Substantial variability exists in hyperbilirubinemia screening and monitoring leading to unnecessary total serum bilirubin (TSB) testing in healthy newborns. We aimed to assess the impact of value-care interventions to decrease the monthly TSB testing rate per 100 patient-days among healthy newborns in our Mother-Baby Unit by 30% by June 2022.
We formed a multidisciplinary team to review the current practice for ordering TSB among housestaff in our Mother-Baby Unit. We identified several themes: variation in clinical practice, fear of hyperbilirubinemia, and desire to act for high-intermediate risk bilirubin levels. The interventions consisted of obtaining faculty buy-in, redesigning the hyperbilirubinemia pathway, educating staff on high value-care, producing an instructional video, and prompting staff to incorporate a bilirubin risk assessment via smart phrases in our electronic health record. The primary outcome was the monthly TSB testing rate per 100 patient-days. Universal predischarge bilirubin screening, length of stay, phototherapy rates, and readmission rates were chosen as balancing measures.
The monthly rate of TSB testing was reduced from 51 to 26.3 TSB per 100 patient-days, representing a 48% reduction. This improvement was sustained for 12 months. The percentage of infants with at least 1 TSB measurement during birth hospitalization decreased from 48% to 30%. Predischarge bilirubin screening, length of stay, and readmission rates were unchanged.
Our quality improvement initiative led to a significant reduction in the monthly TSB testing per 100 patient-days in healthy newborns without evidence of harm.
Neonatal jaundice affects up to 80% of full-term and late preterm newborns.1,2 Although neonatal jaundice is a transient diagnosis with no long-term sequelae for most infants, the extremely rare3,4 but potential development of kernicterus has led to routine bilirubin testing. Since the publication of the 2004 American Academy of Pediatrics (AAP) hyperbilirubinemia guideline,1 new evidence suggests that even significantly elevated bilirubin levels are less harmful than previously thought, with a rate of chronic bilirubin encephalopathy of 0.8 per 100 000 infants.4,5
Although our center used the AAP clinical practice guideline to form an institutional pathway that incorporated both clinical risk assessment and at least 1-time bilirubin screening before discharge,1 it was noted that our institutional pathway called for verifying transcutaneous bilirubin (TcB) levels with total serum bilirubin (TSB) levels at thresholds far lower than the phototherapy nomograms proposed by the 2004 AAP guideline; this is a phenomenon now also addressed by the updated 2022 AAP hyperbilirubinemia guidelines.6 Additionally, our institutional pathway used rapid rate of rise (with serial TSB draws) as an independent indication for initiation of phototherapy. Last, there was substantial practice variability among clinicians regarding screening and treatment of hyperbilirubinemia at our institution. Taken together, these factors led to overuse of TSB draws at our institution.
Decreasing unnecessary blood draws and preventing overuse is 1 of the pillars of providing high-value care because it minimizes waste and prevents overdiagnosis and overtreatment.7,8 The overuse of TSBs when not clinically indicated can lead to increased hospital cost, length of stay (LOS),9 parental anxiety, and infant discomfort.
TcB is a noninvasive procedure that accurately estimates TSB,10–13 and its implementation in revised clinical pathways for the management of infants with neonatal jaundice have been demonstrated to be effective tools in minimizing blood sampling and the likelihood of having a TSB level ≥20 mg/dL.14–17 Thus, we sought to revise our institutional clinical pathway to decrease the monthly TSB testing rate per 100 patient-days among healthy newborns in our Mother-Baby Unit (MBU) by 30% by June 2022 while providing high-value care without compromising outcomes such as universal predischarge bilirubin screening, LOS, phototherapy rates, and hospital readmission.
Methods
Context
This project took place in the MBU of the Women’s Hospital at Jackson Health System Holtz Children’s Hospital, a section of our nonprofit academic medical system in Miami, Florida. The MBU has 2 clinically equivalent well-baby wards staffed by pediatric hospitalists, pediatric interns, medical students, advanced practice providers, nurses, social workers, and lactation specialists. Pediatric hospitalists and pediatric interns drive TSB testing decisions in these resident-run units. Well babies are kept for approximately 24 to 48 hours for vaginal deliveries and 48 to 72 hours for cesarean section deliveries, and outpatient follow-up is scheduled before discharge based on the bilirubin risk assessment.
Planning the Interventions
At the start of the project, a multidisciplinary team was formed in October 2020 to review the current practice for ordering blood work among housestaff. This team included attending physicians, residents, nursing management, and nurses in MBU. Fishbone (Fig 1) analysis was performed to identify factors contributing to the increased number of TSBs. We identified several themes when assessing barriers: variation in clinical practice, fear of hyperbilirubinemia, low to moderate level of evidence in the 2004 AAP guideline, desire to act for high-intermediate risk bilirubin levels, and suboptimal communication between physicians and nurses. Baseline data from July through December 2020 helped us explore the scope of the problem, refine our measures, and create a key driver diagram (Fig 2) to guide our process improvement.
Key driver diagram used to achieve the goal of reducing unnecessary TSB testing in our Mother-Baby units. SMART, specific, measurable, achievable, realistic, and time-bound; TSB, total serum bilirubin test.
Key driver diagram used to achieve the goal of reducing unnecessary TSB testing in our Mother-Baby units. SMART, specific, measurable, achievable, realistic, and time-bound; TSB, total serum bilirubin test.
Interventions
Using the Institute for Healthcare Improvement’s Model, we designed and executed plan-do-study-act (PDSA) cycles18 between January 2021 to January 2022, including (1) obtaining faculty buy-in, (2) revamping the hyperbilirubinemia clinical practice pathway, (3) providing education on high-value care among medical staff during multidisciplinary meetings, (4) devising a smart phrase to include in the resident’s daily progress note to standardized hyperbilirubinemia assessment and management, and (5) creating an instructional video presentation of the revised hyperbilirubinemia pathway for the housestaff on the MBU.
Faculty Buy-In and the Hyperbilirubinemia Clinical Pathway
A Holtz institution-specific pathway for monitoring and managing bilirubin levels in neonates was previously accessible to all pediatric residents through a central hospital application on smart phones or via web access. According to the previous pathway, all neonates underwent TcB at 12 hours of life, 24 hours of life, and before discharge. Each time, the TcB was plotted using the online calculator BiliTool. If the level fell into high-intermediate, high-risk, or phototherapy levels, the next step in the algorithm was to draw a TSB level and assess risk factors. After obtaining faculty buy-in through discussions with our pediatric hospitalists and review of the evidence,4,5 the decision was made to increase the threshold to obtain a TSB only when TcB reached the high-risk zone or crossed phototherapy levels per BiliTool. This allowed for maintenance of universal predischarge bilirubin screening, but with less prompts to draw serum level verifications. Another improvement to the clinical pathway was to start phototherapy only if the TSB level was at or above the threshold to start phototherapy7,13–16 allowing only prophylactic phototherapy (within 1 mg/dL) on the day of hospital discharge. The previous pathway also used rate of rise calculations to trigger phototherapy, which was not continued in the new pathway. Last, thresholds for end-tidal carbon monoxide measurements were raised in the new clinical pathway and used only on day of discharge if indicated. Criteria for NICU admission was highlighted on the new pathway as well (Fig 3).
Clinical pathway for the screening and management of hyperbilirubinemia during birth hospitalization. TcB, transcutaneous bilirubin level; TSB, total serum bilirubin.
Clinical pathway for the screening and management of hyperbilirubinemia during birth hospitalization. TcB, transcutaneous bilirubin level; TSB, total serum bilirubin.
High Value of Care
Biweekly meetings with housestaff and nursing staff were conducted to increase their understanding of overtesting, overdiagnosis, and overtreatment leading to increase health care cost and harms in our pediatric patients.19–21 These sessions allowed medical staff to understand the importance of the rationale for ordering TSB by carefully thinking about why they are ordering the test and if it would provide value to their patient care.
Smart Phrase
A smart phrase was created in the electronic health record to standardize hyperbilirubinemia assessment and management in the daily progress note. This smart phrase was introduced and distributed to oncoming residents at the start of their rotation. This served to facilitate clear communication between medical team members on the inpatient teams regarding the phototherapy risk status of the newborn based on their neurotoxicity risk factors and gestational age. Subsequently, it provided a step-by-step approach for interns to delineate the most recent TcB or TSB levels and categorize those levels based on their risk zone in BiliTool, as well as document its relation to that newborn’s phototherapy initiation level per BiliTool. Last, it provided space for interns to check the new hyperbilirubinemia management pathway to determine when the next level measurement was indicated.
The Hyperbilirubinemia Screening Pathway Instructional Video
The instructional video was started in October 2021 with the primary intended audience being the interns assigned to rotate in the MBU. The 8-minute video served as a step- by-step manual of the pathway from the time of birth of the newborn to the day of discharge from the MBU. During the instructional video, specific case-based examples were reviewed as well. This video was then uploaded to a mobile application accessible to all our pediatric interns and residents to view before their rotation.
Measures and Study of the Interventions
Outcome Measures
We defined the primary outcome measure as the monthly TSB testing rate per 100 patient-days in our MBU; we calculated this measure monthly by dividing the total number of TSB tests by the product of the number of newborns and mean length of stay and then multiplying by a factor of 100. The monthly TSB rate was used because it took into consideration the mean length of stay and census over time. The secondary outcome measure was the monthly percentage of infants receiving at least 1 TSB during birth hospitalization.
Balancing Measures
To ensure we were not causing unintended consequences with our interventions, we measured and tracked universal bilirubin screening, LOS, phototherapy rates during birth hospitalization, and hospital readmissions for phototherapy treatment. Universal bilirubin screening was calculated as the monthly percentage of infants with at least 1 measurement of TcB or TSB before hospital discharge.
Analysis
To assess and understand the impact of the interventions in the process, we used statistical process control charts. We identified special cause variation and shifted the mean centerline when 8 or more consecutive data points were above or below the centerline (shift) after an intervention.22,23 We used U-chart22 to display TSB rates per 100 patient-days, and P-chart for discrete binomial data. We also used the Fisher test and Wilcoxon rank-sum to calculate P values to determine statistical significance between the baseline and intervention period using SAS version 9.4 (SAS Institute, Inc, Cary, North Carolina).
Ethical Considerations
The present initiative fell within the Jackson Memorial Hospital/University of Miami institutional review board’s guidance for quality improvement projects that did not constitute human subject research. We did not obtain external funding.
Results
From July 2020 to January 2022, 3152 newborns were admitted to our MBU. Of these, 802 infants were admitted during the baseline period from July 2020 to December 2020 and 2350 infants were admitted during the intervention period from January 2021 to January 2022.
For the primary outcome measure, the baseline rate of TSB testing was 51 TSB tests per 100 patient-days. A centerline shift in TSB testing was appreciated in February 2021, immediately after PDSA cycle 1, to 26.3 TSB tests per 100 patient-days, representing a 48% reduction from baseline. No changes in the mean centerline were observed for PDSA cycle 2 or 3. However, we were able to sustain this result over a 12-month period (Fig 4).
Annotated control chart (U-chart) depicting rate of TSB performed per 100 patient-days by month from July 2020 through January 2022. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Annotated control chart (U-chart) depicting rate of TSB performed per 100 patient-days by month from July 2020 through January 2022. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
The monthly mean percentage of infants having at least 1 TSB testing from July 2020 to May 2021 was 48%, and after the implementation of several interventions, the mean centerline shifted in June 2021 to 30% (Fig 5). Additionally, comparison between the baseline and study period was found to be statistically significant.
Annotated control chart (P-chart) depicting the monthly percentage of infants receiving at least 1 TSB during birth hospitalization. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Annotated control chart (P-chart) depicting the monthly percentage of infants receiving at least 1 TSB during birth hospitalization. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
For our balancing measures, 100% of infants from the baseline period and 2349 of 2350 infants from the intervention period had at least 1 measure of TcB or TSB during birth hospitalization in accordance with the 2009 hyperbilirubinemia recommendation update for universal predischarge bilirubin screening.24 The only infant who did not have a bilirubin screening had anencephaly and did not receive screening per parental wishes as part of end-of-life care.
There was no difference in the mean LOS between the 2 study periods. The mean LOS was 2.13 days for the baseline and the intervention period. The monthly mean percentage of phototherapy treatment during birth hospitalization from July 2020 to May 2021 was 8%, and after the implementation of several interventions, it decreased to 3.6% (Fig 6). Additionally, comparison between the baseline and study period was found to be statistically significant.
Annotated control chart (P-chart) depicting percentage of infants receiving phototherapy treatment during birth hospitalization. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Annotated control chart (P-chart) depicting percentage of infants receiving phototherapy treatment during birth hospitalization. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Between July 2020 and December 2021, there were 52 readmissions for phototherapy treatment after birth hospitalization. Of these, 12 (1.49%) were from baseline period and 40 from January 2021 to December 2021 (1.76%). Although we saw an increase in the mean percentage, it was not statistically significant (P = .61). Additionally, we did not find special cause variation in our statistical process control chart shown in Fig 7.
Annotated control chart (P-chart) depicting percentage of readmission for phototherapy among infants discharged from the Mother-Baby Units. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Annotated control chart (P-chart) depicting percentage of readmission for phototherapy among infants discharged from the Mother-Baby Units. The centerline reflects the mean, and the upper and lower control limits reflect 3 SDs above and below the mean. LCL, lower control limit; PDSA, Plan-Do-Study-Act; UCL, upper control limit.
Discussion
We describe a single-site process improvement focused on reducing the rate of unnecessary TSB tests per 100 patient-days in our MBU through a combination of interventions, such as obtaining faculty buy-in, redesigning the hyperbilirubinemia clinical pathway, educating on high-value care among medical staff and nurses, creating an instructional video detailing how to navigate the hyperbilirubinemia pathway, and prompting medical staff to use an easy-to-use bilirubin risk smart phrase in our electronic health record system. We achieved a significant reduction in the TSB testing rate per 100 patient-days in our MBU over the study period. Importantly, universal predischarge bilirubin screening was not compromised, no significant increase in readmission for phototherapy was seen, and there was no increase in LOS or phototherapy rates during birth hospitalization. Overall, these findings are echoed in existing literature on the utility of standardizing hyperbilirubinemia pathways to decrease variation in clinical practice, unnecessary cost, and harm to newborns without significant rise in readmission for phototherapy.14–17 The decline in TSB rate per 100 patient-days appears to be driven both by having more infants without any TSB tests and by decreasing repeat TSB tests for those infants who had at least 1 TSB test during the birth hospitalization.
The most impactful interventions leading to a steep downtrend in the TSB testing rate per 100 patient-days were faculty buy-in, the development of a new standardized clinical pathway for hyperbilirubinemia. Faculty buy-in is often considered vital for creation and maintenance of change in academic hospital centers25–28 as the teachings by attending physicians model residents’ clinical practice. We achieved faculty buy-in by creating awareness of the scope of the problem, discussing the impact of variation in clinical practice, showing data on the high correlation between TcB and TSB in healthy newborns.10–13 Thus, this allowed us to create our new clinical pathway focused on prompting TSB only if TcB was in the high-risk zone or at phototherapy threshold level, without having to draw multiple serum levels to calculate rate of rise.
Our emphasis on educating medical staff on high-value care during multidisciplinary meetings and usage of smart phrases to standardize hyperbilirubinemia assessment and management in the daily progress note likely contributed to the sustained reduction in TSB testing rates. The high-value care examples provided during interdisciplinary meeting helped us modify individual TSB-ordering behaviors by creating awareness that overtesting was associated to overtreatment and both are more likely to produce harm than benefits.7,19,20 Published studies in other areas such as radiation exposure or type of laboratory testing have also described that their most significant interventions were those related to education of the housestaff in high-value care and feedback on laboratory ordering practices.17,25,27–30
We did not observe unintended adverse consequences or kernicterus cases throughout the study period. This is consistent with multiple clinical studies and meta-analysis that report TcBs are a reliable measure when compared with TSBs.10–13
Additionally, phototherapy rates decreased because of more judicious use of TSB testing by understanding that the use of phototherapy below published AAP treatment thresholds and the overtesting were more likely to produce harm than benefits.7,9,20,28 Although we observed an increase in readmissions for phototherapy, this was not statistically significant and within the range for the overall increase in national trends describing phototherapy readmission rates doubled during the COVID-19 pandemic.31
Our study has several limitations. First, the aim of this project was to decrease unnecessary TSB testing, but there are no validated tools to determine the necessity of TSB tests without a very labor-intensive chart review and the confounder of assessing retrospectively. Knowing that unnecessary TSB testing was a problem at our institution, we assumed a decrease in the total TSB tests performed without any adverse consequences was representative of less unnecessary TSB testing. In the future, a better assessment of necessity could evaluate TSBs obtained after TcBs and determine if it was indicated based on established guidelines for TSB testing. Second, our data currently track the readmission rates for phototherapy to our own hospital center only; thus, there is a risk of higher rates of readmissions existing for other hospital centers. This can be expanded in future studies by tracking readmission rates to other hospital centers by phone follow-ups with families after discharge. Third, our study does not yet evaluate patient families and personnel satisfaction with the interventions, but on informal feedback through discussions either with families or at multidisciplinary meetings, there was a reported positive response to the changes implemented. This should be further quantified by formal evaluation in the coming months both for families and personnel. An additional limitation is that this is a single-center study; thus, a concern for generalizability can be raised. However, given the high volume of patients assessed at our academic hospital center, the length of time the interventions were tracked and the applicability of these interventions to be integrated to pediatric residency programs, we suspect that our interventions would be impactful in similar academic institutions.
Last, after the conclusion of this project, the AAP published new 2022 clinical practice guidelines on hyperbilirubinemia to replace the previously established 2004 AAP guidelines. Of note, this new guideline affirms that TcB is a reliable, sensitive, and convenient screening test to identify infants who require TSB measurements, and that TSB should be the definite test for guiding phototherapy and escalation of care.6,32 Both principles were also used in the revised hyperbilirubinemia pathway implemented by our project. The new guidelines also removed risk zones and raised the thresholds for phototherapy based on gestational age and neurotoxicity risk factors, relying on evidence that bilirubin neurotoxicity occurs at levels higher than previously used in the 2004 guidelines.1,6 Our project pathway similarly raised the threshold at which point a TSB should be used to confirm TcB based on this same evidence; however, we used obtaining TSB in high-risk zones instead of within 3 points of phototherapy, and use of risk zones was eliminated in the 2022 guidelines.6 Thus, our future efforts will focus on aligning our institutional pathway with the current 2022 recommendations. Overall, we assess that the adjustments made in the 2022 guidelines are in line with our quality improvement initiative to improve high-value care by reducing hyperbilirubinemia overtreatment and its associated harms.6,33
Conclusions
Our quality improvement initiative was able to implement value improvement interventions leading to a reduction in the monthly TSB testing per 100 patient-days in healthy newborns without adversely affecting our universal predischarge bilirubin screening, LOS, rate of phototherapy treatment, or readmission for phototherapy treatment. The most successful interventions were the implementation of the revised hyperbilirubinemia clinical pathway and high value of care coaching. We believe this quality improvement initiative benefited patients, families, and staff and can be reproduced in other centers.
Dr Bayes has conceptualized and designed the study, led the quality improvement interventions, collected, and reviewed data, performed all analyses, drafted the initial manuscript and figures, and reviewed and revised the manuscript; Drs Garcia-Chacon, Lorusso, Sukkar, Jananeh, Harrison, Martin, Grabois, Castillo Ayarza, Raj, Cohen, and Pal have contributed to the design of the study, implementation of key interventions, reviewed data, drafted the initial manuscript, and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have potential conflicts of interest to disclose.
Comments