Improving Time to Antibiotic Administration for Bone Marrow Transplant Patients With First Fever
BACKGROUND AND OBJECTIVE: Timely antibiotic administration in immunocompromised patients is associated with improved outcomes. The aim of our study was to decrease the mean time to administration of antibiotics in hospitalized bone marrow transplant patients with fever from 75 to <60 minutes.
METHODS: By using the Model of Improvement, we performed plan-do-study-act cycles to design, test, and implement high-reliability interventions to decrease time to antibiotics. Nursing, physician, and pharmacy interventions were successfully applied to improve timely antibiotic administration.
RESULTS: The study period was from April 2014 through March of 2017. Through heightened awareness, dedicated roles and responsibilities, a standardized order set specifically used for first fever patients, notification to the pharmacy about newly febrile first fever patients through a dedicated order, the creation of a dedicated sticker (“STAT first dose antibiotic, give directly to nurse”) to be printed when antibiotics were entered via the order set in the pharmacy, and prioritization of antibiotic delivery on arrival on the floor, we saw an increase in the percentage of patients receiving antibiotics within 60 minutes of documented fever from a mean of 40% to over 90%. Our mean time for antibiotic administration decreased from 75 to 45 minutes.
CONCLUSIONS: Implementation of a standardized process for notifying providers of new fever in patients, prioritization of antibiotic preparation in the central pharmacy, and timely antibiotic order entry resulted in improved times to antibiotic administration in the febrile bone marrow transplant population.
- BMT —
- bone marrow transplant
- CCHMC —
- Cincinnati Children’s Hospital Medical Center
- PDSA —
- QI —
- quality improvement
- RN —
- registered nurse
- STAT —
Sepsis and septic shock from bloodstream infections are a growing health care burden in hospitalized pediatric patients.1 Bloodstream infections occur in nearly 30% of patients undergoing bone marrow transplant (BMT)2,3 and are associated with prolonged hospitalization, septic shock, increased rates of intensive care admission, prolonged antibiotic treatment, and increased mortality.2,4–11 Fever is often the first symptom of sepsis,12,13 and the initiation of broad-spectrum antibiotics within 60 minutes of recognition of sepsis or septic shock has been shown to reduce morbidity and mortality.12–14 Considering the impact of sepsis on patient outcomes and the health care system, comprehensive management strategies have been developed to improve early recognition and timely interventions in pediatric patients.14,15
In previous quality improvement (QI) strategies, the rapid administration of antibiotics in immunocompromised patients presenting to the emergency department has been the primary emphasis.16–18 There are limited data, however, describing interventions to improve timely antibiotic administration in immunocompromised, hospitalized pediatric patients.19
The aim of our multidisciplinary initiative was to decrease the mean time to administration of antibiotics in patients admitted to the BMT unit with first fever to <60 minutes. We captured data for all patients admitted to the BMT unit with a temperature of 38°C or higher. We defined BMT patients with first fever as patients with fever who were not currently on treatment doses of antibiotics. We did not include patients receiving antithymocyte globulin as part of their conditioning regimen because a common side effect of that treatment is fever and/or rigors. Before initiating any interventions, we examined data from April 2014 through February 2015. This provided our team with almost 1 year’s worth of data to analyze before initiating our first intervention in March 2015. Our baseline data revealed that <50% of patients with first fever from April 2014 through February 2015 received antibiotics within 60 minutes, with a mean time to antibiotics of 75 minutes (n = 140).
Human Subjects Protection
The present initiative fell within the Cincinnati Children’s Hospital Medical Center (CCHMC) Institutional Review Board’s guidance for QI projects that did not constitute human subjects research.
The BMT unit at CCHMC consists of a 36-bed, highly complex intensive care environment. The BMT program performs more than 100 transplants annually, most allogeneic. The BMT team consists of 14 attending physicians, 17 hematology/oncology and BMT/immunology fellows, 8 hospitalists, and 12 nurse practitioners. The clinical team consists of 112 registered nurses (RNs), 29 patient care assistants, and 8 highly-trained pharmacists. The BMT unit has 2 lead nurses, a charge nurse, and a safety coach, who help coordinate nursing activities on the unit, as well as aiding the bedside nurses and support staff. The hospitalists provide 24-hour coverage for inpatients, because residents do not rotate in the BMT unit.
In December of 2014, we established a multidisciplinary improvement team consisting of key stakeholders, including physicians, nurses, and pharmacists. The primary QI team consisted of 3 physicians, 1 RN, 2 clinical managers, 1 clinical director, and 1 clinical research coordinator. Before initiating interventions, the team reviewed the current antibiotic administration practice and process map, identifying potential areas for improvement from April 2014 through February 2015.
Planning the Intervention
Our improvement team developed a process map outlining the current system for antibiotic administration for febrile patients admitted to the BMT floor not currently on treatment doses of antibiotics (Fig 1). By using Pareto analysis and failure mode effects analysis,20 we found failures could be accounted for by a lack of nurse urgency, difficulty in contacting the ordering providers, and the placement of incorrect antibiotic orders. Delays after order placement included delays caused by a high number of statim (STAT) orders in the pharmacy queue, delays in the preparation of antibiotics caused by a lack of prioritization for patients with first fever, and delays in nurses picking up antibiotics from the medication room after they were sent from the pharmacy. Root cause analysis revealed an insufficient awareness among pharmacy, nursing, and physician staff of the new onset of fever and the need for timely antibiotics, the absence of a standardized antibiotic ordering process, and a lack of provider and nursing accountability for timely antibiotic delivery (Fig 1).
By using the Model for Improvement,21,22 the team first defined their global and specific, measurable, achievable, relevant, and time-bound aims and then developed a key driver diagram to represent their theory of improvement.20 For each driver, we identified potential interventions and their level of reliability and tested each by using plan-do-study-act (PDSA) cycles, which allowed us to test and refine our improvement strategies.20,21 Effective interventions were adopted, and new process maps were developed to represent our new system of care.
We developed a standardized process through small tests of change consisting of the following: (1) the creation of heightened awareness of first fever through automated text reminders during each shift; (2) the creation of a standardized notification of first fever to be delivered by the bedside nurse to the provider, charge nurse, and safety coach; (3) the dedication of a provider responsible for ordering antibiotics for patients with first fever; (4) the implementation of a standardized order set to be used specifically for first fever patients; (5) the creation of a pharmacy notification about newly febrile first fever patients to be delivered through a dedicated order; (6) the creation of a dedicated sticker stating “STAT first dose antibiotic, give directly to nurse” to be printed each time antibiotics were entered via the order set in the pharmacy; (7) a protocol for placement of the sticker on the antibiotics to assist with rapid identification and prioritization on arrival on the floor; and (8) the creation of a safety coach role to anticipate and help prioritize antibiotic delivery (Figs 2 and 3). We also used members of our team as staff caregiver champions to help provide reminders of the importance of timely antibiotic administration when there was a patient with a first fever. Our caregiver champions monitored the process from the time of documented fever to the time of antibiotic order placement, delivery, and administration. If any delays occurred, our champions spoke directly to the bedside nurse, as well as to the pharmacy staff, to obtain real-time feedback on the process. Signs were also placed in the provider workroom reminding them to use the dedicated first fever order set.
Evaluation and Analysis
The outcome measure was defined as the documented time between identification of a patient with a fever through vital sign flow sheets to the start of the antibiotic infusion, as documented by the RN. X-bar and s control charts were used to measure the mean time and SD of mean time, respectively. We defined “timely antibiotic administration” as antibiotics being given within 60 minutes from the time of fever, which was based on the definitions provided in previous studies.16,23 The percentage of patients receiving antibiotics within 60 minutes was managed through a p-chart. We measured 10 consecutive patients in the control charts secondary to the extreme variability in the number of new fevers each month. We identified first fever patients through automated notifications from the electronic medical record. Failures were assessed by the oversight team within 24 to 48 hours. This was achieved by analyzing the electronic medical record, as well as by contacting the ordering provider, the nurse administering the antibiotic, and the pharmacist to analyze what events occurred that may have resulted in delayed antibiotic administration. The patient name, the medical record number, the time of fever, the time at which antibiotics were ordered, the time the order was acknowledged by both the RN and the pharmacy, the time antibiotic was administered, and the reasons for delay were documented in an Excel spreadsheet.
Statistical process control methods were used to monitor changes in care processes. We established a mean, illustrated as the centerline on all control charts with upper and lower control limits. Standard industry criteria were used to determine if observed changes in measures were chance random variation (common cause variation) or caused by a specific assignable cause, in this case the intervention (special cause variation).24,25
The study period was from April of 2014 through March of 2017. During this time, a total of 650 BMT patients developed a first fever (Figs 4–6). Baseline data were obtained from April 2014 through February 2015. During that time, 40% of patients received antibiotics within 60 minutes, with a mean time to antibiotics of 75 minutes (41–146 minutes). Additionally, there was a high degree of variability in the time to antibiotics with an SD of ±75 minutes. In March of 2015, initial PDSA testing and implementation of interventions improved the mean time to administration of antibiotics to 52 minutes (SD ±19 minutes), and 75% of patients received antibiotics within 60 minutes. In November of 2015 the percentage of patients receiving timely antibiotics increased to 85%, with no change in the mean time to administration or SD. Finally, with sustained interventions and processes and no additional interventions, the percentage of patients receiving antibiotics within 60 minutes increased to 95%, with a mean time to antibiotics of 45 minutes (SD ±10 minutes), in December of 2016.
In March of 2016, we experienced significant delays in antibiotic administration outside of our control limits. Analysis revealed delays in antibiotic order placement by the health care providers. With reeducation of providers, the mean time to administration of antibiotics and the percentage of patients receiving timely antibiotics returned to levels within the control limits on the associated control charts.
We implemented and sustained interventions to improve timely antibiotic administration to BMT patients with first fever through small tests of change. After these changes were implemented, the percentage of patients receiving antibiotics within 60 minutes of documented fever increased from a mean of 40% to a mean of 95%. The mean time to antibiotics decreased to 45 minutes, an improvement that is likely to significantly impact outcomes.13,14 Moreover, the variability decreased significantly, from ±75 to 10 minutes.
To ensure the sustainability of our efforts, we created a standardized education module that was given to all nursing and pharmacy staff and provided interval refresher courses. The module included a summary of the literature that supported the administration of antibiotics within 60 minutes, our current process map, our key driver diagram, and an outline of the interventions to be implemented. Information regarding timely antibiotic administration was reviewed with current nursing staff before the start of each shift at the onset of this project. New nurses were provided with educational material in their orientation packets. Updated control charts were posted in the break room monthly, and results were included in the BMT unit dashboard for monthly analysis by leadership.
Identification of first fever by the nursing staff and timely communication with the providers and nursing staff increased awareness and cooperation among providers. The creation of a standardized order set and standardized first antibiotics for first fever decreased the variability in practice. Additionally, alerting the pharmacy through the first fever order set alongside the creation of the sticker to be placed on the antibiotics increased awareness and urgency. We attribute much of the success of our efforts to both the involvement of effective staff caregiver champions, who encouraged a culture shift from complacent routine to active engagement, and a deliberate plan to seek feedback from the front line to improve the process. Discussions of recently reported performance results were common during work hours, indicating a high level of staff member ownership of the initiative. Additional factors contributing to the success of the initiative were physical reminders of the initiative in the providers’ workroom and elsewhere throughout the department.
One limitation of this QI project is that we sequentially implemented the following changes to our process over the intervention period: involvement of the charge nurse and safety coach, development of a standardized order set, and increasing pharmacy awareness of orders for patients with their first fever. During the time of our interventions, the central pharmacy began a separate initiative to help improve processing of all STAT medications. Thus, it is difficult to discern which of these steps most helped to improve our median time to antibiotic administration or if the improvement was attributable to a combination of all of the elements. In addition, we were not able to identify a meaningful balancing measure to evaluate the impact of the intervention on our health care delivery system. Finally, because of the small number of patients, we were not able to evaluate the impact of the intervention on hospital length of stay or health care resources. We hope to evaluate these data in a future study.
When designing interventions to help improve our process, consideration was taken to avoid creating additional work for staff. Our goal was to create interventions that helped facilitate the process of ordering and administering antibiotics with small-scale changes that could easily be incorporated into the staff’s routine.
Although our health care delivery system maintains unique characteristics, we believe that the concepts and interventions are transferrable to other populations requiring complex care and for whom timely interventions by members of a multidisciplinary care team are also needed. For instance, the Children’s Hospital Association is currently engaged in a nationwide hospital collaborative to improve outcomes of pediatric patients with sepsis through early recognition and treatment.26 We believe that the interventions used in our project could be directly used in institutions participating in the collaborative, in that developing a standardized order set such as ours could facilitate the way providers order medical interventions, such as fluid boluses, laboratories, and antibiotics.
This report employed the SQUIRE 2.0 publication guidelines for reporting health care QI research.27 The authors thank the physicians, nurses, patient care assistants, nurse practitioners, hospitalists, fellows, and staff in the BMT unit at CCHMC and also thank, especially, the patients and their families.
- Accepted September 13, 2017.
- Address correspondence to Paulina Daniels, DO, Division of Bone Marrow Transplantation and Immune Deficiency, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, MLC 11027, Cincinnati, OH 45229. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Dandoy CE,
- Hausfeld J,
- Flesch L, et al
- Lane RD,
- Funai T,
- Reeder R,
- Larsen GY
- Bradshaw C,
- Goodman I,
- Rosenberg R,
- Bandera C,
- Fierman A,
- Rudy B
- Dandoy CE,
- Hariharan S,
- Weiss B, et al
- Berwick DM
- Langley G,
- Nolan K,
- Nolan T,
- Norman C,
- Provost L
- Benneyan JC,
- Lloyd RC,
- Plsek PE
- Children’s Hospital Association
- Ogrinc G,
- Davies L,
- Goodman D,
- Batalden P,
- Davidoff F,
- Stevens D
- Copyright © 2018 by the American Academy of Pediatrics