Intervention aims to increase timely referrals to hospice care

July 22, 2019

Hospice care provides many benefits to terminally ill patients, including improving quality of life and reducing the burden of symptoms, yet fewer than half of those who die while on Medicare are actually using hospice services. Of those who choose hospice services, the median length of stay is 24 days, much less than the three months recommended by the National Hospice and Palliative Care Organization.

Dr. Alexis Kirk

Dr. Alexis Kirk

A team led by Alexis Kirk, PhD, who earned both master’s and doctoral degrees at the UNC Gillings School of Global Public Health, is aiming to change these trends. Kirk and her colleagues tested an intervention to improve timeliness of referrals to hospice. Historically, a primary reason that hospice services are underused is that physicians delay in referring patients.

A paper on the team’s full findings, “Pilot Test of an Adapted Intervention to Improve Timeliness of Referrals to Hospice and Palliative Care for Eligible Home Health Patients,” was published online in the May 15 issue of the Journal of Palliative Medicine.

The team adapted an intervention originally used to refer nursing home residents to hospice for use with the home health population. They used the original screening questionnaire — asking about care goals, needs and preferences — but changed the eligibility criteria to screen only high-risk or frail patients, making the intervention more appropriate for home health patients.

Nurses involved in the study then completed three intervention steps: 1) administering screening questions to eligible patients (or their caregivers) during an in-person home health visit; 2) reporting screening results to the patient/caregiver and asking them to authorize follow-up with their physician regarding hospice or palliative care (if the patient screened positive); and, 3) if authorized, initiating appropriate referral.

The nine-week pilot was conducted at two North Carolina nonprofit home health agencies. Of the 28 patients who participated in the intervention, 27 (96.4%) screened positive for potential hospice appropriateness. Of those 27 patients, three entered palliative care and one chose to enter hospice. Two other patients who screened positive intended to follow up with their physician about these options.

Home health staff found that the intervention facilitated end-of-life care conversations by structuring the discussion, especially for patients uncomfortable with this topic or with whom the nurse did not have an established relationship. It was suggested that repeating the screening conversation at multiple time points would allow for follow-up and continued conversations about end of life, which is key as many patients are not ready to make decisions about hospice/palliative care during the first discussion.

While the 100% screening question completion rates suggests that patients are willing to discuss their care needs, goals and preferences, the fact that 23 of 27 patients who screened positive did not consider using hospice/palliative care makes it clear that other considerations factor into the decision.

“This work helped further the field of adaptation,” said Kirk. “It builds on methods and tools that practitioners can use on the ground to make evidence work for them, in their context.”

Kirk, an implementation specialist with The Impact Center at UNC’s Frank Porter Graham Child Development Institute, conducted the research during her tenure as a research public policy analyst in the end-of-life, palliative and hospice care program at RTI International. Co-authors on the paper are: RTI colleague Franziska S. Rokoske, PT, MS; Laura C. Hanson, MD, MPH, from the division of geriatric medicine at the UNC School of Medicine; and Morris Weinberger, PhD, Emily R. Haines, BA, Byron J. Powell, PhD, and Sarah A. Birken, PhD, all of whom are with the Gillings School’s department of health policy and management.

Contact the Gillings School of Global Public Health communications team at

Recent News