Introduction
Globally, every year, of the 140 million newborns born, 10–15 million do not cry after birth and resuscitation is required for these newborn to accomplish spontaneous breathing.1–4 Despite progress and efforts to reduce newborn deaths, over 90% of these deaths still occur in low-income and middle-income countries (LMICs), and most intrapartum deaths can be prevented with effective resuscitation.5 Anxiety and fear among healthcare providers (HCPs), difficulties in assessing the newborn’s condition and providing appropriate clinical response usually delay the initiation of bag and mask ventilation.6
To improve the competency on newborn resuscitation in LMICs, Helping Babies Breathe (HBB) training has been rolled out since 2010.7 8 Following the implementation of HBB training, there have been improvement in HCPs skill competence in newborn resuscitation.9 10 However, there is a rapid skill decay in skill competence of newborn resuscitation over a period of time.11 To tackle this problem, simulated short-term training sessions, such as structured skill drill in newborn simulator, have shown to maintain and retain skill competence on resuscitation.12 13 Despite maintenance of skill competency, implementation in clinical care have been low.14 HCPs have failed to translate their skills into clinical performance, and as a result infants who require resuscitation do not receive timely ventilation15 (figure 1A,B).
To improve the clinical performance, a periodic reviewing method using Plan–Do–Study–Act (PDSA) have been implemented.16 Reviewing newborn resuscitation procedures have shown to be highly beneficial for maintaining and improving skills17–19 and reduce mortality.20 21 However, review of resuscitation procedure is done after intervention and not during intervention. Therefore, during resuscitation, HCPs depend on their cognitive skills, memory and posted visual reminders for actions to be taken. To mitigate this problem, we are currently in the process to develop an automatic guidance to HCPs during resuscitation with the use of deep learning model, MAchine Learning Application (MALA)22 (figure 1C,D).
MALA will be a tablet-based MALA, which will use video and audio activities recorded by a tablet in analysing the event and will guide for next step of resuscitation through visual display and audio prompts in real time during resuscitation. The development of MALA will require a large number of video and audio recordings to train the MALA and currently a preversion of MALA has been developed which records video and audio activity. This current (pilot) version of the application provides a visual display of time on the tablet monitor mounted onto the infant warmer.
To guide the research team on further development of MALA application, acceptability of video and audio recordings as well as the current version of the application is needed.23–25 Therefore, this study aims to assess the usability, feasibility and acceptability of the current version of technology with visual time guidance, video and audio recordings, during newborn resuscitation.