Click the image below to view the poster full screen.


Title of Poster: Reducing Readmission through High Quality ACP Conversations and POLST

Names and Affiliations:

Michael Kersten1, Emelia Altschul1, Ashish Kaul1, Samuel Falkson1,2, Ryan Van Wert MD1

1Vynca, Palo Alto, CA, USA

2Stanford University School of Medicine

Background/Objective/Research Question:

From September 2018 to October 2019, the Marquis Mill Park Skilled Nursing Facility (SNF) in Oregon adopted the cloud-based advance care planning portal, Vynca, for creating, storing, and retrieving end-of-life care POLST forms.  During this time, Mill Park created 147 POLSTs of which 26 indicated DNR/Comfort Measures only.  We set out to determine if POLST creation activity and DNR/Comfort choices correlated with 30-day readmission rates.

Methods/Study Design:

Our data included all discharges from this SNF from September 2018 to October 2019. We filtered for POLST presence or absence and POLST choices against discharge outcomes. Between September 2018 to October 2019, there was a period (March to August of 2019) when Mill Park was not actively engaged in POLST creation. Thus we also compared readmission rates between times when POLST forms were and were not being created.


Patients with POLSTs featuring DNR/Comfort choices had lower 30-day readmission rates (3.8%) compared to patients with no POLST or with POLSTs featuring CPR/Full code choices (9.0%). Additionally, 30-day readmission rates during times with high POLST creation activity were lower (8.4%) compared to times with low POLST creation activity (9.8%).


Having a POLST with DNR/Comfort choices appears protective of 30-day readmissions. While patients should make POLST choices that reflect their values, DNR/Comfort choices often indicate a more thorough conversation regarding end-of-life care.  Our results suggest that scaling in-depth, nuanced end-of-life conversations with nursing home patients can decrease unwanted healthcare utilization, improve end-of-life care, and help SNFs financially.