Supportive Non-invasive Ventilation for Acute Chest Syndrome Prevention in Hospitalized Children with Sickle Cell Disease: Determinants of Implementation

Abstract

Background:

In 2017 Boston Medical Center’s (BMC) general pediatric inpatient unit implemented bi-level positive airway pressure (BiPAP) as supportive non-invasive ventilation for acute chest syndrome prevention (SNAP) to prevent adverse respiratory outcomes among medically stable, hospitalized patients with sickle cell disease. Barriers and facilitators to SNAP implementation at BMC differ from those in other settings.

Objective:

To examine contextual determinants of SNAP implementation across different settings.

Methods:

We conducted semi-structured interviews with inpatient clinicians at three sites (Site 1: extensive implementation, Site 2: limited implementation, Site 3: not implemented) about factors that would affect SNAP implementation. Interviews and analysis were guided by the Promoting Action on Research Implementation in Health Services (PARiHS) framework.

Results:

We interviewed 29 healthcare workers (physicians, nurses, physician assistants, child life specialists, respiratory therapists and psychologists). Five major themes emerged: (1) Communication between staff, particularly between the night and day shift, is critical for BiPAP initiation and success. (2) Nurses are key to successful SNAP implementation. (3) SNAP implementation requires multidisciplinary support, including integration of respiratory therapists and child life. (4) Individual unit level size, culture and workflow influence implementation (5) Hospital resources and leadership support are important for SNAP implementation.

Conclusions:

Successful SNAP implementation across different contexts will require optimized communication between healthcare team members and the integration of respiratory therapy and child life early in the process of BiPAP initiation. Adapting to the unique features of each institution will be critical.

Introduction

Sickle cell disease (SCD) affects approximately 100,000 individuals in the United States and is associated with significant morbidity and early mortality. Acute Chest Syndrome (ACS) a clinical syndrome of respiratory symptoms and a new pulmonary infiltrate, is the second most common acute complication after an acute vaso-occlusive pain episode (VOE) and is a leading case of mortality in children with SCD.^1,2 ACS often develops 1-3 days after the onset of a VOE in the setting of pain (particularly in the chest and back) combined with opioid-induced respiratory depression chest wall splinting, hypoventilation, and atelectasis.^2-4 Incentive spirometry (IS) encourages patients to breathe with maximal inspiratory capacity, offsets the effects of splinting and prevents atelectasis, and is an important component of ACS prevention and management.^5-7 However, IS use is limited to awake, actively participating patients; thus, additional strategies are needed to counteract the hypoventilation and atelectasis that contribute to ACS and its progression to acute respiratory failure, particularly while patients sleep.

Non-invasive ventilation, delivered through a facemask, is a promising strategy to address multiple factors that contribute to the development and worsening of ACS. NIV is commonly used in intensive care unit (ICU) settings for patients with acute respiratory failure to support oxygenation and ventilation, prevent the need for invasive mechanical ventilation, and decrease mortality.^8-10 In 2017, we initiated the use of “preventative” nocturnal bi-level positive airway pressure ventilation (BiPAP) on the general inpatient pediatric unit at Boston Medical Center (BMC) for medically stable inpatients with SCD who were deemed “at risk for ACS” because of a history of recurrent and/or severe ACS, chest or back pain, and/or a new supplemental oxygen requirement. A retrospective cohort study of our experience with “Supportive Noninvasive Ventilation for ACS Prevention” (SNAP) over four years demonstrated that approximately 75% of high risk, hospitalized patients for whom BiPAP was recommended had successful BiPAP use for at least one night.¹¹ We recorded no adverse events and concluded that preventative nocturnal BiPAP was safe, feasible and acceptable for patients with SCD on a general pediatric inpatient unit who were at risk of ACS and associated adverse respiratory outcomes.¹¹

We recognize that our experience implementing SNAP was strongly impacted by contextual factors. For example, given our single pediatric inpatient unit and the small number of pediatric subspecialists at our institution, the inpatient team and nurse manager knew and trusted the pediatric pulmonologist and hematologist and were willing to attempt a trial of nocturnal BiPAP with our index patient. After success with the first patient and with support from the nurse manager, we developed an official policy for use of SNAP for hospitalized pediatric patients with SCD. We recognize that early adoption of a therapy typically reserved for an intensive care unit (ICU) setting may not be the typical experience at many institutions.

Given that our overall goal is to develop a protocol to evaluate the effectiveness and implementation of SNAP in a multi-center intervention trial, we conducted qualitative interviews at three sites with varying levels of SNAP implementation, guided by the Promoting Action on Research Implementation in Health Services (PARiHS) implementation science framework.¹² PARiHS is a conceptual framework used to either explain or predict the success of the implementation of evidence into practice. The framework posits that successful implementation of evidence into practice is determined by “the quality and type of evidence, the characteristics of the setting or context, and way in which evidence was introduced or facilitated into practice”.¹³ In a previous publication focused on the Evidence construct of PARiHS and individual level (patient and provider) barriers and facilitators to BiPAP use,¹⁴ we shared that clinical teams, patients with SCD, and parents of children with SCD perceived BiPAP to be effective at preventing ACS and worsening of mild ACS. Participants perceived that “preventative BiPAP” is appropriate on a general pediatrics unit and does not require transfer to the pediatric intensive care unit (PICU). However, there were concerns about the patient experience with BiPAP and the effort required by overnight bedside nurses to help patients acclimate to using BiPAP.¹⁴ This manuscript presents additional data from those qualitative interviews, focusing on inpatient team members’ perspectives on how (Facilitation construct) and where (Context construct) the intervention is implemented. The objectives of this analysis were to examine unit- and institutional-level contextual factors, facilitators and barriers associated with the implementation of SNAP for medically stable hospitalized pediatric patients with SCD.

Methods

This study is a subsequent qualitative analysis utilizing data from our previously reported study that examined perspectives on the evidence for the intervention.¹⁴ The current analysis focuses on the contextual factors, facilitators and barriers to SNAP implementation at three sites. Methodology has been previously described.¹⁴ Given our future goals of a multi-center trial, we were interested in the perspectives of clinicians from sites with varying levels of experience with SNAP. Briefly, we recruited participants from BMC (where SNAP has been used extensively for high-risk patients since 2017), Children’s Medical Center, Dallas (CMCD, where SNAP has been used in a limited capacity since 2020), and Children’s Hospital of Pittsburgh (CHP, where SNAP has not been introduced). We recruited inpatient clinicians at each site (physicians [hematologists, pulmonologists, hospitalists], nurses, respiratory therapists, physician assistants, child life specialists, and health psychologists), several of whom had clinical leadership roles. All audio recordings and transcripts were de-identified prior to analysis. The institutional review board at the BMC/Boston University Medical Campus designated this study as Exempt with a waiver of documentation of consent.

Interviews

We conducted semi-structured interviews via Zoom (Zoom Video Communications) from November 2022-November 2023. Each interview lasted approximately 30 minutes and was conducted by one of two qualitative researchers, a mixed methods epidemiologist (SAB) and a medical anthropologist (KCJ) (see online supplement for further details). We conducted interviews at each site until we reached thematic saturation. We commenced interviews at BMC (extensive SNAP implementation) and revised the interview guide for subsequent sites based on level of implementation. These interviews aimed to elicit hospital-, unit- and provider-level factors that would affect the SNAP implementation. For example, for the Context construct of the PARiHS framework, we asked about how BiPAP use during sleep would fit into the workflow of the unit and who would need to be involved in the process. Under the Facilitation construct we asked about the importance of an official written policy on SNAP and how the availability of respiratory therapists and other multidisciplinary team members (child life specialists, psychologists, trusted longitudinal outpatient clinicians) during the day would facilitate BiPAP use at night. Results for the Evidence construct and development of the interview guide have been previously published.¹⁴

Details about site selection, participant recruitment, and the interview guides can be found in the on-line supplement.

Data Analysis

Interviews were audio recorded, professionally transcribed, checked for accuracy, and imported into NVivo (QSR International Pty Ltd., Release 1.7.1, 2022) for analysis, utilizing deductive coding guided by the PARiHS framework. Four members of the study team (three research team members [SAB, KCJ, CJW] and the PI [RTC]) refined code definitions for the codebook through the examination of three transcripts. After reaching consensus on the codebook, two of three members of the research team coded each transcript independently. We coded transcripts in order of site completion (BMC, CMCD, then CHP). After we completed coding of BMC transcripts, the study team met to discuss themes and subthemes. We repeated this process for each site, and the study team reviewed the data from each site to determine consistency in themes and the emergence of new themes.

Results

We interviewed 34 participants: BMC (n=16), CMCD (n=6) and CHP (n=12), including 29 clinicians (10 physicians, 11 nurses, 1 physician assistant, 2 child life specialists (CLS), 4 respiratory therapists (RT) and 1 psychologist), two patients with SCD and three parents of children with SCD. We interviewed RTs, nurses, physicians, a physician assistant who was a team member at one site, and either a CLS or psychologist at each site. Given that patient- and caregiver- level experience with SNAP was previously reported,¹⁴ and that this paper focused on unit/institutional implementation of BiPAP, here we present results from the 29 clinicians.

Five main themes about SNAP implementation emerged (Please see Table 1 for themes, subthemes, and representative quotes.):

Table 1.

Themes and subthemes: Implementations barriers and facilitators for BiPAP use as supportive care for hospitalized children with sickle cell disease

Theme	Subtheme	BMC	CMCD	UPMC	Illustrative Quote
Communication is critical for BiPAP initiation and success.
	There is disconnect between day and night shift re: patient tolerance of BiPAP overnight. Construct: Context (Barrier)	++	++	−	Nurse: “People who work in the day shift…they don’t deal with this at night.” Hospitalist: “There was also concern that it was just going to be a lot more work and monitoring by the nurses. But it turns out it's actually been less work, because once the patient's on BiPAP then the nurses aren't having to control them into doing incentive spirometry.” Nurse: “Some people think it’s not a big deal. Kind of like ‘just slap it on, it’s not a big deal’ – and it is more than we might sometimes think.”
	Structured bedside nursing report can facilitate communication between shifts. Construct: Context (Facilitator)	N/A	N/A	+	Nurse: “Yeah, I would say it’s pretty seamless, and we do bedside shift reporting, so the families are involved in everything. So, they also help to add any information in if something’s been missed in the communication.” Nurse: “Well, I think we have a good way of doing our bedside report and saying ‘Look, this patient was up all night, so they're probably gonna be sleeping all day. So you have to get those morning meds in,’ and what not, so communication is great…”
	Multidisciplinary collaboration is important – both in the big picture and day-to-day. Construct: Context (Facilitator)	++	++	++	Nurse: “The doctors just sit in their office and write the order and then they expect you to fight with the kid all night long.” Child Life Specialist: “So I think it's more of just that ongoing communication, or, ‘What have you guys already tried? Do you know what happened last night?’…I really try to find out from a nursing perspective, what has already kind of gone down before I even try to really figure out what my role can be, because then I just want to support them…Our alignment together as a care team is what is the most important piece for kids' buy in.”
	Improved integration of Respiratory Therapy and Child Life is needed (including real-time notification when BiPAP is ordered and daytime collaboration to prepare for nighttime use.) Construct: Facilitation (Facilitator)	++	++	+	Child Life Specialist: “And then I think potentially it has to just be one of those kinds of things, like…‘Can I contact respiratory therapy directly myself and kind of find out what would be recommended later for this patient at night?’” Respiratory Therapist: "It doesn't show up anywhere. No, it doesn't just pop up on our page - you have to go in and look for the orders… So if you check it at 9:00 and there's no new orders…and then there's a new order at 10:30 - we won't know about that unless we look it up again or unless the nurse or doctor pages us."
Nurses are key to the success of the implementation of SNAP	Night nurses currently have the heaviest lift Construct: Facilitation (Barrier)	++	++	+	Nurse: “…Generally, it's just me. There's nobody who helps me in the middle of the night.” Nurse: “You play the game of on/off, on/off, at some point you’re just like ‘Ok, let me talk to the doctor and see what else they want to do because this is not working.’” Nurse: “I think sometimes we just have really high acuity on the floor. And a patient that you have to always go in and remind them to put their BiPAP on, I'll hear frustration the next morning of like, ‘…they were giving me a run for my money’… So I think in high acuity times, people are racing around giving a lot of care to a lot of patients.” Charge Nurse: “I think that nursing kind of has the brakes on it about this topic. Just because I think nursing can feel that we are expected to do a lot already and…just trying to not take on too much and to protect our license and the patient.”
	If we started to implement this during the day also (because some patients sleep a lot during the day), then the burden of getting acclimated be spread more evenly to daytime staff. Construct: Facilitation (Facilitator)	++	++	+	Nurse: “So, I think unfortunately both shifts would have to feel very comfortable with using that BiPAP and getting it on them when you see that they're sleeping.” Child Psychologist: “I think ideally it would be something that is put in as an order set, during daylight hours and put in prior to transition. And I think it would need to be something that it's already almost set up during daylight - RT makes sure equipment and everything is in room. And then at night, I would think that turning it on, putting the kids actually on the machine, could be an easier ask on the night RTs or nursing…I think it would take a collaboration between the daylight and the nighttime shifts…I don't think it could be exclusively dependent on the night shift.” Pulmonologist: “But I do wonder if reinforcement with daytime nursing is especially crucial, because the daytime nurse is there with the patient, the patient is awake. And if the daytime nurse has buy-in and kind of reinforces the benefits, I think that could be gold.”
SNAP implementation requires multidisciplinary support	RT and Child Life support are limited at night. Construct: Context (Barrier)	++	++	++	Hospitalist: “…We have less Child Life resources in the evening, and so it's dependent on the nurses being able to spend time at the bedside, kind of acclimating the child to using the mask, which can be really tough.” Nurse: “It depends on the night…They are covering usually a lot of floors in addition to ours…sometimes they can come relatively quickly and sometimes it might take them a while. I mean, we can escalate if it's a more urgent issue but as far as switching a mask - that might be lower on their priority list. Especially if it's not because of an escalation of care.”
	There is, however, underutilization of child life resources during the day. Construct: Facilitation (Barrier)	++	++	N/A	Child Life Specialist: “I'm usually looped in very much by…almost inserting myself and asking questions…or if I hear about a patient in rounds or/and I kind of just have some forward thinking, I'll talk to the nurse directly. I'm not necessarily consulted outside of [the pulmonologist] mentioning it to the medical team and saying ‘You should call [Child Life] about this.’” Child Life: “Typically they don't call us unless there's an issue, which I think could potentially be changed as well, so that we can get ahead of the problem…versus when the patient is maybe already escalated or frustrated about wearing the mask… If we found out ahead of time, maybe we could help provide that education and maybe that would prevent some of the compliance issues.”
	Including Child Life/Child Psychologist early in the process will be crucial Construct: Facilitation (Facilitator)	++	++	+	Nurse: “…We don't have a lot of child life coverage at night. So doing the first time putting on a mask and having that kind of claustrophobic feeling at night is not ideal… Our plan is to hopefully get child life involved during the day. Make this be a very planned event… with Child Life and RT doing a mask fitting sometime in the afternoon before we start doing a trial during the day is what we're trying to do.” Child Psychologist: “I would want to be contacted early…I round with the team like 2-3 days a week. So I think I have the benefit of that and that I'm directly involved in the conversations with the multidisciplinary team… I also get overnight sign-out …so I think just because of how embedded I am in the division and I think that I would probably have awareness of it. And I also think I'd probably end up being a point person in terms of linking in other psychosocial team members.”
Individual unit level size, culture and workflow influence implementation	Small unit and close relationships between medical teams help with staff acceptance. Construct: Context and Facilitation (Facilitator)	+	N/A	N/A	Hematologist: “The close relationship between all of us in pediatrics: Everybody already trusted [the pulmonologist]. Everybody, I think already trusts us in hematology… It wasn't somebody out of the blue saying this, it was somebody that the teams had already known and already trusted and that is crucial…” Hospitalist: “Having someone who is a recognized expert who is very accessible and able to talk to individual clinicians, not just in a large Zoom meeting.” Hospitalist: “… Just the sheer size of this place makes things like this difficult. So that's just something to keep in mind that our training process and buy-in process is going to be more unwieldy.”
	Making sure the patients needing BiPAP get admitted to the specific SCD unit is important to success Construct: Facilitation (Facilitator)	N/A	+	+	Pulmonologist: “Now, the fact that we're going to be focused probably on one geographic ward, which is like 30 beds, that is in our favor that they already try to cluster these kids into one area. And so that hopefully will help us.”
	The unit where this will be implemented is open to innovation and has experiences with clinical trials Construct: Context (Facilitator)	N/A	N/A	+	PA: “We are in a teaching hospital that's very research-centered. We have people sign a registry to be part of any other studies and… the vast majority - if not all - of our patients sign it. They're always willing to talk to new people coming in…I can't think of any families that would be against hearing about things that could help their kid.”
Hospital resources and leadership support are important for SNAP implementation	Having an official hospital policy is crucial for health care team acceptance. Construct: Context (Facilitator)	+/−	++	++	Hospitalist: “Yes. I think any time there's any written protocol that can be referenced, it's incredibly helpful for ensuring that there aren't going to be delays or hang-ups related to somebody not knowing what the expectations are.” Nurse: “I've been putting kids on BiPAP for well as long as they've had BiPAP… No, I don't need to read a policy, I know exactly what it's for and what it does and how to do it.” Respiratory Therapy: “Here in this institution, nursing - the only thing that they are, per policy, supposed to… mess with on the ventilator, or adjust, is Fi02. Anything else is supposed to be left up to the RT's, and we're very big sticklers on that.”
	RT staffing limitations are/will be an important barrier to successful implementation. Construct: Context (Barrier)	+	++	+	Nurse: “I think they'll be a little taken back, to be honest. Their staffing isn't the greatest right now and I think they'll worry that they're going to be pulled in more than they want to be, or more than they're able to right now…I think they'll have a lot of questions as well as how heavily they'll have to be involved.” Nurse: “It's a hit or miss during the day or at night. It's also who's ever on. You'll sometimes get RTs that are great and that will help you and that will come right away. Others, obviously, are really busy…They are short-staffed as well most of the time.” Respiratory Therapist: “A barrier sometimes could be possibly the staffing, only because we would have to really work around a lot to make somebody be available, and we always do, but that would be a barrier as well as the staffing just for initiation, because it does take such a long time.”
	Buy-in: leadership was supportive, overall belief that it works, normalization of BiPAP as preventive care on the unit (e.g. help residents and newer staff learn about preventative BiPAP). Construct: Context (Facilitator)	+	N/A	N/A	Hospitalist: “Having a nurse champion also made that project successful. So there's one thing I've learned by doing some of these projects is having a nurse champion is indispensable to making this work.” Hematologist: “I mean, there's no way to overstate how important the nurse manager's role was in this…she supported all of her nurses…she listened to them, she validated their concerns, but she didn't let that stand in the way of implementing this… So her role is probably the most important in all of this in terms of getting this accepted.” Hospitalist: "I think it's just an expectation…It's a part of the culture of our unit that we know that data now supports its use. I think there's never any pushback. I think it's even led to people being I think more open to other kinds of respiratory support on the floor for patients with acute as opposed to chronic illness.”

ACS: Acute chest syndrome; BiPAP: bi-level positive airway pressure; BMC: Boston Medical Center; CMCD: Children’s Medical Center, Dallas; N/A: not a theme at that site; ICU: intensive care unit; PA: physician’s assistant; RT: respiratory therapy; ++: explicit agreement by many participants; +: explicit agreement by some participants; +/−: some participants agree, others disagree; − : explicit disagreement by some participants; −−: explicit disagreement by many participants

1. Communication is critical for BiPAP initiation and success. There was strong agreement across all three sites that communication among the different clinical staff involved in patient care is important for successful SNAP implementation. At BMC and CMCD where SNAP has been implemented, there was concern about the “disconnect” between day and night shifts regarding patient tolerance for BiPAP overnight. Night shift staff reported that those on the day shift did not appreciate the difficulty in getting patients to use the BiPAP overnight (“People who work in the day shift – they don’t deal with this at night.” - Nurse). Conversely, CHP (where SNAP has not yet been implemented) uses a detailed bedside nursing report process about patient care between shifts; they did not anticipate barriers to communication about BiPAP use between day and night teams. All three sites agreed that cross-disciplinary collaboration and communication was important for both big picture and day-to-day SNAP implementation. BMC and CMCD clinicians highlighted a need for improved and earlier communication with RT and CLS to help facilitate intervention success. For example, RTs and CLS stated they were often not alerted during daytime hours when BiPAP is ordered for the upcoming night.

2. Nurses are key to SNAP implementation success. Night shift nurses at BMC and CMCD reiterated that currently they carry the burden of implementing SNAP, given that they are the ones at the bedside who ensure that patients wear BiPAP (“Generally it’s just me. There’s nobody who helps me in the middle of the night.” – Nurse). For patients with difficulties tolerating BiPAP, this can be burdensome and competes with other nursing duties. At CHP, nursing staff felt that SNAP implementation could add to their workload. Night nurses suggested that extending BiPAP use to the day (for brief daytime trials to get acclimated and during naps, especially for adolescents who sleep more during the day) would be beneficial. This would allow for more participation by other team members including CLS, and the burden of getting patients acclimated to BiPAP would be shared by both day and night staff (“If the daytime nurse has buy-in and kind of reinforces the benefits, I think that could be gold.” - Physician). Although this would be helpful to night staff, it would require additional training for daytime nurses.

3. SNAP implementation requires multidisciplinary support. A major hurdle to multidisciplinary support across the three sites was the lack of RT and CLS support at night. Although CLS are present during the day and they expressed that helping patients get acclimated to BiPAP is squarely within their scope, they reported being underutilized (“Typically they don't call us unless there's an issue, which I think could potentially be changed…so that we can get ahead of the problem.” - CLS). CLS and psychologists stated that it would be beneficial if they could collaborate with RT during the day to help prepare patients for BiPAP use at night. This was echoed by nursing staff who also felt that doing daytime trials would be beneficial for patients as they would have access to both RT and CLS to aid in adjusting to BiPAP.

4. Individual unit level size, culture and workflow influence implementation. At the unit level, clinical staff at BMC stated that having a small unit and close relationships between medical teams helped with staff acceptance (“It wasn't somebody out of the blue saying this, it was somebody that the teams already knew and trusted and that is crucial.” – Physician). Staff at CMCD and CHP, both free-standing children’s hospitals, stated that cohorting patients within a specific inpatient unit that routinely provides SCD care would optimize familiarity with this intervention and improve the likelihood of success. At CHP, clinical staff indicated that an important facilitator to future SNAP implementation is the culture of innovation that exists on their inpatient hematology unit. Due to their experience with clinical trials, clinicians believe that patients, caregivers, and staff on their hematology unit would be open and accepting to implementing SNAP.

5. Hospital resources and leadership support are important for SNAP implementation. There was general agreement that an official institutional policy would be crucial for healthcare team adoption of SNAP. At BMC, where there is a policy for SNAP, nurses had mixed views on its importance; some were not aware of the policy’s existence but still found SNAP acceptable, and others knew of the policy but did not feel like it strongly influenced their acceptance of SNAP. However, at CMCD and CHP, clinical staff stressed that an official policy with clear parameters for SNAP would be essential to ensure patient safety and for the acceptance of SNAP by the healthcare team. There was universal agreement that RT staffing limitations is an important barrier to successful implementation at BMC and CMCD and would be an important barrier to implementation at CHP (“Their staffing isn't the greatest right now and I think they'll worry that they're going to be pulled in more than they want to be, or more than they're able to right now.” - Nurse). BMC clinicians explained that buy-in from leadership (the inpatient unit medical director and nurse manager) was critical to staff acceptance; this helped with normalizing BiPAP use (typically an ICU-level intervention) on the general pediatrics unit (“Having a nurse champion is indispensable to making this work.” - Physician).

Discussion

In this qualitative study of healthcare workers in diverse roles from three institutions, participants shared that successful SNAP implementation would require good communication among clinical staff and a multidisciplinary team approach, including optimizing the integration of RT and Child Life into the SNAP workflow. Participants emphasized that nurses are critical to the successful implementation of SNAP, although there were mixed views on the additional burden this would place on nursing staff. There were specific institutional needs and unit features to guide or aid implementation such as additional RT staff, buy-in from clinical leadership, or, in some cases, an institutional policy. However, respondents’ views on the impact of some factors varied based on the level of implementation at each site. For example, including a hospital policy guiding implementation was very important at CMCD and CHP where there was either limited or no use of SNAP. Conversely, some BMC clinicians were unaware of the existing policy, given the unit culture of using BiPAP as a preventative treatment.

No prior studies have reported the implementation of NIV for ACS prevention on a general medical inpatient unit, but a small number of studies have reported the use of NIV in ICUs and intermediate care/ICU “step-down” units as treatment for “early ACS.” One prospective study of NIV for adults with “early-stage ACS” was conducted in an ICU step-down unit rather than a general medicine floor.¹⁵ Ninety four percent of participants tolerated the NIV protocol, and no adverse events in association with NIV use were reported.¹⁵ Heilbronner and colleagues describe their single-center experience using early NIV as first-line treatment for ACS among patients 5-20 years of age in their PICU.¹⁶ NIV was initiated 95% of the time with a nasal mask, which was perceived as easier to monitor and had better patient acceptance compared to facial masks. The authors observed “a better overall compliance to NIV in SCD patients than in other respiratory diseases.”¹⁶ A French national survey of pediatric intensivists representing 17 PICUs and pediatric intermediate care units examined use of NIV for ACS in 2015.¹⁷ Three centers designated “high-experience centers” treated 65% of the 137 ICU cases of ACS. NIV was perceived to be well-tolerated by patients with SCD, though its use varied substantially such that 71% of patients in the 3 “high-experience” centers received NIV compared to only 17% of patients in the 14 “low-experience” centers. NIV was most often applied in PICUs rather than intermediate care units because of physician and nurse experience with ventilators and higher staffing levels.¹⁷

More broadly, the real-world feasibility of implementing NIV use on a general ward setting was first demonstrated in the multi-center Yorkshire Noninvasive (YONIV) trial,¹⁸ in which 236 adult patients with acute exacerbations of chronic obstructive pulmonary disease (COPD) who were hospitalized on 25 general wards in 14 hospitals in the United Kingdom were randomized to receive standard COPD treatment or standard treatment plus NIV. Twenty two of the 25 wards had no experience with NIV, and the median nurse:patient ratio was 1:11. Each ward received an average of 7.6 hours of formal training by a research physician and nurse at the beginning of the trial, with approximately one hour of ‘refresher’ training per month to maintain skills. In addition to demonstrating that the addition of NIV was effective at preventing the need for intubation and in-hospital mortality, the YONIV trial showed that NIV is feasible in the real world, outside the ICU.¹⁹ A subsequent 2015 international survey of 151 respondents from 51 countries about NIV use in general wards²⁰ demonstrated that NIV was being used on general wards for patients with acute respiratory failure (most often for COPD exacerbations) in 66% of hospitals and had been increasing over time. Limited training and staffing limitations were the most common reasons reported by hospitals not using NIV on the general wards.²⁰ More recently, high volumes of patients with acute respiratory failure combined with the limited number of ICU beds and ventilators during the COVID-19 pandemic led to further expansion of NIV for adults outside intensive care settings.²¹

Given our findings that bedside nurses play a key role in the success SNAP implementation, ensuring their comfort with this technology outside the ICU is essential. In an anonymous survey study of 90 nurses responsible for managing NIV for patients with acute respiratory failure on 4 general wards in a hospital in Milan, Italy nurses reported that they had not received adequate training in its use.²² A review of 8 studies describing nurses’ experience with NIV found that lack of adequate training to safely care for patients on NIV was a common theme.²³ To address this need during the COVID pandemic, Bambi et al described a 2 session, 3 hour-per-session on-the-job training program delivered to ward nurses by an “NIV Team” of experienced ICU nurses.²⁴ Post-training, nurses self-assessed proficiencies across a range of skills increased, and the authors felt that on-the–job training was a valuable and fast approach to implement NIV management skills among nurses outside the ICU.²⁴ An NIV education and hand-on simulation program, consisting of six 30-minute sessions for nurses on a general internal medicine ward in Korea was associated with greater increases in knowledge and self-efficacy about NIV nursing skills compared to a control group who only received a lecture about “ventilator basics.”²⁵ Participants reported improved understanding of the patient experience, increased confidence and reduced anxiety regarding applying NIV, but they also expressed the desire for ongoing training. Our findings and those described above are consistent with a systematic review of 43 published studies about facilitators and barriers to implementation processes for hospital-based interventions.²⁶ Lack of confidence or ability and self-reported lower skills were significant but addressable barriers to implementation of a proposed intervention; training, awareness campaigns, and regular in-services to maintain skills and train new staff were helpful facilitators of implementation success in hospital-based settings.²⁶

In addition to nurse training on the use of NIV, optimizing multidisciplinary collaboration also emerged as an important facilitator of successful SNAP implementation. Farmer and colleagues published an evidence-based “interprofessional algorithm” to improve NIV utilization for patients with COPD exacerbations.²⁷ Their algorithm included clear delineation of roles among physicians, nurses, and respiratory therapists with regards to each step of NIV delivery including patient assessment, NIV set up, monitoring, and recognition of the need to escalate care, and offered approaches to interdisciplinary collaboration and communication for each aspect of NIV care delivery.²⁷ Given the well-described role of CLS in helping patients and families prepare for and cope with difficult procedures,²⁸ expanding this model of multidisciplinary collaboration on pediatric units to systematically include the integration of child life specialists and/or psychologists in helping pediatric patients acclimate to NIV – particularly during daytime trials - has the potential to alleviate some of the workload burden²⁹ as repeatedly described by the nurse participants in our study.

Separate from the specific implementation of NIV, previous studies have demonstrated that unit-level implementation of inpatient interventions for hospitalized patients with SCD and VOE is both feasible and potentially effective at reducing the incidence of ACS. One pre-post multimodal intervention study found that implementation of a standardized order set for the management of VOE, which included automatic ordering and nursing staff monitoring of incentive spirometry, criteria for supplemental oxygen, and guidelines for opioid use, was associated with a 50% decrease in the incidence of ACS and an increase in the time to ACS development from 0.8 to 1.7 days (p=0.05).³⁰ A second pre-post retrospective study examining outcomes associated with the implementation of mandatory IS for patients hospitalized with VOE noted a decline in the incidence of ACS and the need for blood transfusions, specifically among patients who presented with back pain.⁷ Similar to our study of provider perceptions,¹⁴ the providers in this study believed that “atelectasis prevention is paramount.”⁷ Finally, Morrone and colleagues presented the results of a multimodal quality improvement initiative for hospitalized children with SCD who did not have ACS at the time of admission. The study team described a detailed checklist that was implemented using an Epic^™ Smartphrase and included multiple interventions related to opioid dosing, IV fluid administration, addressing oxygenation status, and IS use and documentation. They described a comprehensive implementation strategy including ongoing nurse and resident education, competitions, and regular multidisciplinary meetings. They reported a non-significant reduction in ACS incidence ACS with significant improvement in the documentation of IS use.³¹ Taken together, these studies suggest that inpatient interventions aimed at improving oxygenation and preventing atelectasis were feasible, acceptable, and potentially associated with a reduction in the development of ACS.

Our study had several strengths. We conducted interviews with healthcare workers from three sites with varying levels of implementation and different institutional structures. By gaining perspective from a site where SNAP had been implemented we know that implementing SNAP is feasible. Furthermore, by conducting interviews across multiple sites, we learned valuable knowledge about barriers and facilitators across different settings. We had diverse representation of clinician roles at each site, which is important given the multidisciplinary approach needed for successful implementation of SNAP. We also note some limitations. Although we recruited broadly at BMC, we used targeted sampling among RTs. We also used targeted sampling via clinician referral to recruit providers at our partner sites. Given our approach, we did not have information on those providers who did not participate in the study. Thus, there is the potential for selection bias as those who participated may have perspectives that do not represent the entire staff. Additionally, while our study included three sites, our findings may not be generalizable to institutions with different contextual factors at play (e.g. setup of the unit, hospital policies). In terms of implementation, some barriers may not be easily addressed and could inhibit the implementation of SNAP. Across all sites, addressing RT staffing shortages are important given their role in BiPAP initiation. This issue was faced by several institutions after the COVID-19 pandemic, and building the RT work force will take time. At BMC, where SNAP has been implemented, nursing staff felt comfortable troubleshooting minor BiPAP issues if RT was unavailable. However, nursing staff at other sites would need some BiPAP training, which may require a change in hospital policy and potentially result in an increased burden on the nursing staff.

Our prior work^11,14 and the data we present here suggest that implementation of SNAP on a general pediatric inpatient unit is feasible, albeit with strategies that address the local context and needs of each institution, including staffing to support introduction and acceptance of BiPAP by patients with SCD “on the general floor.” There will be a specific need for increased RT support, communication across clinical staff and shifts and integration of child life into the clinical care model. Although sites differed on the importance of a hospital policy to guide the implementation of SNAP, it was clear that one would be needed to improve acceptance, particularly if the implementation of SNAP would require a change in clinical processes at the institution. These findings will guide the development of a multicenter hybrid effectiveness-implementation trial of SNAP as a potential strategy to reduce the incidence of ACS among hospitalized patients with SCD.

Abbreviations:

ACS

acute chest syndrome

BiPAP

bi-level positive airway pressure

BMC

Boston Medical Center

CHP

Children’s Hospital Pittsburgh

CLS

child life specialist

CMCD

Children’s Medical Center Dallas

PARiHS

Promoting Action on Research Implementation in Health Services framework

PICU

pediatric intensive care unit

RT

respiratory therapy

SCD

sickle cell disease

SNAP

Supportive Non-invasive Ventilation for ACS Prevention

VOE

vaso-occlusive episode