Subspecialists who care for dementia patients near the end of life are far more likely to introduce a feeding tube than hospitalists who follow such patients.
Compared with nonhospital generalists, hospitalists were 22% less likely to tube feed hospitalized nursing-home residents - and even less likely to tube feed patients who were the most severely impaired (35%). In contrast, subspecialists were five times more likely to insert a tube. When a mixed group of physicians was on the case, rates were even higher, with a 9-fold increase overall and a 9.5-fold increase for severely demented patients.
The findings clearly illustrate that many physicians could benefit from some education about the most appropriate interventions when patients near the end of life enter a hospital, Dr. Joan Teno and her associates reported in the April issue of Health Affairs (2014;33:675-82).
"It may be that subspecialists do not have adequate knowledge about the risks and benefits of using feeding tubes in people with advanced dementia," said Dr. Teno of Brown University, Providence, R.I., and her coauthors. "Hospitals should educate physicians about the lack of efficacy of PEG [percutaneous endoscopic gastrostomy] feeding tubes, compared with hand feeding, in prolonging survival and preventing aspiration pneumonias and pressure ulcers in people with advanced dementia. In addition, hospitals should examine how they staff the role of attending physician and ensure coordination of care when patient hand-offs are made between different types of attending physicians."Such education would bring all physicians up to speed with position statements against tube feeding for this group of patients. The issue sits atop the Choosing Wisely lists of both the American Academy of Hospice and Palliative Medicine and the American Geriatrics Society. The American Academy of Hospice and Palliative Medicine states that "feeding tubes do not result in improved survival, prevention of aspiration pneumonia, or improved healing of pressure ulcers. Feeding tube use in such patients has actually been associated with pressure ulcer development, use of physical and pharmacological restraints, and patient distress about the tube itself."
Dr. Eric G. Tangalos agrees with the concept that tube feeding can impose even more distress on both these patients and their families. "As a medical profession and a society, we have yet to accept some of the futility of our actions and continue to ignore the burdens tube feedings place on patients, families, and the health care system once a hospitalization has come to its conclusion," he said in an interview.
Who's the attending?
Dr. Teno and her team looked at the rate of feeding tube insertion in fee-for-service Medicare patients with advanced dementia who were within 90 days of death and hospitalized with a diagnosis of urinary tract infection, sepsis, pneumonia, or dehydration. The study examined decisions made by four groups of physicians who cared for these patients: hospitalists, nonhospitalist generalists (geriatricians, general practitioners, internists, and family physicians), subspecialists, and mixed groups that included a subspecialist and either a hospitalist or nonhospitalist generalist.
The cohort comprised 53,492 patients hospitalized from 2001 to 2010. The patients' mean age was 85 years. About 60% had a do not resuscitate order, and 10% had an order against tube feeding.
The rate of hospitalists as attending physicians increased from 11% in 2001 to 28% in 2010. The portion of patients seen by a mixture of attending physicians increased from 29% in 2001 to 38% in 2010.
The rates of tube feeding were lowest when a hospitalist or nonhospitalist generalist was the attending physician (1.6% and 2.2%, respectively). Subspecialists had significantly higher rates (11%). The highest rate occurred when there were mixed groups of physicians involved in the patient's care (15.6%).
Using the nonhospitalist generalists as a reference group, the researchers found that hospitalists were 22% less likely to insert a tube overall and 35% less likely to do so when the patient had very severe cognitive and physical impairment.
Conversely, subspecialists were five times more likely to commence tube feeding for all patients and for very severely impaired patients. The mixed groups were the most likely to begin tube feeding - almost 9 times more likely overall and 9.5 times more likely for the most severely impaired patients.
"Our finding that subspecialists had a higher rate of insertions of PEG feeding tubes might reflect their lack of experience in providing care for people with advanced dementia," the authors wrote.
The fragmention factor
The mixed-physician group could be seen as a proxy for discontinuity of care among the attending physicians, they noted. Prior studies have found that such discontinuity was associated with longer hospital stays.