Reach, implementation fidelity, and safety of bubble continuous positive airway pressure (bCPAP) therapy in children with severe pneumonia in Pakistan

This prospective cohort study in Pakistan reveals that while bubble continuous positive airway pressure (bCPAP) was administered to 53% of children with severe pneumonia, significant gaps in implementation fidelity—particularly regarding continuous monitoring and feeding practices—were associated with a 22% treatment failure rate and highlighted the need for targeted strategies to improve reach and safety.

The Gist

Imagine a child's lungs are like a pair of tiny, waterlogged balloons that have trouble inflating. When a child has severe pneumonia, these balloons get heavy and stiff, making it hard for them to breathe. In many parts of the world, doctors use a clever, low-tech tool called Bubble CPAP to help. Think of it as a "breathing helper" that blows a gentle, steady stream of air into the child's nose, keeping those lung-balloons puffed up so they can work properly. It's like inflating a deflated tire with a steady hand so the car can drive again.

This study is like a "report card" on how well this breathing helper is actually being used in two hospitals in Pakistan. The researchers wanted to know three things:

1. Reach: Did enough sick kids actually get to use the breathing helper?
2. Fidelity: Was the helper used correctly, like following a recipe perfectly?
3. Safety: Did using the helper cause any new problems?

The Story of Two Hospitals

The researchers looked at two different hospitals, which is like comparing a well-stocked private kitchen (Aga Khan University Hospital) with a busy, crowded public community kitchen (Abbasi Shaheed Hospital).

• The "Reach" (Who got the tool?):
  In the public kitchen, almost everyone who needed the breathing helper got it. In the private kitchen, many kids got a more expensive, high-tech version of the tool instead. It's like if a community garden only had one type of shovel, so everyone used it, but a fancy landscaping company had rakes, shovels, and power-tillers, so they chose the power-tiller for some jobs.

• The "Fidelity" (Was it used right?):
  This is where the story gets tricky. Using the breathing helper is like trying to keep a delicate sandcastle intact while a storm is blowing. You need constant attention.

  • The Storm: The study found that the "sandcastle" often got disrupted. The tubes would get knocked loose, or the bubbling would stop. About two-thirds of the time, the machine was working perfectly, but that means one-third of the time, the help stopped unexpectedly.
  • The Watchers: To keep the sandcastle safe, you need someone watching it 24/7. In the private hospital, the kids were in a "VIP zone" with nurses watching them closely. In the public hospital, many kids were in a regular ward where there was only one nurse for 40 kids. It's like having a lifeguard for a pool of 4 people versus a lifeguard for 40 people; the 40-person pool is much riskier.
  • The Feeding Problem: The breathing helper works best if the child's stomach is empty, like a car engine running best without too much fuel in the tank. However, the study found that half the time, kids were still being fed by mouth while wearing the mask. Sometimes doctors ordered it, but often, worried parents sneaked in food because they wanted to feed their hungry child. This is like trying to fix a leaky pipe while someone is still pouring water into it.

• The "Safety" (Did it hurt?):
  The breathing helper is generally safe, but because it wasn't always watched perfectly, some accidents happened.

  • Aspiration: About 1 in 10 kids accidentally breathed some food or liquid into their lungs (like choking on a sip of water while wearing a snorkel).
  • Pneumothorax: One child had a tiny air leak in their lung (like a small puncture in a tire).
  • Leaving Early: A big problem was that 16% of the families left the hospital against medical advice while their child was still on the breathing helper. It's like a mechanic telling you, "Your car is half-fixed, but you need to wait another day," and you drive off anyway because you can't afford the bill or you're worried about something else.

The Big Lesson

The main takeaway is that having the tool isn't enough; you need the right environment to use it.

Imagine you give a world-class chef a great knife (the breathing helper). If the kitchen is dark, the chef is exhausted, and there's no one to watch the stove, the meal might still burn. The study suggests that to save more lives, we don't just need to build more breathing machines. We need to:

1. Train the "Watchers": Teach parents and nurses how to spot when the machine stops working, so they can fix it immediately.
2. Manage the "Feeding": Find better ways to feed sick kids without taking the mask off, so they don't choke.
3. Keep Families Close: Understand why families leave early (money, fear, frustration) and help them stay until the child is truly safe.

In short, the breathing helper is a life-saving invention, but it needs a team of attentive people to make sure it works its magic. Without that team, even the best tool can't save the day.

Technical Summary

1. Problem Statement

Pneumonia remains the leading cause of death for children aged 1–59 months globally, with the highest burden in low- and middle-income countries (LMICs) like Pakistan. Bubble continuous positive airway pressure (bCPAP) is a low-cost, life-saving respiratory therapy that has shown significant mortality reductions in several LMIC trials. However, a major trial in Malawi demonstrated increased mortality in the bCPAP group compared to low-flow oxygen, attributed to poor implementation fidelity (e.g., lack of monitoring, inconsistent feeding protocols, and inadequate staffing).

While bCPAP is widely available in Pakistan, there is a lack of data regarding its reach (proportion of eligible children receiving it), implementation fidelity (adherence to safety protocols like monitoring and feeding), and safety outcomes in real-world hospital settings. Understanding these gaps is critical to preventing the adverse outcomes seen in the Malawi trial and optimizing care in resource-constrained environments.

2. Methodology

• Study Design: A prospective cohort study conducted between February and May 2025.
• Setting: Two contrasting hospitals in Karachi, Pakistan:
  • Aga Khan University Hospital (AKUH): A private, tertiary teaching hospital with high nurse-to-patient ratios (1:6 in wards, 1:1 in PICU) and access to advanced respiratory support (HFNC, mechanical ventilation).
  • Abbasi Shaheed Hospital (ASH): A public, tertiary teaching hospital with significantly lower staffing (1 nurse per 40 ward beds at night) and limited advanced respiratory support.
• Participants: 165 children aged 1–59 months presenting with severe pneumonia (defined by WHO criteria: hypoxemia, severe respiratory distress, or general danger signs) within 24 hours of ED presentation.
• Data Collection: Trained research assistants conducted daily follow-ups via chart review, caregiver surveys, and physical exams until discharge, transfer, death, or loss to follow-up.
• Outcomes Measured (based on the RE-AIM framework):
  • Reach: Proportion of eligible children receiving bCPAP.
  • Fidelity: Four dichotomous measures: (1) Uninterrupted therapy (no unplanned disruptions), (2) Continuous monitoring (pulse oximetry), (3) Location in a highly monitored area (ED, HDU, or PICU), and (4) Avoidance of oral feeding while on bCPAP.
  • Safety: Aspiration events and pneumothorax.
  • Effectiveness: Composite "treatment failure" (death, escalation of respiratory support, or leaving against medical advice) and in-hospital mortality.
• Analysis: Descriptive statistics and mixed-effects multinomial logistic regression models with robust standard errors to estimate average predicted probabilities of fidelity measures over time, adjusting for study site.

3. Key Contributions

• Contextual Implementation Data: This is one of the few studies to systematically measure bCPAP fidelity (adherence to safety protocols) rather than just clinical outcomes in a real-world LMIC setting.
• Site Comparison: By comparing a well-resourced private hospital with a resource-constrained public hospital, the study highlights how infrastructure and staffing directly impact therapy delivery.
• Caregiver Role: The study quantifies the role of familial caregivers in identifying and resolving therapy disruptions, a factor often overlooked in clinical trials.
• Feeding Practices: It provides empirical data on the prevalence of oral feeding and lack of nasogastric tube placement in children on bCPAP, addressing a specific safety concern linked to the Malawi trial's failure.

4. Key Results

• Reach:
  • Overall, 53% (88/165) of eligible children received bCPAP.
  • Reach was significantly higher at the public hospital (ASH: 78%) compared to the private hospital (AKUH: 41%), where many children were treated with High-Flow Nasal Cannula (HFNC) instead.
• Implementation Fidelity:
  • Disruptions: Unplanned disruptions occurred frequently; the average predicted probability of uninterrupted therapy was 66%.
  • Monitoring: Continuous monitoring was highly variable. At AKUH, it was 93%, but at ASH, it was only 12%. Overall, only 56% of bCPAP days involved continuous monitoring.
  • Location: 85% of children were in highly monitored areas (ED/HDU/PICU), though some were in general wards.
  • Feeding: Avoidance of oral feeding was poor. Only 46% of children avoided oral feeds while on bCPAP. Nasogastric tube placement was rare (14% overall), and caregivers frequently administered oral feeds despite medical orders (12% overall).
  • Disruption Resolution: Caregivers were the primary identifiers (80%) and resolvers (54%) of therapy disruptions, often more so than nurses or doctors.
• Safety & Effectiveness:
  • Adverse Events: 10% (9/88) experienced aspiration events; 1.1% (1/88) experienced pneumothorax.
  • Treatment Failure: 22% (19/88) experienced treatment failure.
    • Leaving Against Medical Advice (LAMA): 16% (14/88), a significant driver of failure.
    • Escalation: 6.8% (6/88) required higher respiratory support.
    • Mortality: 1.1% (1/88) died.
• Demographics: Children receiving bCPAP were younger (median age 9 months) and had higher rates of malnutrition compared to those who did not.

5. Significance and Implications

• Fidelity is Critical: The study confirms that bCPAP outcomes are heavily dependent on implementation fidelity. The high rate of oral feeding and lack of continuous monitoring at the public site mirrors the conditions associated with harm in the Malawi trial.
• Caregiver Engagement: The finding that caregivers are the primary detectors of therapy failures suggests that targeted education and empowerment of caregivers are essential strategies for improving safety in resource-limited settings where nursing staff are scarce.
• Feeding Protocols: The high rate of oral feeding despite risks suggests a need for better communication between clinicians and families regarding the dangers of aspiration during bCPAP, and potentially a shift toward more consistent use of nasogastric tubes.
• Policy Recommendations:
  • Strategies to reduce "Leaving Against Medical Advice" (LAMA) are needed, as financial constraints and dissatisfaction drove many families away.
  • Future trials should not just test efficacy but must include implementation outcome measures to ensure safety protocols (monitoring, feeding) are maintained.
  • In settings where continuous pulse oximetry is unavailable, protocols must be developed to train caregivers to recognize signs of respiratory decompensation.

Conclusion: While bCPAP is a vital tool for severe pneumonia in Pakistan, its safety and effectiveness are compromised by gaps in reach, frequent therapy disruptions, inadequate monitoring, and risky feeding practices. Addressing these implementation gaps through team- and individual-targeted strategies is essential to prevent complications and reduce child mortality.