Comparing Outcomes Between Direct and ED Admissions for Neonatal Hyperbilirubinemia Free

We conducted a single-center, retrospective study of newborns admitted for the management of neonatal hyperbilirubinemia at a large, academic, freestanding children’s hospital between January 1, 2017, and December 31, 2021. This study was approved by the institutional review board of the hospital.

Infants between the age of 24 hours and 14 days who were admitted with a primary diagnosis of neonatal hyperbilirubinemia were included in the study. The International Classification of Diseases, 10th Revision, Clinical Modification, codes included were as follows: Neonatal jaundice from breast milk inhibitor (P59.3); Neonatal jaundice, unspecified (P59.9); Rh isoimmunization of newborn (P55.0); ABO isoimmunization of newborn (P55.1); Other hemolytic disease of the newborn (P55.8); Hemolytic disease of newborn, unspecified (P55.9); Neonatal jaundice due to other specified excessive hemolysis (P58.8); Neonatal jaundice from other specified causes (P59.8); and Disorder of bilirubin metabolism, unspecified (E80.7).

We excluded patients with a gestational age less than 35 weeks because these patients follow a separate hyperbilirubinemia treatment guideline.¹⁸  Other exclusion criteria included patients with a primary or secondary diagnosis of sepsis or sepsis evaluation, bacteremia, urinary tract infection, biliary obstruction (eg, biliary atresia, choledochal cysts), a metabolic disorder (eg, hypothyroidism, galactosemia), or major congenital anomalies. We did not include birth hospitalization encounters. Patients were also excluded if there were concerns regarding data integrity owing to missing electronic health record (EHR) documentation (Supplemental Fig 4).

From the data set, based on inclusion and exclusion criteria, we performed additional chart reviews on patients who did not receive phototherapy during the hospitalization. We did this to ensure that patients were not admitted unnecessarily with below phototherapy initiation bilirubin levels. There were a total of 3 patients (1 ED, 2 DA) who were admitted for hyperbilirubinemia that did not receive phototherapy given their below threshold bilirubin level. One of these patients (ED) was admitted over concern for a rapid rate of rise. Another (DA) was admitted for concern over abnormal behavior in addition to elevated bilirubin levels. The final patient (DA) had a repeat bilirubin level on reaching the floor that was no longer above phototherapy threshold. All 3 patients were excluded from the analysis.

Demographic information, including age, sex, race and ethnicity, language preference, and primary payer, was extracted from the EHR for each patient. Race and ethnicity data were collected based on standard operating procedures to assess whether inequities existed in rates of DA.²¹  “Other” was included as a race and ethnicity category and comprised non-Hispanic American Indian and non-Hispanic Native Hawaiian (because of small sample sizes), patients for whom the guardian was unavailable, or was listed as unknown. The language preference category of “Other” represented patients with preferred languages that had small sample sizes (Supplemental Table 4). Patients were also characterized by their initial bilirubin level, gestational age, and feeding method (breastmilk, formula, or both). Initial bilirubin level was defined as the first level obtained after arrival to the hospital (Supplemental Table 5).

Primary outcome variables included LOS and time to clinical care (Supplemental Table 5). Time to clinical care encompassed both time to initiation of phototherapy and time to first bilirubin level collection. Secondary outcomes included the total number of bilirubin levels obtained, as well as the following binary variables: IV fluid use (maintenance or bolus fluids), laboratory testing (complete blood count [CBC], reticulocyte count, blood type and direct antiglobulin test [DAT], albumin, glucose-6-phosphate dehydrogenase [G6PD], and other), and lactation consultations. We also compared NICU transfers and 7-day readmissions for phototherapy between the 2 groups.

Descriptive statistics were used to analyze the patient characteristics for each admission group of interest. Pearson’s χ² test, Fisher exact test, and Wilcoxon rank-sum test were used to explore associations of patient variables with the admission group. We then calculated the adjusted estimates of admission group differences using linear regression models for 2 primary outcomes: LOS and time to initiation of phototherapy. Each model was adjusted for age at admission, sex, race and ethnicity, gestational age, initial bilirubin level, DAT result (positive, negative, or not obtained), and primary payer. We used a likelihood ratio test to assess the modification of effect between direct and ED admissions with the variables “DAT result” and “initial bilirubin level.” Because the LOS and time to initiation of phototherapy were right-skewed, both outcomes were fitted using their log-transformed values. For each model, the predicted marginal means for the admission groups were calculated and summarized with 95% confidence intervals. Marginal means were back-transformed onto the original unit scale of the outcome, and thus may be interpreted as the predicted geometric mean duration of hours. Statistical significance was assessed at α = 0.05. All data preparation and analyses were performed using R version 4.3.0.²² 

A total of 1098 neonates met the inclusion criteria, including 276 admitted from the ED (25.1%) and 822 directly admitted to the hospital (74.9%).

There were no significant differences between the 2 groups with regard to sex or gestational age; however, there was a significant difference in the age of the patient on admission (Table 1). Additionally, there were no differences in race and ethnicity or language preference between direct and ED admissions. However, there was a statistically significant difference in the primary payor, with a higher percentage of patients insured by Medicaid being admitted through the ED and patients with commercial insurance being admitted directly to the hospital (Table 1).

The median LOS for patients who were directly admitted to the hospital was 2 hours shorter than that of patients admitted from the ED (P < .001), despite no significant difference in the initial total serum bilirubin level (P = .91), gestational age (P = .87), DAT result (P = .53), or albumin level (P = .29). Patients who were directly admitted to the hospital also had a shorter median time to initiation of phototherapy by 3.4 hours (P < .001) and a shorter median time to first bilirubin level collection by 0.25 hours or 15 minutes (P = .003) (Table 2).

Adjusted analyses showed similar differences between the admission groups when controlling for factors inherent to the patient on arrival to the hospital. Patients directly admitted to the hospital had a shorter LOS compared with those admitted from the ED (geometric mean: 26.2 vs 28.7 hours; P < .001, respectively) (Fig 1). Likewise, the time to phototherapy initiation was shorter for patients directly admitted than for those admitted from the ED (geometric mean: 1.0 vs 4.4 hours; P < .001, respectively) (Fig 2).

There were significant differences in resource utilization between the 2 groups. First, patients admitted via the ED were more likely to receive IV fluids than those admitted directly (51.4% and 8.9%; P < .001, respectively). Patients directly admitted to the hospital were more likely to have a CBC, reticulocyte count, blood type, and DAT obtained, whereas patients admitted through the ED were more likely to have an albumin level obtained. There was no difference between the groups regarding the percentage of G6PD screening tests or quantitative levels obtained. Additionally, patients admitted directly to the hospital had fewer total serum bilirubin levels drawn and fewer hemolyzed or “insufficient quantity” results when compared with ED admissions (Table 3). When comparing laboratory results between the groups, there were no statistically significant differences except for hemoglobin level (g/dL), which was slightly lower in the ED group (median, 18.1 g/dL versus 18.6 g/dL; P = .009) (Table 3).

The route of admission was not associated with a significant difference in the percentage of patients who received a lactation consultation, despite a higher number of patients who were directly admitted receiving breastmilk in some capacity (direct breastfeeding, expressed breastmilk, and/or breastfeeding with formula supplementation).

There were a total of 16 patients (1.4%) who required readmission for phototherapy within 7 days of discharge, with no difference between patients directly admitted to the hospital and those admitted through the ED (P = .90).

Only 1 patient, a DA, required transfer to the NICU after arrival to the floor (1/822 = 0.12%). This patient had a 7.5 mg/dL increase in their total bilirubin level in the 6.5 hours between their outpatient pediatrician visit and the initial laboratory draw on admission to the floor, equating to an uncommonly high rate of rise of greater than 1 mg/dL/h. This patient did not require exchange transfusion as the bilirubin level decreased with phototherapy and IV fluids, and they were ultimately transferred back to the general floor the next day.

This is the first study to compare direct and ED admissions of patients with a primary diagnosis of neonatal hyperbilirubinemia. In this single-center, retrospective study, we demonstrated that directly admitting patients for the management of neonatal hyperbilirubinemia is a preferred alternative to ED admission as our study showed that DAs had a shorter time to clinical care, shorter LOS, and less unnecessary resource utilization with no difference in 7-day readmissions for phototherapy.

In addition to having a shorter time to clinical care, patients directly admitted to the hospital had fewer bilirubin levels drawn and peripheral IVs placed. They also had fewer laboratory draws that were unusable because of hemolysis or insufficient quantities. This decreased the need for additional phlebotomy attempts, which can be very distressing for both patients and caregivers. Yet, our study also demonstrated that DAs were more likely to have a CBC and blood type/DAT obtained. Outside of obtaining a reticulocyte count, this was management consistent with the 2004 AAP CPG, which was the most current version at the time of this study.²³  Similarly, there was a significant difference in IV fluid use between the groups, with a larger number of ED patients receiving IV fluids. Because we did not obtain information on hydration status or weight loss from birth, it was difficult to assess whether IV fluid use was necessary. However, this could demonstrate a potential area in which easily accessible, locally instituted pathways could help to provide high-value care. Interestingly, there was no difference in patients who received a lactation consultation despite more patients who were directly admitted receiving breastmilk in some capacity.

Directly admitting pediatric patients has been shown to be a safe alternative to admitting patients via the ED.^7,11,24  Our study further supports these data because only 1 patient who was directly admitted required transfer to the NICU, and this was due to findings specific to this patient. Additionally, despite a shorter LOS for patients directly admitted to the hospital, there was no significant difference in 7-day readmissions for phototherapy between the 2 groups. DAs can improve ED overcrowding by decreasing the ED patient load, which could improve hospital throughput of patients.²⁵  Bypassing the ED can also limit infectious exposures for this high-risk population. Our study builds on previous literature, demonstrating that directly admitting patients to the hospital, instead of using the ED, is a safe and effective mode of admission.^7,8  Directly admitting patients to the hospital improves resource utilization and decreases unnecessary health care spending by eliminating an unnecessary layer of care for the majority of patients with neonatal hyperbilirubinemia.

Previous DA research has shown differences with respect to race and ethnicity when comparing direct and ED admissions, which our findings did not demonstrate.⁷  One hypothesis for these previously documented differences is that the majority of pediatric DAs originate from their primary care provider and there are well-documented racial and ethnic disparities in children’s access to primary care^26,27 ; thus, certain race and ethnicities may have decreased access to the DA process. However, for neonatal hyperbilirubinemia specifically, the effects of unequal access may be at least partially mitigated by the common birth hospital recommendation that a newborn follow-up appointment be scheduled before discharge. If birth hospitals ensure linkage to primary care in this manner, a neonate could be more likely to attend a newborn primary care visit, where most neonatal hyperbilirubinemia is diagnosed.²⁸  A second potential explanation for the lack of racial and ethnic disparities in our specific study is that management of neonatal hyperbilirubinemia is largely driven by national CPGs. Guidelines that protocolize the decision regarding the need for treatment, and therefore admission, could potentially remove some implicit biases from making the decision to directly admit a patient. Interestingly, our study had similar findings to those of previous DA research with regard to primary payor.^7,8  This may demonstrate that there are other variables at play contributing to this difference, other than the common hypothesis that children with public insurance are less likely to have a medical home.^7,8 

This study has several limitations that should be considered. First, this was a single-center, retrospective study that may not be generalizable to all hospitals or health care systems. At our institution, we have a well-established, streamlined process for DAs including a physician call center and 24/7 physician presence within the hospital to ensure timely assessment of DAs. At other institutions that may not have these formalized processes, time to clinical care may be faster in the ED. Institutions can consider using the recent AAP policy statement recommendations regarding DAs to adopt a process specific to the capabilities and needs of their respective institutions to optimize safety and efficacy.¹¹  Second, this study solely evaluated DAs for the diagnosis of neonatal hyperbilirubinemia. Although DAs can be used for other diagnoses, clinicians should carefully consider the safety and appropriateness of bypassing the ED when implementing this alternative method of admission.¹¹  Third, calculation of the phototherapy threshold for all patients would be manually intensive and potentially impossible for patients without documentation of the presence or absence of neurotoxicity risk factors. As a proxy, we evaluated the differences in the median gestational age, initial bilirubin level, DAT result, and albumin level between the 2 groups. Despite the use of this measure, the study did not evaluate whether the initial ED or DA bilirubin levels met phototherapy initiation thresholds. However, for our institution’s DA process, a patient being directly admitted for neonatal hyperbilirubinemia would have a serum bilirubin level obtained as an outpatient and the decision to admit would be based on whether the level met the phototherapy threshold. Finally, our population was predominantly non-Hispanic White and had a preferred language of English which may also limit the generalizability of the study.

Future studies should compare outcomes between direct and ED admissions for other common pediatric diagnoses. Additionally, a dedicated study to understand the inequities found in this and prior DA studies could help create a foundation for subsequent work in minimizing or eliminating these disparities.

Directly admitting patients for the management of neonatal hyperbilirubinemia is a preferred alternative to ED admission as our study showed that DAs had a shorter time to clinical care, shorter LOS, and less unnecessary resource utilization with no difference in 7-day readmissions for phototherapy. Because the initiation of care in the ED does not appear to provide any clinical or economic benefit, there may be very few cases in which admission through the ED is necessary for the management of neonatal hyperbilirubinemia.