Experiences with implementation of continuous positive airway pressure for neonates and infants in low-resource settings: A scoping review

Background Continuous positive airway pressure (CPAP) is the gold standard of care in providing non-invasive positive pressure support to neonates in respiratory distress in high-resource settings. While safety has been demonstrated in low-resource settings, there is a lack of knowledge on the barriers and facilitators to proper implementation. Objective To identify and describe the barriers, facilitators, and priorities for future implementation of CPAP for neonates and infants in low-resource settings. Methods A systematic search (database inception to March 6, 2020) was performed on MEDLINE, Embase, Web of Science, CINAHL, Global Health, and the WHO Global Index Medicus using PRISMA-ScR guidelines. Original research articles pertaining to implementation of CPAP devices in low-resource settings, provider or parent perspectives and experiences with CPAP, cost-benefit analyses, and cost-effectiveness studies were included. Inductive content analysis was conducted. Findings 1385 article were screened and 54 studies across 19 countries met inclusion criteria. Six major themes emerged: device attributes, patient experiences, parent experiences, provider experiences, barriers, and facilitators. Nasal trauma was the most commonly reported complication. Barriers included unreliable electricity and lack of bioengineering support. Facilitators included training, mentorship and empowerment of healthcare providers. Device design, supply chain infrastructure, and training models were imperative to the adoption and sustainability of CPAP. Conclusion Sustainable implementation of CPAP in low resource settings requires easy-to-use devices, ready access to consumables, and holistic, user-driven training. Further research is necessary on standardizing metrics, interventions that support optimal provider performance, and conditions needed for successful long-term health system integration.


Introduction
The World Health Organization has declared the reduction of neonatal mortality a global priority [1]. Each year, two and a half million infants die in their first month of life and the majority of these deaths occur in low resource settings [2]. While considerable progress has been made over the last few decades, respiratory distress syndrome (RDS) remains a leading cause of neonatal mortality worldwide [1][2][3][4]. RDS usually develops in the first 24 hours after birth in premature newborns due to a lack of surfactant within the lungs, and often requires positive pressure ventilation for treatment [5]. Continuous positive airway pressure (CPAP) is considered to be the gold standard, treatment for preterm neonates experiencing RDS and is recommended by WHO [6-9].
Forms of CPAP can vary across a number of factors including the patient interface, sophistication, and how they generate pressure. Bubble continuous positive airway pressure (bCPAP) is a common mode of CPAP delivery for newborns that uses a bubbler instead of a ventilator to generate pressure [6][7][8]. Since bCPAP systems are considered at least as efficacious and are considerably lower cost than ventilator-derived CPAP devices, they may have significant potential to improve access to non-invasive ventilation in low-resource regions worldwide [7, 10,11]. While reviews of all forms of CPAP [12,13] have described the efficacy of the treatment, there has been a specific focus on bCPAP therapies suggesting that bCPAP may be safe and effective in low and middle income countries (LMICs) [14][15][16]. These reviews called for further research on effectiveness and sustainability of bCPAP therapy in lowresource settings [13][14][15][16]. A recent systematic review on barriers and facilitators to implementation of neonatal bCPAP among health facilities in sub-Saharan Africa found that staffing ratios, provider knowledge, and device maintenance were crucial to the success of the intervention [17]. However, more information is needed to understand optimization and guide further implementation of all forms of CPAP, including bCPAP, across low-resource settings. Consideration of implementation factors such as successful CPAP device attributes, provider and parent acceptance, and systems uptake must be better understood. Additionally, a broader picture that considers qualitative factors is needed to understand how to create lasting sustainable uptake of CPAP. To explore these factors the following research question was formulated: What are identified barriers, facilitators, and priorities for future implementation of CPAP for neonates and infants in low-resource settings? To answer this more qualitative and nuanced question, a scoping review was chosen to broadly map knowledge gaps and evidence [18].

Search strategy
The scoping review framework was adopted in order to present an overview of all the evidence relating to experiences with CPAP implementation [19]. A scoping review protocol was developed according to the Joanna Briggs Reviewer's manual [20] and this review is reported in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for Scoping Reviews (PRISMA-ScR) checklist (S1 File) [21]. The final protocol was registered on Open Science Framework (https://osf.io/qwvgs/). The search query (S2 File) was run on six databases (MEDLINE, Embase, Web of Science, CINAHL, Global Health, and the WHO Global Index Medicus) from database inception to March 6 th , 2020.

Selection of studies
Search results were uploaded to an online program (Covidence, Veritas Health Information, Melbourne, Australia) to allow for collaborative screening by multiple authors. Four reviewers (SD, RS, HA, AS) independently screened a sample of ten titles and abstracts and agreed on criteria for inclusion and exclusion. Two blinded reviewers (SD, RS) independently screened all articles by title and abstract. Conflicts were resolved by an independent arbiter (RH). Two blinded reviewers (SD, RS) then screened articles by full text for potential eligibility. A final arbiter (RH) resolved conflicts of agreement on inclusion for the final dataset. Original peerreviewed research articles of any study design on implementation of CPAP devices in lowresource settings as defined by the World Bank Classification at time of study, provider or caregiver perspectives and experiences with CPAP, and cost-benefit analyses or cost-effectiveness studies were included. Grey literature, reviews, and research articles that solely focused on safety and efficacy of CPAP were excluded.

Data extraction
Three reviewers (SD, HA, AS) independently extracted data from each study using the Covidence data extraction form. Extracted data included: study year; study type/method and setting; population; sample size and method; study objectives; characteristics of CPAP intervention or treatment; complications, barriers, and facilitators. Findings were coded into broad themes by two independent reviewers (SD, HA) using an inductive content analysis on NVivo 12 (QSR International, Melbourne, Australia). An inductive analysis was used in order to uncover patterns and themes in the experiences and perceptions of CPAP implementation [22,23]. Once all studies were uploaded into NVivo, the two reviewers coded a sample of the studies until data saturation was reached. The individual codebooks were compared and discussed in order to create a final codebook which was then applied to the full dataset.

Synthesis of results
Studies were grouped by intervention. Broad categories were developed from extracted data related to experiences with implementation of CPAP treatments and results were synthesized across articles. Due to the high variation in study designs and in order to capture and present all of the existing data, studies were not excluded based on quality; and therefore, critical appraisals were not conducted.

Overview of included studies
Of the 1385 identified studies, 54 were included in the final analysis (Fig 1-PRISMA

PLOS ONE
Implementation of CPAP for neonates and infants in resource-poor settings: A scoping review studies were excluded based on language (six non-English papers were translated using Google Translate). Findings were coded into six main categories: device attributes, patient experiences, provider experiences, parent experiences, barriers, and facilitators (Inter-rater reliability kappa score 0.91).

Description of included studies
The 54 included studies were conducted in 19 countries over five regions: Africa (n = 23), Asia (n = 15), Central & South America (n = 9), Middle East (n = 6), and Oceania (n = 11). Studies ranged from analysis of CPAP treatments (n = 34), training processes (n = 7), patient interfaces (n = 5), nasal protection (n = 3), body positions (n = 2), pain relief (n = 1), and general knowledge or perception surveys (n = 2). Most included studies were randomized controlled trials (n = 10), followed by observational (n = 8) and prospective cohort (n = 6) studies. The most common study populations were term and preterm neonates (n = 18), followed by only preterm neonates (n = 11), and healthcare providers (n = 15). Four studies described their sample population with the general term "infants," which refers to ages 1-12 months, so unless specifically mentioned, the following findings refer to preterm and term neonates, defined as under one month of age.

Device attributes
Fourteen different CPAP devices were described across the included studies, including Fisher & Paykel (n = 14), Pumani (n = 8) and locally-made or improvised devices (n = 9). CPAP devices varied in price, features, and patient interfaces. Price was one of the most common themes overall. Five studies emphasized that affordability and cost-effectiveness of different CPAP devices encouraged implementation [25][26][27][28][29] while five studies cited that if a CPAP device was expensive, cost was a barrier to implementation [25-27, 30, 31]. Commercial CPAP devices were noted to have other challenges. For example, A number of studies also reported on techniques to reduce nasal trauma through application of protective dressings and use of various patient-device interfaces. In two studies, hydrocolloid dressings, a soft gel-based dressing, effectively reduced nasal injuries [50,51]. Two of the four studies that compared nasal prongs to nasal masks concluded nasal masks were associated with statistically significant lower incidences of nasal injuries [(36% vs 58%] [52] (33% vs 92%) [53]].
Seven studies described pain or discomfort experienced by a patient on CPAP treatment [37, 45, [54][55][56][57][58]. These studies noted different levels of reported pain (assessed using validated pain assessment tools) based on device type and patient position. Khan et al. found that neonates in a local low-cost CPAP (J-CPAP) group had significantly lower average Neonatal-Pain Agitation and Sedation Scores (N-PASS) than those in a Fisher & Paykel bCPAP group [45]. Osman et al. reported higher pain scores in an nCPAP group compared to high flow nasal cannula [58]. Jabraeli et al. compared pain scores across supine, prone, and facilitated tucking (fetal) positions with nCPAP and described that the lowest pain scores were recorded when the neonate was in a fetal position [56]. Two additional studies found that when neonates received CPAP in a prone position, heart rates and respiratory rates were lower [59], but there were higher rates of nasal prong displacement (56% required repositioning) [60].

Parent experience
Four studies reported on parents' experience when their newborns underwent CPAP treatment [33,49,50,61]. These studies emphasized that communication between HCPs and parents is important. Parents should be taught about CPAP and engaged in their neonate's care [61]. Two studies described parents' fears related to CPAP treatment [49,61]. Nyondo-Mipando et al. stated: "Study participants reported that caregivers sometimes had fears that the many tubes interfered with breathing and that oxygen therapy was associated with death-a perception that may have been influenced by the lack of clear, effective communication between providers and caregivers" [49]. These two studies also reported on parent interactions with their babies while on CPAP treatment. Participation in their infant's care, such as checking for bubbling in the device, was associated with decreased anxiety and worry [49,61].

Provider experience
Multiple studies discussed providers' knowledge of CPAP, device assembly, and patient selection for CPAP treatment. HCPs were more confident in their ability to use CPAP when the devices were simple and accompanied by quality training [

Barriers
The primary barriers to CPAP implementation were a lack of HCPs and insufficient facility resources. HCP turnover and scarcity were often cited as limitations to effective training and quality patient monitoring [ [67].
Facility resource constraints included lack of uninterrupted electricity, compressed air, oxygen blenders, specific CPAP protocols [72], and computers for record keeping [65]. Reliable electricity was the most frequently described facility infrastructure barrier that affected both patient care [34, 42, 49] and training [68]. In some instances, facility backup generators were not reliable during power outages [49]. Equipment shortages at medical facilities and in supply chains were the most commonly noted of all physical barriers [27,32,34,49,71,72] A major finding from this review was that it is essential that CPAP devices are easy to assemble, use, maintain, and have simple bioengineering support [33,39]. Evidence has shown how devices designed in high resource settings are not sustainable as once they break, there is no bioengineering support to fix them [74,75]. While CPAP devices have traditionally been designed in high resource settings, the unique contexts of low resource regions need to be considered when implementing CPAP across these settings. For example, the polite bCPAP device was specifically designed after surveying Nigerian HCPs on their preferences. Affordability, transportability, and simplicity were the most essential characteristics [25]. The essential takeaway here is that a device's success is dependent on the users and their settings and therefore it is imperative to involve the target audience in the design and implementation process. Such a human-centered design approach has a greater potential to create sustainable, context-based solutions [76]. Incorporating human-centered design facilitates local ownership of CPAP devices and programs by creating a system that may be more appropriate and sustainable [77].
In addition to engineering devices to match their settings, the sustainability of their consumables must also be considered [74]. It is well understood with any device that without available consumables devices will be unusable and only generate waste. That is why it is essential future interventions go beyond facility introduction of CPAP devices to comprehensive integration into health systems in order to ensure sustainability and scale. This includes engaging local manufactures and supply chains. Another solution includes understanding what components could be safely cleaned as reused. Two studies in this review did so for nasal prongs [48,78], but there is a need to determine safe and standardized reprocessing procedures that are feasible across facilities with different levels of resources. These factors should also be considered in the initial design of devices as mentioned above [76].
Quality training and mentorship were identified as vital facilitators of successful CPAP implementation [17]. Providers must feel confident, empowered, and knowledgeable about CPAP to support and encourage long-term implementation. There is a need for more evidence on different models of training and mentorship, especially taking into account limitations on staff availability. The findings from this review suggest that training models should be integrated into the flow of work with interval in-service training and simulation. As with device design, the development and implementation of training materials should be co-created with local healthcare provider leaders in the settings where they will be used. This will not only foster engagement, but also further adapt education and use to the particular setting in which it will be used [79].

Limitations
A limitation of this review was the significant variation in study design across the included studies. By setting out to capture a wide range of experiences, we incorporated studies with varied interventions and outcomes. For example, the subset of papers on complications and interventions associated with nasal injury were challenging to compare with studies that reported on the effectiveness of different CPAP devices.

Conclusion
Inconsistent parameters and outcomes between studies to-date have prevented meta-analyses [13][14][15][16]. The study designs, interventions, and objectives in our included studies were also remarkably diverse. Each of the studies in this review addressed an aspect of CPAP implementation that is important to consider when planning for long-term integration of this treatment. While implementation factors are often addressed separately from efficacy and safety in highresource settings [80], the breadth of experiences described in this review indicates how these measures must be considered concurrently in low-resource settings. Future effectiveness studies should consider not only the short and medium term population outcomes, but also factors that influence sustained integration of CPAP into health systems. A standardized set of implementation outcomes for future research-common barriers and facilitators to study-could allow for improved data synthesis and guidance on optimal care and future research questions.
Successful implementation and integration of CPAP devices across health systems in lowresource settings require appropriate devices, reliable supply chains to replace consumables, and innovative training models that engage users. Each of these elements have one key connection: they each require a deeper engagement of healthcare workers and health systems using these devices. From start to finish CPAP design and implementation should be driven by the final users and the system in which they operate. Combined, it is the hope that these efforts can empower and promote device use, rather than perpetuate potentially unsustainable implementation processes for CPAP use in low-resource settings.