When it comes to what some consider overly aggressive treatment at the end of life, the attitudes of both cancer patients and their oncologists matter, a new study from Rutgers University concludes.

"This is the first time that it has been shown that physician attitudes affect the quality of end-of-life care independent of patient treatment preferences," said lead author Paul Duberstein, chair of the department of health behavior, society and policy at the Rutgers School of Public Health.

Previous studies have not looked at the two groups in tandem, he said. Duberstein, who came to Rutgers in November, conducted the study while at University of Rochester School of Medicine and Dentistry, working with researchers at the University of California-Davis, Tulane University and Weill Cornell Medical College.

While small — it followed 265 patients with advanced cancer and 38 of their oncologists — the study establishes an association between treatment and both physician and patient attitudes. It was not designed to measure the strength of the relationships.

The study, which was published in the Journal of Pain and Symptom Management, looked at rates of chemotherapy use and hospitalizations or emergency department visits in the last month of life. Those forms of treatment have been associated with worse patient quality of life, tougher bereavement for caregivers, and greater moral distress, burnout and turnover among doctors. The study found that rates were higher when patients were willing to try experimental treatments with no known benefits when all else had failed and when doctors were more comfortable making decisions for patients, what the researchers called “biomedical paternalism.” The study found there was more aggressive care in the last month if patients were treatment risk takers or had negative views on palliative care, or if a paternalistic physician was involved.

All of the patients in the study had incurable cancer. Duberstein said that aggressive care in their last month is not always bad care. However, the study said most individuals prefer to die at home rather than in a hospital.

Patients, Duberstein said, often have no way of knowing whether their oncologist has a history of pursuing treatment to the bitter end. "The attributes of individual doctors matter," he said. "What's interesting just from a policy perspective is that patients do not have access to this information." He thinks that information should be available.

While he has no way of knowing whether the physicians in this study actually were asked by patients to make late-stage decisions about care, he theorized that the most paternalistic among them might be less disposed toward learning about patient values. Once in charge, doctors are beset by incentives that reward more care. "The default, I think, is always to treat," Duberstein said.

He thinks that greater attention to patients’ beliefs, particularly their anxiety about death, might lead to better decision-making by both patients and doctors.