating physicians, for example, were not a nationally representative group because they were taken from the CanCORS database. The responders may have been inherently different from the nonresponders in that responders may have been more likely to be trial participants and may potentially have overestimated their trial participation. In addition, the survey obtained data from physicians self-reporting about their participation and their objective characteristics, not including anything subjective regarding beliefs and attitudes toward trials. Further, patient characteristics were not collected.

“This study raised some interesting and important issues, but really did not drill down into the specifics of why physicians, who are part of the infrastructures already established, are not fully participating,” says Lori Minasian, MD, chief of the NCI's Community Oncology and Prevention Trials Research Group. However, the study is a survey that is meant to be informative and help identify issues, which could be addressed, she adds.

Clinical Implications

The authors state that this study helps better define the characteristics of physicians who do and do not participate in clinical trials. These data, they say, identify possible areas of intervention to help improve physician participation and improve trial accruals.

For example, even though this study reinforces others showing that affiliation with an academic center or cooperative group increases physician trial participation, only two-thirds of physicians with those affiliations refer patients to clinical trials. Because these physicians already have the infrastructure in place that facilitates clinical research, targeted outreach may be the most efficient way to increase trial accrual.

Targeting beyond the oncology specialties may also help accrual; only one-third of surgeons referred their patients to clinical trials. However, the fact that there are more drug trials than surgical trials may be part of the reason. Further, this study did not examine other specialties such as gynecology, internal medicine, or family practice and certainly the issue goes beyond surgeons, Dr. Minasian says. The authors wrote that all specialties could be targeted by more emphasis being placed on research during medical education or by establishing multidisciplinary cancer care teams to enhance communication between oncologists and other specialists.

An accompanying editorial (J Natl Cancer Inst. 2011;103:357-358) by Dr. Minasian and her colleague Ann O'Mara, MD, cites prior data showing that more than 70% of the general public was likely to consider participation in a clinical trial, but only 6% of their physicians offered it to them. They also note that one implication of the study is already being addressed through a new standard by the Liaison Committee on Medical Education, the accrediting body for medical schools in the United States and Canada, expanding the time that must be spent in training learning about the clinical trial process.

According to Dr. Ganz, the take-home message of this study is that physician referral, or lack thereof, is probably one of the biggest barriers to clinical trial enrollment. “If new treatments could not be given without patients enrolling in a trial, we would answer questions much faster. However, neither patients nor doctors are willing to face the fact that sometimes we do not know what the right answer is, and that a study is the best way to find out,” she says.