Quality Improvement Initiative to Improve Timing of Enteral Feeds in Pediatric Acute Pancreatitis

The project took place at Boston Children’s Hospital’s main campus, a tertiary care academic center that has an annual volume of 60 000 emergency department (ED) visits and 25 000 inpatient admissions per year. The ED and inpatient services are staffed by attending physicians, fellows, residents, and advanced practice providers. Patients with AP are admitted to either the gastroenterology or the general pediatrics service with gastroenterology consultation depending on patient censuses of both services.

We designed a QI intervention to standardize and improve the care of children with AP. This project was conducted according to local institutional standards for QI initiatives and, therefore, institutional review board approval was not obtained.

To address our aim of standardizing and improving the care of pediatric patients hospitalized with AP, a multidisciplinary team came together and created a key driver diagram to identify the primary drivers of provider variability in the management of AP (Fig 1). We then targeted these key drivers with 3 interventions: (1) the development of a clinical care pathway, (2) the development of a complementary electronic order set, and (3) provider education. Our QI interventions were implemented beginning January 1, 2017, and patients were reviewed from January 2013 through September 2021 (including pre- and postintervention). A comprehensive description of each intervention is detailed below.

An AP clinical care pathway was created by a group of pediatric pancreatologists along with the department of pediatrics QI program on the basis of a review of the literature and expert opinion. The clinical pathway was presented to key stakeholders, including representatives from multiple departments: gastroenterology, nutrition, emergency medicine, radiology, surgery, critical care medicine, pediatrics, and nursing. Feedback and review from stakeholders resulted in several revisions of the clinical pathway.

The clinical care pathway was designed to standardize the management of patients hospitalized with AP, providing recommendations for initial admission triage (ie, general floor vs ICU), fluid resuscitation, and nutritional support (Fig 2). Inclusion criteria for the pathway, based on the diagnostic criteria from INSPPIRE (International Study group of Pediatric Pancreatitis: In Search of a ure) consortium,¹³  includes patients with 2 of 3 of the following: (1) abdominal pain +/− emesis, (2) serum amylase and/or lipase 3 times greater than the upper limit of normal, and (3) imaging suggestive of AP. For patients who meet the criteria for AP, the first part of the pathway highlights recommendations for initial fluid resuscitation and identification of those at increased risk of severe AP and who require admission to an ICU (Fig 2A). Next, the pathway provided recommendations for nutritional support, including the initiation of enteral feeds within 48 hours of admission, clinical expectations as patients are advancing feeds, and a reminder to involve dietitians in these patients’ care (Fig 2B).

The pathway was distributed throughout the hospital on January 1, 2017 and included a link on our internal hospital homepage and printed versions posted in provider workrooms. Starting in January 2018, the pathway was also made available to all providers through a hospital-sponsored clinical pathway mobile app. Additionally, beginning in August 2018, the pathway could be accessed via a nested link within our AP order set.

Before our QI initiative, there were multiple order sets specific to the management of AP available to ED and inpatient providers. We reviewed the existing order sets, modifying and updating these to standardize the ordering process for patients with AP. Through collaboration with clinical informatics, we developed an electronic health record-based AP order set that complemented and aligned with our clinical care pathway. This included both an ED and inpatient AP order set. Major updates included initial laboratory tests to be ordered for all patients with AP as well as options for additional tests as clinically relevant (including blood gas analysis, lactate, and blood/urine cultures), intravenous fluids (IVF) with instructions for rate calculations and recommendations to give the first fluid bolus within 90 minutes, standard dosing for nonopioid analgesics (such as acetaminophen, ibuprofen, and/or ketorolac), and orders to consult appropriate services (including gastroenterology and nutrition for all patients with AP). Additionally, the updated order sets included reminders for providers to refer to the pancreatitis clinical pathway and an embedded link to the pathway for reference. The modified order sets were implemented in August 2018.

To increase provider awareness of recommendations for the management of pediatric pancreatitis, an education campaign for hospital providers was developed. Education of pediatric emergency providers occurred at divisional meetings during the clinical pathway development, which included a review of the literature and standard of care practices. We also presented at Gastroenterology Divisional Grand Rounds in August 2017 and November 2020, reviewing management guidelines and introducing the clinical care pathway and electronic order sets. Additionally, we created educational posters separate from the clinical care pathway that were placed throughout the hospital in July 2019, which highlighted key aspects of pancreatitis management and included a reminder for providers to use the clinical care pathway.

Medical records were reviewed starting in January 2013, identifying eligible patients with an admission diagnosis of AP. Using an automated search of our electronic health record data warehouse, we identified all patients with an inpatient discharge diagnosis of AP. We then performed a manual review of those medical records, identifying all patients <21 years old admitted for AP. Children with ≥1 hospitalization with a primary admission diagnosis of AP were included for each eligible hospital stay. AP was defined as any patient with 2 of the 3 criteria¹³ : (1) symptoms consistent with AP, (2) elevated serum amylase and/or lipase ≥3 times the upper limit of normal, and/or (3) imaging suggestive of AP. We excluded children with AP transferred from an outside institution and children with an admission diagnosis other than AP but who developed AP while hospitalized.

Data were abstracted from the medical records of eligible patients, including patient demographics, time to initial fluid bolus, time to first enteral feed, and hospital LOS. Initial data obtained from January 2013 until December 2016 served as our baseline measure. After our clinical care pathway launch in January 2017, data were reviewed on a quarterly basis in an effort to monitor the impact of our interventions.¹⁴ 

Our primary outcome measure was the percentage of patients receiving enteral nutrition within 48 hours of admission, measured as the time from initial ED triage (or “time of admission”) to the time of first enteral feed. Our secondary outcome measures included median hospital LOS and time to first IVF bolus. Hospital LOS was the measured time (in days) from time of admission to time of discharge, and time to first IVF bolus was defined as the time (in minutes) from initial ED triage time to time of first IVF bolus administration. We choose to measure time to first IVF bolus given the evidence that early IVF boluses improve outcomes in AP.^15–17 

Our primary balancing measure was readmission rate to ensure that the initiation of early enteral nutrition and reduced LOS did not result in an increase in hospital readmissions. This was tracked by measuring the frequency of admissions within 7 days of discharge for each patient encounter. Additional balancing measures include admission and transfer to the ICU.

Statistical process control charts were used to analyze changes in all outcome and balancing measures over time. The percentage of patients receiving early enteral nutrition, readmission rate, and ICU admission and transfer rates were analyzed by using p-charts with control limits set as 3 standard deviations.¹⁸  Hospital LOS and time to first fluid bolus were analyzed by using a run chart. A run chart (rather than x-bar and S chart) was used to account for extreme outliers because LOS and time to first bolus were not normally distributed. Significant changes were recognized by shifts in the centerline during the project period by using the established 8-point rule for identifying special cause variation.¹⁸  The baseline project period was from January 2013 to December 2016, and the postintervention period was from January 2017 to September 2021. SQCPack version 7.0 (PQ Systems, Dayton, OH) was used to create the p-charts and median charts. We also compared patient characteristics before and after the implementation of the intervention. Comparisons of categorical and numeric data were assessed by using the X² test and Mann-Whitney test, respectively.

During the 93-month project period, we identified 652 eligible patients with an admission diagnosis of AP. Of these patients, 322 (49%) were included in the postimplementation period. Patients were similar in both study periods (Table 1).

First, we evaluated the time to initiation of enteral feeds. During the baseline period, the percentage of patients receiving enteral nutrition within 48 hours of admission was 40%. Statistical process control p-chart analysis revealed special cause variation with a significant increase in the mean to 84%. Initially, we saw an increase in patients receiving early enteral nutrition to 58% after the implementation of our clinical pathway, which further increased to 84% after additional interventions, including the distribution of our educational posters (Fig 3). This improvement remained stable throughout the project period.

We then analyzed our secondary outcome measures, including hospital LOS and time to first IVF bolus. For median hospital LOS, there was a significant shift from a median LOS of 5.5 days (baseline period) to 4 days (postimplementation period) on the run chart (Fig 4). This improvement began shortly after the implementation of the clinical pathway and was sustained throughout the QI initiative. We found that time to first IVF bolus did not statistically change within the study period, with a median of 120 minutes (Supplemental Fig 6).

The balancing measures were additionally analyzed. Despite the decrease in hospital LOS, hospital readmission rates did not change between the pre- and postintervention project periods (Fig 5). We saw no change in the percentage of patients requiring admission to the ICU (Supplemental Fig 7A). However, we did see a significant decrease in patients requiring ICU transfers from 5% to 1% after multiple interventions (Supplemental Fig 7B).

Through the implementation of our QI initiative, we were able to improve and standardize the care of pediatric patients with AP. In this project, we were able to increase early enteral nutrition rates and reduce hospitalization time by introducing a high-quality, evidence-based clinical care pathway with an integrated electronic order set and direct education of providers. By utilizing core QI principles, we implemented multiple interventions that changed our hospital’s culture around enteral nutrition in AP. Specifically, we found that the percentage of patients receiving early enteral nutrition rose to 84%, which was sustained over a 2-year period and surpassed our initial goal of 60%. Additionally, these improvements were associated with shorter hospitalization times but not increased hospital readmission rates.

We feel that we were successful in achieving our initial goal of increased early enteral nutrition rates because of our multiple interventions, including the development of our clinical care pathway, integration of an electronic order set, and direct provider education. We felt that, by employing a multifaceted approach, we could ensure our project’s success, an approach that has previously been shown to be effective.^19–21  We found that there was an immediate improvement in early nutrition rates shortly after the implementation of our clinical care pathway, adding to the growing evidence that clinical pathways help standardize care and improve outcomes for common conditions.²²  However, we found that this initial improvement in early nutrition rates was rather modest and noticed a more significant improvement in Q2 2019 (>1 year after our pathway launch date), at the time that our educational posters were distributed (Fig 3). We also noticed a significant decrease in ICU transfer rates around this same time, as well (Supplemental Fig 7B). This likely reflects the cumulative effects of our multiple interventions over those initial 12 months of the project, including the widespread distribution of the pathway through multiple avenues and our interventions targeting provider education. Thus, highlighting a cultural shift in the approach to AP management, as well as increased adherence to our clinical pathway during that time frame.

Over the past several decades, there has been a paradigm shift with regard to enteral nutrition and AP, specifically the importance of early enteral nutrition in patients with AP.^{5,7–9,11,23}  Despite these recommendations, there continues to be significant variability and slow uptake in actual clinical practice, with some studies suggesting that providers are reluctant to initiate early enteral feeds in patients with AP because they do not want to worsen pain or prolong hospitalization times.^12,24  This is reflected in the present project because, initially, only 40% of patients received early enteral nutrition. Given this gap between evidence-based care and clinical practice, we created our clinical pathway to standardize care and reduce variation, as well as provide a framework for management for providers less familiar with AP. To address potential concerns from providers, we incorporated evidence-based information into our clinical care pathway and educated providers directly on the safety of early enteral nutrition. We also ensured additional supports for providers were in place, including the involvement of our gastroenterology consultation service (for patients admitted to other services), as well as dietitians.

AP also represents a significant financial burden within the pediatric population. Recent data estimates that the average hospitalization cost for pediatric AP is approximately $2500 per day.^1,4  Although we did not perform a cost-effectiveness analysis in this present study, the estimated cost savings of our intervention (based on our 1.5-day reduction in LOS and average daily hospitalization costs for AP),⁴  is conservatively approximated to be $3750 per patient. This could equate to annual cost savings of approximately $300c000 within our own institution, based on our average number of AP admissions each year. However, future studies incorporating cost measurements are necessary to determine actual health care cost-effectiveness.

An area of future focus is time to first fluid bolus. Fluid resuscitation has been known to be critical in the treatment of AP, helping to limit organ hypoperfusion and counterbalancing the interstitial fluid losses that occur in AP.^16,17,25,26  Despite our success at increasing early enteral nutrition, we found no difference in the timing to first fluid bolus with our interventions. Because the timing for initiation of fluids depends on many different factors, including ED staffing and census, laboratory delays, and delays in alerting providers of laboratory results, there are reasons why our initiative did not yield any significant improvements in this measure. Future steps to address these barriers could include integrating automated notifications in patients’ records who have an elevated serum amylase and/or lipase, alerting providers that these results could indicate a diagnosis of AP, and recommendations for initiating IVF.

There are several limitations to our study. First, this study took place in a single institution, so the results may not be generalizable. Second, because this was a retrospective review, we were unable to determine provider reasoning for the timing of initiation of feeds (ie, clinical status of patient or patient preference). Lastly, because of limitations in our patient sample size each month, we choose to analyze our outcome measures on a quarterly basis rather than monthly, which may have limited our ability to visualize which interventions made the biggest difference.

A multidisciplinary QI initiative was effective at improving and standardizing the care of pediatric patients with AP without increasing hospital LOS. Through the implementation of multiple QI interventions, including developing a clinical care pathway, as well as an integrated electronic order set and direct provider education, we were able to successfully increase the rates of early enteral nutrition and decrease hospitalization times without increasing readmission rates, improvements which have been sustained over a 3-year period.

We thank the members of the BCH Department of Pediatrics quality team, as well as the BCH Clinical Pathway team for their efforts with this project. We thank Victor Fox, Menno Verhave, and Sue Hue for their contribution in the development of the clinical pathway and implementation of this project. We also thank Taruna Banerjee for her help in the establishment of the clinical care pathway during its early development.

AP

acute pancreatitis

ED

emergency department

IVF

intravenous fluids

LOS

length of stay

QI

quality improvement