Optimization of Antibiotic Prophylaxis Delivery for Pediatric Surgical Procedures

C.S. Mott Children’s Hospital is a 255-bed pediatric tertiary care center within the larger Michigan Medicine health system. Approximately 19 000 surgical procedures for pediatric patients are performed annually. The PIPS Committee is tasked with developing and maintaining structured and collaborative efforts to improve care for children with surgical needs in alignment with the institution’s quality and safety strategy. The PIPS Committee began this project in July 2015 as one of its initial improvement efforts. We report results of this effort through December 2019 in this article. Of note, our existing antibiotic selection guidelines were based on the 2013 ASHP guidelines and did not change significantly over the study period (Supplemental Information).

Root cause analysis, with identification of a lack of reliable systems for ordering, preparing, and delivering antibiotics to the operating room (OR), was performed to understand discrepancies in antibiotic administration across surgical services, as summarized in a fishbone diagram (Fig 1). An interdisciplinary team of surgical faculty, anesthesiologists, pediatricians in antimicrobial stewardship and hospital quality leadership roles, nurses, pharmacists, and health information technology staff performed a structured needs assessment using A3 problem-solving, which is a validated process improvement schematic that has been demonstrated as an effective form of structured problem-solving in health care and incorporates the plan-do-study-act cycle.^12,13 

To understand barriers to timely administration, we created a current state process map of perioperative antibiotic delivery, as shown in Fig 2. On the basis of this workflow diagram, the most common barriers to timely perioperative antibiotic administration included (1) failure to order antibiotics in the electronic health record (EHR) before the surgical date and (2) lack of availability of the desired antibiotic in the OR at the time of administration. Although we also identified antibiotic selection as an area for improvement, the team decided to focus first on addressing timely administration, while intentionally developing a process and framework that would support future work to optimize antibiotic selection.

The team conducted several plan-do-study-act cycles to sequentially target the most common reasons for failure by iteratively adjusting workflow processes.¹² 

Through the A3 process schematic, the interdisciplinary task force identified varied processes for ordering and administering preoperative antibiotics among surgical services as well as inconsistent indications for and choices of preoperative antibiotics among providers. Different definitions of appropriate timing for prophylactic antibiotic administration were also noted across surgical services.

The health system surgical antimicrobial prophylaxis guidelines were therefore reassessed by the institutional ASP. Although ASHP guidelines allow for most antibiotic infusions to begin 0 to 60 minutes before incision, emerging data suggest that the optimal start time for infusion may be closer to 15 to 30 minutes before incision for particular procedures.¹⁴  Furthermore, data clearly reveal that prophylaxis administered after incision is less effective.^15,16  Given pharmacokinetic theory that maximal tissue concentration at the time of incision requires starting antibiotic infusion before incision, the ASP revised guidelines to specify that antibiotic infusions should be started 15 to 60 minutes before incision. Antibiotics (eg, vancomycin) requiring longer infusion times should be initiated 60 to 120 minutes before incision (further revised to 45–90 minutes in September 2019).

Surgical providers expressed reluctance to ordering antibiotics at the time of case booking because of concerns that interval weight gain would result in underdosing of antibiotic prophylaxis. To address this concern, the team publicized and encouraged the use of weight-based antibiotic dosing ranges, which had been previously developed (Supplemental Information). These ranges limit the impact of interval weight gain because most patients do not outgrow their dose range before the procedure. Additionally, patients at the top of each range receive standard weight-based doses; thus, the few patients who advance out of the range are minimally underdosed. As part of an educational session, analysis of retrospective data was also shared, illustrating that changes in patient weight between case booking and the procedure would not lead to inappropriate dosing.

To integrate guideline-based antibiotic selection into provider workflow at the time of case request placement, a hard stop requiring an appropriate preoperative antibiotic order was developed and integrated into the EHR case request order set. This step was important to address early in this QI initiative because additional interventions depended on the antibiotic being ordered before the surgery date. Input from each service line regarding provider workflow was solicited in the development and rollout of order sets, and targeted education regarding the order set was provided through direct communication to surgical provider teams. Choice of antibiotics offered was based on evidence-based, procedure-related SSI prevention and current weight-based dosing ranges, as stipulated by the recently updated institutional guidelines. An option for “no antibiotics are indicated” was included to enable intentional appropriate declination of antibiotic prophylaxis while still fulfilling the hard-stop requirement. Additionally, if a patient’s weight was not previously recorded in the chart, it could be entered directly into the antibiotic order. Services were required to use these order sets to book cases rather than individual case request orders to ensure antibiotic decision-making occurred before the day of surgery. Feedback on use of the case request form was reviewed regularly by the PIPS Committee with surgical division chiefs.

To ensure that antibiotics ordered through the case request form were available in the OR for appropriate administration, additional staff were engaged to ensure process measures were developed, implemented, and audited. Specifically, a team consisting of preoperative charge nurses, supervisors, the quality manager, and a dedicated OR pharmacist designed a process for validation and review of each patient with an antibiotic order placed in advance to ensure the appropriate dose was prepared, delivered, and available at the bedside before patient transport to the OR. Within the EHR, we added a specific icon for antibiotics in the preoperative medication column visible on the daily master OR schedule. Each morning, the preoperative charge nurse reviewed the master daily schedule to ensure that antibiotic orders were prepared by the pharmacy and delivered to the preoperative holding space. Gaps identified through this review were shared at the daily morning perioperative services huddle so that specific plans could be made immediately to reconcile and obtain the appropriate antibiotic before the patient went back to the OR suite. This ongoing dialogue between the perioperative huddle and pharmacy staff also allowed for identification of gaps in the current process and optimization strategies for the future. Preoperative nurses were also integral in verifying orders and confirming that the correct drug and dose was available at the bedside.

To address the lack of OR team situational awareness of antibiotic administration timing, automatic whiteboard antibiotic timers linked to the anesthesia electronic documentation system were introduced in the ORs. This visual cue timer displays red text until 15 minutes has lapsed after documentation of first antibiotic administration, at which point the timer text changes to green to indicate that the incision can be made. For antibiotics with a longer required preincision time, the timer adjusts appropriately, and alerts for intraoperative redosing are displayed when required.

All patients ≤17 years undergoing a surgical procedure from July 2015 to December 2019 were eligible for inclusion. Excluded cases included (1) those in which antibiotics were not indicated, (2) those in which patients were already receiving antibiotics either as inpatients or transfers from the emergency department, (3) those in which there was a plan to administer antibiotics after cultures were drawn in the OR, and (4) emergent and urgent cases because the new workflow depended on advanced placement of antibiotic orders.

Our primary outcome measure was the proportion of surgical procedures for pediatric patients during which timely delivery of surgical antibiotic prophylaxis occurred. As a secondary outcome measure, SSIs were identified by the institution’s infection prevention team using their standard real-time processes, which incorporate diagnostic criteria established by the Centers for Disease Control and Prevention National Healthcare Safety Network. Process measures included rates of (1) ordering antibiotics preoperatively, (2) pharmacy verification and filling of antibiotics preoperatively, and (3) placement of antibiotics at the bedside by the preoperative nurse.

Statistical process control methodology was used to assess the impact of interventions on the outcome and process measures. Multiple instances of special cause variation were identified, and control chart limits and center lines were reset according to standard statistical process control rules for health care. Differences between the proportion of cases with appropriately timed prophylaxis in the pre- and postintervention periods were compared by using χ² tests.

This QI project received exemption by the institutional review board.

Of 45 401 surgical procedures performed during the study period, 32 192 (71%) were included. There were no significant differences in the types and numbers of operations in each surgical category performed per month over the study period.

The resulting process for perioperative antibiotic delivery and administration after implementation of these interventions is shown in Fig 3.

As demonstrated in the Fig 4 p-chart, the proportion of pediatric cases in which administration of surgical antibiotic prophylaxis occurred in the appropriate time frame increased from 38.6% per month at baseline to 94.0% at the conclusion of rollout of the sequential interventions (P < .001).

Baseline SSIs occurred at a rate of 0.8% (64 of 8331) per year. The postintervention SSI rate in 2019 was statistically unchanged at 0.7% (35 of 5327) per year. Of note, a shift in hospital reporting metrics occurred in 2018, and surveillance of high-risk and high-volume procedures was prioritized, resulting in fewer surgical cases surveyed for SSIs. Institutional SSI rates are shown in Supplemental Fig 5.

Rates of adoption of process measures are shown in Table 1. There was a significant increase in the number of appropriate preoperative antibiotics that were ordered before the day of surgery (P < .001) over the study period. Data limitations include that components of the hospital staff engagement bundle, such as pharmacy workflow optimization and preoperative antibiotic placement at the bedside by nursing, were not routinely measured before the study intervention and were obtained via a convenience sample during the intervention period. However, this intervention bundle was successfully adopted, with a 94% completion rate for eligible cases per month.

We successfully standardized timing of appropriate preoperative antibiotic prophylaxis across a large tertiary care pediatric operative program, increasing adherence from 38.6% to 94.0% by the end of the study period. We believe the success of this QI effort is due in large part to effective implementation of interventions informed by human factors concepts and strong partnerships with institutional antimicrobial stewardship efforts. Through the A3 process schematic, the interdisciplinary task force identified the variable sequence of events leading to inconsistent choice and timing of prophylactic antibiotic delivery. This led to the design of 4 corrective interventions: (1) updating institutional surgical antimicrobial prophylaxis guidelines, (2) implementing a hard-stop antibiotic order within EHR case requests, (3) engaging pharmacy and OR and/or PACU nursing staff, and (4) using an electronic antibiotic whiteboard timer. This QI approach may be helpful for other institutions that may similarly lack standardization of the process of preoperative antibiotic ordering and administration and are seeking to champion a focus on QI at the institutional level.

Previous retrospective analyses have revealed that compliance with antibiotic prophylaxis protocols may reduce risk of SSI for specific surgical indications.^10,17,18  However, in these analyses, the investigators do not specifically study means of achieving institutional compliance across multiple surgical services and procedure types, which was the primary objective of this study. The sustained improved outcomes across different surgical services, with variable case mix and patient populations, lend credibility to this QI program’s generalizability.

There is increasing recognition in human factors literature of the different levels of team and group behaviors, such as perception of responsibility and coordination and communication, that can contribute to care delivery. With these interventions, we sought to optimize system design to ensure appropriate and timely antibiotic delivery independent of individual provider variation across service lines in the health system.¹⁹  Iterative continuous performance improvement identified the opportunity for further engagement of hospital staff, such as pharmacy and nursing staff, in antibiotic retrieval and delivery. For example, we believe that one critical part of our success was task integration into existing provider workflow. We changed the choice architecture of a case request order set by adding a hard stop for antibiotic choice, which was an opt-out–forcing function as opposed to the previous opt-in decision-making workflow. This implementation and ongoing feedback to provider groups about compliance with the case request order set were essential because without an order for antibiotics before the date of surgery, the other interventions would have been less effective.

Although timely antibiotic administration has been linked to reductions in SSIs, we did not see a reduction in our SSI rate. Because most antibiotics were already given before incision, although not within the recommended time frame, this change in timing may not have been sufficient to decrease SSI rates. Conversely, the impact of our intervention may have been diminished by other factors that affect SSIs incidence^3–6  as well as the change in hospital SSI surveillance, which through exclusion of lower-risk procedures, might have artificially increased postintervention SSI incidence. Furthermore, because our interventions were performed across our pediatric institution’s entire surgical population, the expected impact based on studies in limited higher-risk populations may have been diluted.^10,17,18 

Although we did not observe a decrease in SSIs, this new process offers significant return on investment, with multiple safety benefits to our patients, by reducing the potential for medication errors. Before implementing this process, we learned of instances of dosing errors and inappropriate administration of antibiotics to patients with allergies when antibiotics were drawn up in the OR. The implemented standardized process for ordering antibiotics in advance of surgery now ensures that safety checks, such as dosing verification and cross-checking of allergies and medication interactions, by the EHR, pharmacist, and preoperative nurse reliably occur. We also learned that in the baseline state, only cefazolin was stocked in the ORs so cases were sometimes delayed while stat antibiotic orders were being prepared by pharmacy staff, and on occasion, attending anesthesiologists would need to leave the OR around the time of induction to retrieve antibiotics from the pharmacy. We also learned that because attending surgeons were not present for the preinduction time-out to confirm antibiotic recommendations, surgical residents would sometimes make incorrect preoperative antibiotic decisions; anecdotally, the most common error was to recommend antibiotics when they were not necessary. Standardization of EHR order sets with the inclusion of antibiotic clinical decision support addresses these gaps and has also provided a framework for future efforts to further optimize antibiotic selection and stewardship. Ultimately, this project illustrated that longitudinal investments in QI efforts at the institutional level have potential long-term benefits much broader than those achieved by any single QI project.

As further validation of the impact of this project, this work is currently being replicated by adult surgical service lines in our health system leveraging these new workflows and EHR case requests. Further investigation regarding appropriateness of antibiotic prophylaxis is indicated, as well as assessment of additional interventions to reduce risk of SSI in the pediatric population. To sustain success in timely prophylactic antibiotic administration, the interdisciplinary team continues to participate in bimonthly conferences to share data and lessons learned.

By nature, statistical process control methodology precludes detection of unmeasured confounders or evaluation of individual interventions in isolation. Given the visibility of the program across the institution, staff behaviors beyond the interventions may have contributed to overall improved timely antibiotic delivery. These particular interventions may not be generalizable to other institutions performing children’s surgery; however, the approach we used may inform similar efforts at other sites.

We demonstrated the utility of the A3 problem-solving schematic to successfully optimize prophylactic antibiotic timing and delivery in the surgical setting for pediatric patients by implementing 4 systems-based interventions.

We thank Brian Briggs; Matt Claysen; Timur Dubovoy, MD; Steve Harrington; Nathan Kirkpatrick; and Kelly Schultz, RN, MS, for developing and implementing the antibiotic timer.

ASHP

American Society of Health-System Pharmacists

ASP

antimicrobial stewardship program

EHR

electronic health record

OR

operating room

PIPS

performance improvement and patient safety

QI

quality improvement

SSI

surgical site infection