Implementation
To quantify the perceived delay in antibiotic prescribing, a pre- and postprotocol retrospective chart review of patients who presented with FN to the RLRVAMC ED was conducted. Patients were identified through the electronic health record (EHR) based on 3 criteria: recorded/reported fever as defined above, ANC < 1000 cells/mm3, and administration of cancer treatment (IV and oral) within 4 weeks. The data collected in the postimplementation phase were identical to the pre-implementation phase. This included timing of blood cultures, choice/appropriateness of antibiotics based on guidelines, and length of admission. The pre-implementation period started on August 1, 2018, and ended on August 1, 2019, to allow for an adequate pre-implementation sample size. The protocol was then implemented on October 1, 2019, and data collection for the postimplementation phase began on October 1, 2019, and ended on October 1, 2020.
The protocol was accompanied by EHR order sets initiated by both nurses and health care practitioners (HCPs), including physicians, nurse practitioners, and physician assistants. The nursing order set consisted of vitals and appropriate laboratory monitoring, and the practitioner order set housed medication orders and additional clinical monitoring for more patient-specific scenarios. On identification of at-risk patients, the nursing staff could initiate the neutropenic fever protocol without consulting an HCP. The patient was then assigned a higher acuity rank, and the HCP was tasked with seeing the patient immediately. In conjunction with a complete physical assessment, the HCP ordered appropriate antibiotics through the designated order set to streamline antibiotic selection. Antibiotic options included cefepime or piperacillin-tazobactam, and vancomycin when clinically indicated. Alternatives for patients allergic to penicillin also were available. The protocol intended to streamline workup and antibiotic selection but was not designed as a substitute for solid clinical decision making and complete assessment on behalf of the HCP; therefore, additional workup may have been necessary and documented in the EHR.
Findings
This patient population comprised 17 patients pre-implementation and 12 patients postimplementation, most of whom had solid tumor malignancies (88.2% and 83.3%, respectively) receiving platinum, taxane, or antimetabolite-based chemotherapy. In the pre-implementation group, most patients (70.5%) coming through the ED were treated with palliative intent. Only 25% of these received any prophylactic granulocyte-colony stimulating factor (G-CSF) based on risk for FN. The mean time from triage to the first dose of antibiotics decreased from 3.3 hours before protocol implementation to 2.3 hours after. Only 6% in the pre-implementation group compared with 17% in the postimplementation group received the first dose of antibiotics within the recommended 1-hour interval from triage. The most common antibiotics administered were cefepime and vancomycin. Eleven patients in each group (65% and 92%, respectively) were admitted to the inpatient service for further care, with 10 and 8 patients, respectively, experiencing a hospitalization > 72 hours. Of note, 41% of patients died pre-implementation vs 17% postimplementation.
Interpretation
The goal of this protocol was to optimize ED care of patients presenting with FN to better align with guideline-recommended time lines and antibiotics. The mean time from triage to administration of antibiotics decreased by 1.0 hour from the pre- to postimplementation phase, similar to the study by Mattison and colleagues.3 When removing an outlier from the postimplementation group, the mean time from triage to first dose further decreased to 1.8 hours. The percentage of patients receiving antibiotics within 1 hour of triage nearly tripled from 6% to 17%. Additionally, the percentage of patients empirically treated with appropriate antibiotics consistent with NCCN/ASCO/IDSA guidelines increased from 65% to 83%. Although goals for the optimization of care have not yet been reached, this protocol is the first step in the right direction.
