Reducing IV Antibiotic Duration for Neonatal UTI Using a Clinical Standard Pathway Free

This study was conducted at a tertiary, university-affiliated, freestanding children’s hospital in the Pacific Northwest. Our 364-bed hospital has an annual admission volume of ≥14 000 patients and ≥100 encounters for UTIs. Since 2002, the institution has developed and implemented >50 clinical standard pathways for common medical conditions⁹  and used pathways as part of quality improvement efforts.^8,10  Pathways are presented as flowcharts or algorithms and linked to the diagnosis-specific electronic order sets. Pathways are formally reviewed at least every 5 years to ensure they are consistent with the current medical literature and national guidelines. Reviews are conducted by individual pathway committees, which are led by 1 to 2 physicians, as well as multidisciplinary stakeholders. Once a review is complete and any changes verified, revisions are posted on the institutional pathway Web site and are immediately available to users. The UTI pathway committee is led by a pediatric hospitalist (author PC), as well as an emergency medicine physician, and comprises representatives from urology, radiology, nursing, and pharmacy.

Our primary intervention was the revision of our UTI clinical pathway. The pathway underwent a formal review starting in late 2019, which included a comprehensive literature review and meetings with the pathway committee, as well as the infectious disease division. For the literature review, the hospital’s medical librarian performed searches for all pediatric UTI articles in PubMed, Embase, and national guideline clearinghouses published in the previous 5 years. Two trained literature reviewers then summarized and appraised the quality of evidence as it pertained to each key recommendation in the clinical pathway. The pathway committee used these evidence summaries to decide on the pathway changes. If evidence was insufficient from the literature, recommendations were made on the basis of the expertise of the pathway committee members and infectious disease division.

A revised algorithm was released on February 20, 2020.¹¹  Changes included decreasing the recommended IV antibiotic duration from 7 days to a minimum of 3 days for infants 0 to 28 days of age with an Escherichia coli UTI (including bacteremic E coli UTI), transitioning to oral antibiotics after 3 days if afebrile and back to baseline for >24 hours, and identifying the sensitivities of the available uropathogen. For Staphylococcus aureus or Pseudomonas UTIs, the pathway recommends an infectious disease consultation for treatment recommendations, and for all other non-E coli UTIs, the pathway recommends considering an infectious disease consult to discuss IV antibiotic duration. The total antibiotic duration (IV plus oral) was unchanged at 14 days (Supplemental Fig 4).

The revised algorithm was posted on the hospital’s clinical pathway Web page^9,11  and linked in the electronic medical record (EMR) UTI order set. Notification of the clinical pathway revision was sent by e-mail to residents and hospital medicine attending physicians on the day of the revision go-live and announced on the home page of the hospital’s internal Web site (February 2020). Our additional interventions involved clinician education, rounding workflow, and attending feedback. We delivered training on the new pathway recommendations to the hospital medicine attending physicians during a faculty meeting in April 2020 and to the residents via noon conferences in October 2020 and again in November 2021. In October 2020, a new “transition IV to enteral meds (yes/no)” item was added to the rounding checklist, which is completed daily for each patient. Finally, during the intervention period (March 2020–March 2022), we monitored for UTI admissions and sent a real-time inquiry if a neonate with a UTI was hospitalized for >3 days and met all the criteria for switching to oral antibiotics (key driver diagram, Fig 1).

The primary outcome measure was IV antibiotic duration, defined as the number of days between administration times of the first and last doses. Intramuscular ceftriaxone was considered an IV antibiotic. The secondary outcomes were hospital LOS and cost per hospitalization. Hospital LOS was calculated on the basis of the admission and discharge times in the EMR. Costs were obtained from the hospital’s data warehouse and were Medical Care Consumer Price Index-adjusted to 2022 dollars.¹²  The balancing measures were readmission for UTI, bacteremia, or meningitis within 30 days of hospital discharge and all-cause readmission within 30 days of hospital discharge. Outcomes were assessed for 4 years before (baseline period January 2016 to February 2020) and 2 years after (intervention period March 2020 to March 2022) the pathway revision.

We performed a retrospective medical record review on neonates 0 to 28 days of age at the time of hospital admission with a primary discharge diagnosis of UTI, which was defined as having a positive urinalysis (pyuria and/or bacteriuria) and a positive urine culture with ≥10 000 colony forming units/mL of a uropathogen. Neonates who were admitted to the neonatal or pediatric ICU or with a hospital LOS of >30 days were excluded because these are unlikely to be otherwise healthy infants with a primary diagnosis of UTI. Although infants with genitourinary abnormalities are off-pathway during clinical care, we did not exclude infants with an International Classification of Diseases, Tenth Revision (ICD-10) code of a genitourinary abnormality because many of these abnormalities are likely to have been diagnosed after a neonatal UTI hospitalization. To ensure that all neonates who met inclusion criteria would be identified from the hospital’s EMR data warehouse, we initially identified neonates from the EMR using these 2 criteria: (1) ICD-10 diagnosis code of cystitis or pyelonephritis (N39.0, P39.3, N10, N30.00, N30.01, N30.80, N30.81, N30.90, N30.91) and (2) no aforementioned ICD-10 code but with a positive urinalysis result, positive urine culture result, or a temperature ≥38°C plus antibiotics prescribed at discharge. The lead author then reviewed 15% of the encounters identified using ICD-10 codes and 20% of encounters identified using the alternative method of random selection. All the neonates identified using ICD-10 codes met the inclusion criteria, and none of the neonates identified using alternative methods met the criteria. Thus, quarterly data extractions thereafter were based solely on ICD-10 codes. For any identified neonates with a hospital LOS of 15 to 29 days, a manual chart review was performed quarterly to ensure that they met the inclusion criteria.

We summarized subject characteristics by using descriptive statistics. We used statistical process control x-bar and S charts to examine average trend and variation in outcomes and balancing measures over time and determined special cause variation using the usual standards.¹³  Because pathway use is common at our institution and updated pathways are used in real-time, we elected not to include a washout period in the analyses. All analyses were performed by using R statistical computing software version 4.3.0 (R Core Team [2023]).¹⁴ 

The hospital’s institutional review board approved this study.

There were 93 neonates in the baseline period and 41 neonates in the intervention period who met the eligibility criteria. There were no statistically significant differences between the baseline and intervention groups in their age at admission, sex, or race. There were statistically significant differences in ethnicity and preferred language of care (Table 1). Consistent with previous studies,¹⁵  the majority of neonates were males (baseline period 84%, intervention period 88%), and most UTIs were due to E coli in both the baseline (84%) and intervention periods (88%). There were 10 neonates (10.8%) with bacteremic UTIs in the baseline period and 3 (7.3%) in the intervention period.

After the UTI pathway revision, we noted special cause variation with a significant decrease in IV antibiotic duration from a mean of 4.7 days in the baseline period to 3.1 days in the intervention period (Fig 2A). There was also a significant reduction in the hospital LOS from a baseline of 5.4 to 3.6 days (Fig 3A). There was no significant difference in cost between the baseline and intervention periods (Supplemental Fig 5A). We had special cause variation in the SDs of all 3 outcomes in quarter 4 of 2020 (Figs 2B and 3B; Supplemental Fig 5B). This was due to 1 neonate who was hospitalized with 14 days of IV antibiotics to treat a multidrug-resistant E coli UTI that did not have an oral antibiotic alternative. The neonate required infectious disease consultations, as well as a peripherally inserted central catheter by interventional radiology.

There were 7 neonates (7.5%) who were readmitted within 30 days of hospital discharge during the baseline period: 5 for recurrent UTI and 2 for non-UTI diagnoses. There were 0 readmissions during the intervention period.

This study reveals the impact of a clinical pathway on reducing IV antibiotic duration and hospital LOS for neonates with UTIs. After implementing new pathway recommendations, we successfully reduced the IV antibiotic duration and hospital LOS, and an overall shorter IV duration and LOS were sustained for 2 years after the change.

Before the pathway revision, there was a slight trend toward shorter IV antibiotic duration and a wide variation in the duration used, which was likely because the first key publication that supported shorter IV antibiotic treatment of neonatal UTI was published several years before the pathway revision. In addition, there has been national attention to increasing stewardship efforts around the timely transition of IV to enteral antibiotics.^16,17  However, without formal national guidelines, many clinicians may have been hesitant to apply these publications to their practice, whereas those who were ready to prescribe shorter IV durations may have felt bound by the unchanged local guidelines. Thus, once the pathway was revised, clinicians appeared to quickly adopt the new recommendations. This study highlights the importance of establishing local guidance as new evidence emerges that changes patient management. The authors of previous studies have estimated that it takes an average of 17 years for research evidence to translate into clinical practice.¹⁸  Clinical pathways are 1 strategy to improve evidence-based patient care, and local pathways have the potential to be nimbler in incorporating new evidence than national guidelines.

At our hospital, in which clinician workflow is pathway-driven for a variety of common health conditions, clinical pathways have improved care delivery. A pathway for acute gastroenteritis led to decreased IV fluid use and LOS in the emergency department,⁸  and a pathway for neonatal jaundice increased the efficiency of phototherapy initiation, decreased IV fluid use, and decreased LOS and charges.¹⁰  Although the context of successful pathway implementation at our hospital may limit generalizability, similar successes have been demonstrated at other hospitals, such as the implementation of a care process model that significantly decreased variation and increased the evidence-based care of febrile infants.¹⁹ 

Interestingly, although both the IV antibiotic duration and LOS significantly decreased after the pathway revision, hospital costs did not significantly decrease. This may be because the pathway revision occurred at the beginning of the coronavirus disease 2019 pandemic, which resulted in personnel and supply shortages that drove up costs.²⁰  Although we excluded patients who were in the ICU, medically complex patients on the medical floor may have incurred higher costs that affected the overall cost distribution, especially given our small sample size. Other possible positive patient-centered effects from the reduction in IV antibiotic days could be fewer peripheral IV lines replaced, the number of peripherally inserted central catheters placed, and the burden of longer hospitalizations for families, although we did not measure these outcomes in this study.

Despite the shorter IV antibiotic duration, there was no increase in readmissions within 30 days of hospital discharge, suggesting there was no increase in treatment failure for this patient population. This is consistent with the findings of the authors of previous publications who examined varying IV antibiotic durations for UTI in young infants. In a study that examined an administrative database, there was no difference in treatment failure rates between infants <2 months of age who received longer (≥4 days) versus shorter IV courses.¹  In a study of 172 infants <1 month of age with UTI, the median IV duration was 4 days (range 2–12 days), and no infant experienced a treatment failure or relapse.⁵  In addition, 3 retrospective studies of infants ≤2 months²¹  and <3 months^3,22  of age with bacteremic UTI found no association between IV antibiotic duration and recurrent or relapsed UTI. Studies specifically focused on the 0- to 28-day age group have been rare, and this study adds to the growing body of evidence that reveals that outcomes are excellent for neonates with UTIs who receive short IV antibiotic courses. Future quality improvement efforts should focus on total antibiotic duration.²³ 

This study has several limitations. First, this clinical pathway revision was implemented at a single freestanding children’s hospital that is accustomed to clinical pathways and, thus, may not be generalizable to other institutions. Although the implementation period was over 2 years, our patient population was still small, which may have limited our ability to capture readmissions as a balancing measure, and we could not assess if infants were readmitted to another hospital. In addition, we do not yet know if the use of shorter IV treatment will be sustained in the long term. Finally, we do not know if there were any socioeconomic, language, racial, or ethnic disparities in the outcomes, which should be evaluated in future studies.

The implementation of a revised clinical pathway significantly reduced IV antibiotic treatment duration and hospital LOS for neonatal UTIs without an increase in hospital readmissions.