Evaluation of a required palliative care rotation for internal medicine residents

Internal medicine programs are now required to integrate palliative care into teaching conferences, and palliative care content is already tested on the American Board of Internal Medicine examination. Previous research has focused on integration of palliative care into existing rotations and seminars, but none has studied a required inpatient palliative care rotation.

We evaluate the impact of a required palliative care rotation for internal medicine residents in four domains: pain management, non-pain symptom management, communication/ethics, and terminal care.

All second-year internal medicine residents (R2s) completed a required 2-week rotation in inpatient palliative care, and were asked to complete a previously validated palliative care examination immediately before and within 2 weeks after the rotation. During the same year, all interns (R1s) were also asked to complete this examination, as were third-year residents (R3s) who had completed the rotation one year earlier, to provide a basis for comparison. Participation in the examination was voluntary for all.

All interns (100%), and 71% and 87% of R2s and R3s, respectively, completed the examination. Mean examination scores improved by 12.4% between internship and start of the palliative care rotation, and by an additional 9.7% by the end of the rotation. Mean examination scores were consistent from the R2 to the R3 year. At all levels, residents scored well above the national average on the examination. Scores improved in all domains measured.

Palliative care knowledge, as tested by objective examination, improves during internal medicine residency at our institution and specifically over the course of a required, 2-week palliative care rotation. Further study is warranted to determine the relative contributions to this improvement from the palliative care rotation itself, the institutional culture and/or Rochester residents' preexisting interest in the bio-psychosocial model.