“Help Me Sleep”: A Quality Initiative to Reduce Overnight Vital Signs

This quality improvement (QI) project occurred at a large, urban, academic children’s hospital on 3 PHM teaching services. Each team includes a PHM attending physician, pediatric and family medicine residents, and medical students. A PHM fellow rotates on the service periodically. At our institution, obtaining patient vital signs is primarily within patient care assistant (PCA) workflow. PCAs are responsible for providing basic supportive care to patients and families under the direct supervision of a registered nurse. For our 24-bed medical unit, 3 to 4 PCAs are present each shift and are assigned to 6 to 8 patients.

For this project, the target low-risk patient populations were identified through consensus opinion of our improvement team and institutional PHM leadership. The team chose patients admitted for neonatal indirect hyperbilirubinemia requiring phototherapy and failure to thrive, because these patient populations were deemed to be at low-risk for clinical deterioration and had consistent admission rates. The time frame of 12 am to 6 am was determined after collaborative discussions with nursing staff and family representatives through our institutional family advisory council. These groups believed this 6-hour period provided the most impactful time frame to abstain from vital sign measurements to promote rest. Before starting this QI initiative, standard inpatient hospitalization practice and diagnosis-specific admission order sets were defaulted to obtain vitals Q4H. At our institution, neonates admitted for hyperbilirubinemia are placed in isolettes while undergoing phototherapy and, per hospital policy, axillary temperatures are checked every 4 hours.

A multidisciplinary improvement team composed of a PHM fellow, PHM attending physician, QI service line coordinator, nursing clinical leader, and pediatric residents was formed in October 2019. To obtain patient and parent perspectives regarding overnight vital signs, the QI team collaborated with our institution’s family advisory council. Using the Institute for Healthcare Improvement Model for Improvement methodology, the team developed a key driver diagram (Fig 1) to identify drivers and potential interventions.¹⁹  The team used plan-do-study-act (PDSA) cycles to test hypotheses and statistical process control charts to review results on a monthly basis.

PDSA 1 started in March 2020 with project dissemination to key stakeholders. This involved a description of the project in the weekly resident newsletter and an E-mail communication to the resident team before starting the PHM rotation. Additional in-person presentations by a QI team member to on-service PHM attending physicians and residents occurred at the beginning of each 4-week rotation. These discussions centered on project rationale, current state of overnight vitals, and project goals. Directed education for nursing staff was completed via E-mail communication and monthly reminders at strategic partner staff meetings.

We also leveraged the electronic health record (EHR) (Epic, Verona, WI) to change ordering behavior. At our institution, customized order sets exist for patients admitted with the diagnosis of failure to thrive and hyperbilirubinemia. The vital sign order was preselected for vitals Q4H. The QI team modified admission order sets for our low-risk patient populations to require a provider to manually select a vital sign frequency on admission. In addition, the vital sign order was modified to include preset radio buttons for obtaining vitals only while awake (Fig 2). This order already existed within the EHR, making it an easily accessible option for conveying the project’s goal. These EHR modifications were intended to promote more deliberate critical thinking when ordering vital sign frequency.

On the basis of initial feedback, PDSA 2, starting in August 2020, focused on increasing provider awareness of the project and delivering targeted outreach to PCAs. Visual reminder placards were placed at resident and attending workstations to remind staff of the sleep-friendly initiative and patient inclusion criteria for the QI project. Manual chart reviews from PDSA 1 patients revealed Q4H while awake orders were being used without a change in vital sign frequency or timing. QI team members attended required PCA in-service education sessions to discuss the project, encourage careful review of vital sign orders, and identify any potential barriers.

PCAs felt that the language of the Q4H while awake vital sign order lacked necessary specifics given that both patient populations have irregular sleep cycles with nighttime awakenings and daytime naps. Therefore, PDSA 3, starting in November 2020, focused on additional refinement of vital sign orders to include specific language in the comments section about abstaining from vital signs between 12 am and 6 am if the Q4H or Q8H while awake order was selected. Comments also stated to contact the primary team for any clinical concerns. Standardized note templates for these low-risk patient populations were also updated to encourage documentation of the ordered vital sign frequency.

The primary outcome measure was the percentage of patient nights without a vital sign measurement between midnight and 6 am for those admitted for hyperbilirubinemia or failure to thrive. Q4H temperature checks were continued for patients with hyperbilirubinemia in accordance with hospital policy; therefore, the absence of all other vital signs during this period was counted as a success. Our process measure was the percentage of our low-risk patients who had the vital sign order indicating a vital sign frequency (ie, Q4 or Q8) while awake on a nightly basis. Patients were identified on the basis of their International Classification of Diseases, 10th Revision discharge codes.

Balancing measures were focused on adverse patient events, specifically code blue activations outside the ICU and emergent transfers. Emergent transfers are defined as the need for mechanical ventilation, vasoactive agents, or cardiopulmonary resuscitation within 60 minutes of arriving to the pediatric ICU.²⁰  In addition, given the rarity of these events in our low-risk patient populations, our team also tracked an institutional metric known as “watcher” status. The watchstander program is a situational awareness initiative to identify patients at risk for clinical deterioration to intervene before decompensation. We tracked whether any of our low-risk patients were designated a watcher at any point during their hospital stay, regardless of the vital sign order.

Statistical process control p-charts were used to analyze and display outcome and process measures. Published statistical process control rules and the aggregate point rule were used when determining special versus common cause variations.²¹  Data were taken from our institution’s EHR on a monthly basis and converted to a spreadsheet for analysis by the team quality improvement service line coordinator.

This quality improvement work was reviewed and deemed not to be human subjects research by the institutional review board.

Baseline data were obtained from January 2019 to February 2020. The study period ran from March 2020 through April 2021. During the 14-month implementation period, there were 449 low-risk admissions for a total of 1550 inpatient nights. Of these, 287 admissions were for hyperbilirubinemia with 467 inpatient nights, and 162 admissions were for failure to thrive with a total of 1083 inpatient nights. The study consisted of 171 (38%) female patients. The median age was 2.9 months, and the median length of stay was 1.1 days. For comparison, the baseline period consisted of 444 patients, with 195 (44%) female patients, a median age of 2.9 months, and a median length of stay of 1.2 days.

During our study period, the percentage of patient nights with vital sign checks decreased from 98% to 38% by January 2021 (Fig 3). Special cause variation was identified using the aggregate point rule. PDSA 1 led to a centerline shift from 98% to 83% starting in March 2020. A second centerline shift was seen after PDSA 2 to 60% starting in August 2020. A third shift occurred after PDSA 3 to 38%. In total, this translated to 708 patient nights during which routine vital sign checks were not obtained between 12 am and 6 am.

Our process measure, the use of the order “Q4H (or Q8H) while awake” increased from 0% to 31% in March with PDSA 1. Special cause variation was seen again in August to 87% and was sustained for the remainder of the project (Fig 4).

During our baseline and implementation periods, there were no emergent transfers or code blues outside the ICU in our patient populations. Of the 449 patients in the implementation period, 9 were identified as a watcher during their hospitalization. Two of these patients did not have vital signs obtained overnight. Reasons identified for increased staff concern included “gut feeling of parent or caregiver/team” (n = 6) and “unfamiliar/high risk treatment or diagnosis” (n = 4). One patient had both reasons identified. There were no adverse patient outcomes in any of the patients meeting watchstander criteria.

This quality improvement initiative successfully reduced the percentage of patient nights with a vital sign check between 12 am and 6 am from 98% to 38% for 2 low-risk pediatric populations. More importantly, this decreased monitoring did not result in any significant adverse patient safety events. Previous work dedicated to improving inpatient sleep has focused on vital sign order placement or the elimination of specific vital signs (ie, blood pressure monitoring) as a primary outcome metric.^3,7,16,22  To our knowledge, this is the first pediatric study with a primary outcome metric being the elimination of all aspects of an overnight vital sign check when possible, because patients undergoing phototherapy still received temperature checks. In addition, instead of focusing on specific clinical inclusion or exclusion criteria, our QI initiative targeted preidentified low-risk patient populations and relied on provider critical thinking to determine appropriateness for decreased overnight monitoring. With our project, we were able to minimize disruptions among patient populations that already have fragmented and unpredictable sleep patterns because of their young age.

Changing our institution’s culture around vital signs required a multifaced approach. Initial efforts focused mostly on EHR modifications and medical team education. Our most notable improvement, however, occurred when we expanded collaboration with nursing and PCA team members. Engaging with these stakeholders provided them the platform to share important feedback around the initial EHR changes and led to more clear and specific language in the vital sign orders. The subsequent EHR modification promoted a shared mental model among all providers, nurses, and PCAs for the previously ambiguous vital sign request of “while awake,” thereby leading to the vital sign order more consistently producing the desired effect. Additionally, bedside staff are likely better at identifying barriers to sleep and are more comfortable with foregoing routine bedside monitoring overnight.⁷  Our experience is mirrored in the adult literature, in which the empowerment of nursing staff led to a more robust reduction in nocturnal disruptions and improvement in sleep quality than with EHR modifications and physician education alone.¹⁶ 

Combining routine cares, such as vital sign checks at the time of feeds, offers an alternative opportunity to reduce patient disruptions. Initial chart reviews, however, identified vital signs documented at the prescribed frequencies. Given the sporadic timing of patient care needs, this finding suggested that the request to obtain vital signs while awake did not produce clustering of care. Furthermore, given the absence of meaningful information routine vital sign checks provide, we sought to promote increased comfort around foregoing this practice altogether.

While order entry is key in vital sign workflow, there is still a significant potential for nonadherence given how well-entrenched Q4H vital signs are in today’s clinical practice. This phenomenon is well-illustrated in the discrepancy between our process measure, vital sign ordering, and our primary outcome metric. Given that our global goal was to improve inpatient sleep quality, we decided to measure success as the actual absence of vital signs, rather than only as the placed order, because we felt it better reflected the patient and caregiver’s experience.

Q4H vital signs are perceived as best practice, but evidence does not support their ability to screen for clinical deterioration.¹¹  This project highlights how reducing excessive monitoring for stable patients is unlikely to lead to adverse clinical outcomes and that the frequency should be individualized to each patient and their clinical course. Furthermore, eliminating vital signs on clinically stable patients hypothetically provides greater flexibility to nurse and PCA workflow to devote more time to higher acuity patients.

This study had several limitations. It was a single center project; therefore, results may not be generalizable. We focused on only 2 low-risk groups, and our interventions are likely not applicable to many other diagnoses. However, the concept of individualized vital sign monitoring and the reduction of excessive monitoring is broadly applicable. While our study did not conduct comprehensive patient satisfaction surveys at our local institution, previous research has demonstrated poor patient and family satisfaction around overnight vital signs and an improvement in sleep once these are reduced or eliminated.^5,7,13,16,23  These vital sign checks, however, are one of many possible nocturnal disruptions, and we did not evaluate other interruptions or the amount or quality of patient sleep. Lastly, infants admitted for hyperbilirubinemia at our institution still required temperature checks every 4 hours while undergoing phototherapy. This is a practice that would benefit from reevaluation.

A multidisciplinary team safely and effectively reduced the percentage of patient nights with a vital sign check in predetermined low-risk pediatric populations. Increasing the period of time without nocturnal disruptions and expanding this project toward additional cohorts of patients should be considered.