AMA Journal of Ethics^®

Coulourides Kogan A, Brumley R, Wilber K, Enguidanos S. Physician factors that influence patient referrals to end-of-life care. Amer J Manag Care. 2012;18(11):e416-e422.

As evidence has made the benefits of end-of-life (EOL) care clearer, the provision of appropriate, high-quality EOL care has become part of the national health care agenda. EOL care has been associated with reduced costs, improved quality of life, and, in some instances, longer survival [1-3]. Furthermore, EOL care is highly desired by patients and their families, who may prefer to avoid aggressive and burdensome care in the context of a poor prognosis [4]. Both California and New York have enacted laws requiring that EOL information and counseling be offered to patients with terminal illnesses [5-7], and other states are considering similar laws [8]. Yet access to adequate EOL care is variable. Patients, families, clinicians, and institutions contribute to underutilization of EOL care [9] and data suggest substantial variation in physicians’ referral to EOL care [10-13].

Coulourides Kogan et al. delve into the issue of physician-related barriers to end-of-life care by examining whether physicians’ personal experiences or level of comfort with discussing end-of-life care affect referral patterns [14]. Using a cross-sectional sample of physicians from a large health maintenance organization in Southern California, the authors administered a survey to assess self-reported patterns of referral to end-of-life care services, which included home-based palliative care and hospice. Data on demographics, medical practice background, physicians’ comfort with discussing end-of-life care, and their personal experience with hospice were also collected. The authors then used logistic regression models to identify factors associated with referral to end-of-life care.

Eighty-three (83) percent of respondents reported at least a single referral in the last year and 50 percent reported four or more referrals over the past year. Eighty (80) percent of respondents also expressed comfort conducting end-of-life conversations. Less than half (43 percent) reported having had a family member receive hospice care.

The authors found that being a family or internal medicine practitioner and self-reported comfort with having end-of-life care discussions were significantly associated with referral, and age was inversely associated with likelihood of referral. Family and internal medicine practitioners were twice as likely to refer frequently than were specialty and emergency medicine practitioners, and those who reported comfort with end-of-life care discussions were five times as likely to refer frequently than those who did not. Previous personal experience with hospice did not significantly affect referral patterns.

What to make of these findings? The fact that physicians’ comfort level with conducting EOL conversations was positively associated with referral to EOL care services is hardly surprising. What makes the finding noteworthy is that physician comfort with having EOL discussions is a modifiable factor and, hence, a possible target for educational interventions. The authors also found that younger physicians were more likely to refer patients to EOL care services. As the authors surmise, this may reflect changes in medical education over the past decade; training in EOL care is now a mandated part of the medical school curriculum. Family and internal medicine practitioners are perhaps more likely to initiate EOL discussions because they have a “closer relationship” with their patients [14].

While these findings are intriguing, they also warrant qualification. The study has significant limitations. Given the survey design and the chosen outcome measure of self-reported referral to EOL care, the results are subject to both non-response and recall bias. The rate of referral to EOL care and rate of physicians who reported feeling comfortable with having EOL discussions were quite high (80 percent). Physicians likely to respond to this type of survey may also be more likely to refer to EOL care services than nonresponders. Also, because referral to EOL care was self-reported, the validity of the authors’ conclusions would be strengthened by administrative data or some other objective measure demonstrating that physicians’ actual practices correlated with their reported practice.

The authors conclude that, on an organizational level, physician comfort with having EOL discussions should be a target for quality improvement since this may lead to increased referrals to EOL care. However, changing physician behavior may not be so straightforward. A large multicenter trial of a complex intervention to improve EOL care in intensive care units did not appear to improve the majority of palliative care elements; nor did it improve family satisfaction or nurse and family ratings of quality of death [15]. Although changes in clinician attitudes were not measured, objective measures of physician practice and outcomes were not significantly changed by the intervention. The results of this trial exemplify how resistant physician behaviors may be to change. While it is tempting to accept Coulourides Kogan and colleagues’ conclusion, whether referrals to EOL care can be increased by simply promoting physician comfort with EOL discussions is uncertain.

Furthermore, physician-related factors are only one type of barrier to EOL care [9]. Because of these barriers, other methods to increase referral to EOL care, such as the use of screening criteria or triggers, have been advocated. The Center to Advance Palliative Care has published criteria for identifying hospitalized patients for whom a palliative care needs assessment would be appropriate [16], and the IPAL (Improving Palliative Care) Project offers resources to structure and implement palliative care initiatives [17]. Although many hospitals are moving towards developing screening criteria specific to their patient populations, referral to EOL care may or may not be automatic when patients meet them.

Physicians are the “gatekeepers” to EOL care services, and since prior studies have demonstrated variability in referral to EOL care, ways to circumvent or remove barriers to physician referral merit consideration. The work of Coulourides Kogan and colleagues adds to an ongoing discussion of the challenges and barriers to quality EOL care and highlights the fact that physicians and the systems in which they provide care may need to be targets of interventions. Future studies will be necessary to determine what interventions will most effectively result in increased physician referral of appropriate patients to EOL care.