Standardization of Inpatient CPR Status Discussions and Documentation Within the Division of Hematology-Oncology at UPMC Shadyside: Results From PDSA Cycles 1 and 2.

Abstract

In December 2016, 49% of patients admitted to inpatient oncology services at University of Pittsburgh Medical Center Shadyside Hospital had cardiopulmonary resuscitation (CPR) status discussion documentation before discharge. The aim of this project was to improve the rate of CPR status conversations. During Plan-Do-Study-Act (PDSA) cycle 1, a stakeholder workgroup was formed in January 2017 by oncology faculty, fellows, nurses, advance practice providers (APPs), medicine housestaff, and palliative care faculty. All oncology clinicians and inpatient team members were reminded weekly to discuss and document CPR status preferences. APPs received training on efficient and effective CPR status assessment from palliative care faculty. Oncology leadership received monthly e-mail updates of CPR status documentation rates and endorsed CPR status best practice guidelines. For PDSA cycle 2, patient charts without CPR status documentation in March 2018 were reviewed, and themes were shared with oncology leadership and reviewed with APPs. After PDSA cycle 1, CPR status assessment rates increased from 49% to greater than 80%. In 2017, more than 1,500 more CPR status discussions were documented than in 2016. The percentage of patients discharged with "comfort measures only" or "do not resuscitate" orders increased from 14.2% (95% CI, 9.5% to 19.0%) to 19.8% (95% CI, 15.6% to 24.0%). For PDSA cycle 2, charts of 60 patients without CPR assessment were reviewed. Of these, 52% were admitted overnight by nocturnists and 48% by daytime APPs. Fifty-five percent of patients (n = 33 of 60) had metastatic disease. CPR status was documented on previous admissions for 53% of patients (n = 31 of 60) in the past 12 months. Fifteen percent (n = 11 of 60) were admitted for scheduled inpatient chemotherapy. A multipronged approach significantly increased CPR status assessments. More patients transitioned to comfort measures only or do not resuscitate when their preferences were clearly assessed and documented.