Implementation of a Clinical Decision Rule for Treatment of Neonatal Jaundice in the Emergency Department

Background The rate of hospital admissions for hyperbilirubinemia (jaundice) has increased
over the past 10 years. Jaundice is a condition caused by elevated serum levels of bilirubin
due to an increase in bilirubin production or normal production with problems eliminating it
from the body. Serum levels of bilirubin in excess of 5 mg/dL signify clinical jaundice; more
than half of full term infants experience these levels within the first week of life.1 It is
important for health care providers to distinguish infants experiencing physiologic jaundice
as part of otherwise healthy development from those experiencing pathologic jaundice which,
if not treated aggressively, can lead to irreversible damage of the central nervous system.2
Since 2000 there has been a 160% increase in the number of children readmitted to hospitals
for jaundice after discharge from the hospital nursery.2 This dramatic increase is believed
to be associated with the change in birth hospital culture as newborns are now typically
discharged 48 hours after delivery while bilirubin levels are still on the rise. Bilirubin
levels typically peak between the third and fifth day of life.3 The change in hospital
discharge procedures results in circumstances where parents may present for medical care in
an ED when their newborn starts to show symptoms of jaundice. This trend has been observed
despite the implementation of routine bilirubin screening in many newborn nurseries. Similar
to national trends, at Children's there has been a substantial increase in ED evaluations for
jaundice; in 2012, 274 neonates younger than 7 days old were treated for jaundice in
Children's EDs. Families may seek care in the ED as they may not have an established
relationship with a Pediatrician or they may believe that their child requires an
intervention that was only available in the ED setting. While some of these patients are
ultimately admitted to the hospital for monitoring and treatment others are sent home for
phototherapy. At Children's, approximately 61% of the newborns evaluated for jaundice were
admitted to the hospital for further care.4

In 2004 the American Academy of Pediatrics (AAP) implemented evidence-based recommendations
for the treatment of jaundice, that incorporate not only the infant's serum bilirubin
concentrations but also gestational age, postnatal age, symptoms and risk factors for
pathologic hyperbilirubinemia.5, 6 The AAP recommendations are the basis of all current
jaundice clinical algorithms. The main treatment options according to the AAP guidelines are
phototherapy, and in rare circumstances, exchange transfusions or pharmacologic treatments.2
Phototherapy is the most common treatment for jaundice and may be used in either a home-based
setting or after admission to a hospital.6 Phototherapy reduces serum bilirubin levels by
changing the shape and structure of bilirubin to one that is water soluble and can more
simplistically be eliminated from the body through urine and stool.1 For those patients who
fail phototherapy treatment or in whom the levels are exceedingly high, the use of exchange
transfusions is an option within the hospital setting. Phototherapy has been successfully and
safely utilized in the home-based setting to reduce the likelihood of hospital readmissions.6
Although pediatricians have successfully utilized home-based phototherapy, ED utilization of
this option has been limited by access to phototherapy equipment vendors and ED provider
knowledge of the criteria for outpatient management. As a result of this knowledge and access
gap, the current ED care for infants with jaundice results in variation in care and
over-hospitalization of infants who may be safely treated at home.

While the AAP guidelines educate physicians on the importance of evaluating infants who are 3
- 5 days of age, there is a lack of adherence to these recommendations and follow-up is often
delayed. As an example, Profit et al. evaluated a large group of urban and suburban
pediatricians to determine adherence with the AAP recommendations for follow-up of healthy
newborn infants ≥35 weeks gestational age. Only 37% of infants in the study were seen by a
pediatrician within the first 6 days of life.7 This finding is consistent with national
trends that show infants frequently present to EDs when symptoms of jaundice emerge in lieu
of the recommended screening.1 For example, in 2007, EDs across the United States evaluated
almost 46,000 patients for neonatal jaundice. In order to reduce time to phototherapy and
decrease ED length of stay8 when faced with the emergent presentation of jaundice some
hospitals have implemented clinical pathways. Gomez et al. developed and implemented such a
pathway through the use of a CDS tool to accurately diagnose and treat jaundice as well as to
mitigate the costs that are associated with readmissions to hospitals. The authors found that
the clinical pathway was useful for clarifying the decision process for clinicians by helping
to educate clinicians about treatment options for jaundice them gain while highlighting
inconsistencies in their routine protocols.9 Other authors have similarly reported that
clinical pathways can help to reduce irregularities and discrepancies in neonatal jaundice
care.10 Thus, the available literature would support that clinical pathways for neonatal
jaundice can be useful and can lead to meaningful improvements in patient care.

Key strategies for improving adherence to current jaundice clinical pathways are to increase
ED clinician knowledge and improve access outpatient phototherapy. An electronic,
Cerner-based CDS tool, such as the one we propose, would provide clinicians with information
at the point of care and would be incorporated into the ED workflow. This approach utilizes
algorithms developed from evidenced based guidelines to standardize the treatment of neonatal
jaundice and stratify patients into diagnostic risk groups based on particular aspects of the
history, physical examination and laboratory data. The goal is to standardize the care in
order to promote safe, cost-effective as well as appropriate care for all infants. This is
especially important in the ED, where there is a high volume of acutely ill patients for whom
management decisions must be made both accurately and quickly. This includes the rapid
recognition of who is at risk of more serious complications and conditions such as
kernicterus and who can safely be discharged home.3 Finally, the ordering, delivery, and
set-up of home phototherapy equipment must require minimal effort by the ED clinician. We
propose to leverage our electronic health record (EHR) system to streamline ordering and our
well established home nursing organization to conduct home nursing visits to address these
barriers to adherence with jaundice clinical pathways.

Research Question Will a clinical decision support tool for jaundice promote appropriate use
of home photo therapy and reduce hospital admissions for patients at risk for jaundice?

Design This is a prospective, interrupted time series trial.

Methods We will measure resource utilization (i.e. laboratory use, hospital admission)
retrospectively for 48 months pre-intervention in children identified at risk for jaundice
presenting for care in a Children's Emergency Departments. We will then implement the
jaundice-CDS over a 2-month period and measure resource utilization for 18 months post
intervention. Rates of unplanned ED or hospital visits, use of exchange-transfusion and
mortality will be monitored as safety outcomes.

Jaundice-CDS is composed of two components. The first, MD Alert, triggered at initial
provider login will cognitively prime the ED provider toward use of Jaundice-CDS. The second
consists of a pre-populated order-set which provides targeted recommendations and guidance on
use of photo-therapy and criteria for hospital admission. Both are based on the currently
accepted pediatric guideline for management of this condition.

MD ALERT:

A provider EHR-based alert will trigger for infants identified by triage who are meet
eligibility criteria. This alert will serve to remind the clinician of the existence of a
Jaundice order set, the availability of home nursing and home photo-therapy.

ORDER SET:

When the clinician begins the process of ordering any laboratory tests, they will be
encouraged to utilize the Jaundice order-set. The order-set will contain recommendation on
the most appropriate laboratory tests to obtain, criteria for admission to the hospital, as
well as pre-defined orders for obtaining a home nursing referral and delivery of home
photo-therapy equipment. A link will also be provided to "bilitool.org", an online resource
for calculating a neonates risk for jaundice.

Analysis We will use logistic regression to assess, after controlling for relevant clinical
factors, the odds of ordering a specific laboratory test (e.g., CBC, BMP, Coombs, etc.) in
the group of patients seen during implementation of the CDS compared to the odds of such an
order among patients arriving prior to the implementation of the CDS. Similar to the analysis
of laboratory tests, we will use logistic regression to estimate odds ratios for admission
and for home-based phototherapy among patients presenting to the ED with jaundice symptoms
before and after the implementation of the CDS. We will use linear regression to compare
total cost per encounter for all patients seen in the ED before the implementation of the CDS
to total costs for patients seen in the ED post-implementation. In addition to ED costs,
total costs of care will include any inpatient costs for patients admitted, and they will
include costs for all lab tests. We will also compare costs pre- and post-implementation for
those admitted to the hospital and for those discharged home.

Inclusion Criteria:

- Patients are birth to 7 days post-natal age

- Present in Minneapolis or St Paul ED during the pre-implementation or implementation
period with a chief complaint of jaundice.

Exclusion Criteria:

- Children born prematurely

- Concern for sepsis or bacterial infection

- In-born error of metabolism

- Metabolic disease

- Positive newborn screen (e.g. Sickle Cell Disease, Cystic Fibrosis)