Technology-Based Newborn Health Learning Initiatives for Facility-Based Nurses and Midwives in Low- and Middle-Income Countries: A Scoping Review

BACKGROUND

It has been argued that the greatest test of health system function is the ability to provide timely care at birth (World Health Organization [WHO], 2005). The reality is that worldwide, of the 140 million babies born each year (WHO, 2017), almost one million newborns die on their first day of life, and close to one million die during their first week of life (United Nations Children's Fund [UNICEF], 2017). An additional one million of the surviving infants develop neurocognitive problems such as cerebral palsy and learning difficulties (Ersdal & Singhal, 2013).¹ An estimated 99% of these deaths occur in low-and middle-income countries. Although some regions have made progress in decreasing neonatal mortality, the highest mortality regions have seen insufficient progress in the last decade. “Early neonatal deaths” are those that occur in the first week of life, while “perinatal deaths” are comprised of early neonatal deaths and intrapartum (“fresh”) stillbirths; “neonatal deaths” are those deaths occurring in the first 28 days of life (Lawn, Lee, et al., 2009). As post-neonatal mortality (in children under 5 years of age) has decreased dramatically over the past 15 years globally, there has been slower progress in reducing the neonatal mortality rate, with neonatal deaths accounting for an increasing proportion of under-5 mortality. In fact, the share of neonatal deaths is projected to increase from 47% of under-five deaths in 2015 to 52% in 2030 (UNICEF, 2017). The three major causes of neonatal deaths are preterm birth complications, intrapartum-related “birth asphyxia,” and infections, which account for more than 85% of newborn deaths (Tuncalp et al., 2015).

The time of greatest risk of mortality and morbidity for both the mother and baby is at birth, when a large percentage of the newborn deaths occur (UNICEF, 2018). Intrapartum-related neonatal deaths, formerly referred to as “birth asphyxia” (i.e., failure to initiate spontaneous respirations and/or 5 min Apgar score of less than 7), are closely linked with maternal complications and death (Lawn, Lee, et al., 2009).² In reality, the vast majority of neonatal deaths can be prevented or treated by relatively straightforward, effective interventions practiced by health workers starting with primary prevention such as skilled attendance at birth and early recognition and timely management of obstetric complications (Chandrasekaran, Thukral, & Deorari, 2014; UNICEF, 2017). Secondary prevention focuses largely on early neonatal care and neonatal resuscitation, while tertiary prevention centers on management of the baby with complications (Lawn, Lee et al., 2009; Ersdal & Singhal, 2013).

The WHO estimates that while 5%–10% of newborns worldwide do not breathe immediately at birth and require some level of assistance to initiate respirations, 3%–6% require basic neonatal resuscitation consisting of simple initial steps and assisted ventilation via bag-mask or face-mask ventilation (WHO, 1998). Less than 0.1% of neonates require advanced resuscitation techniques such as chest compressions and vasoactive medications (Wall et al., 2009). Simple immediate newborn care includes warming, drying, stimulation, hygiene and thermal care. Providing these immediate steps to newborns in all settings is known as essential newborn care, and can even be performed by family members or community members in births at home (Wall et al., 2009). Basic neonatal resuscitation training can be effectively performed by a wide range of health providers, resulting in reductions in intrapartum-related mortality (Lawn, Kinney, et al., 2009). Evidence from several observational studies show that facility-based basic neonatal resuscitation alone may avert 30% of intrapartum-related neonatal deaths in full-term babies, plus 5%–10% in preterm babies (Lawn, Kinney, et al., 2009; Wall et al., 2009).

Despite an increased global emphasis on institutional births, for the estimated 44 million women who give birth annually without skilled care at birth, mostly at home, there is an inverse relationship between access to skilled care at birth and intrapartum-related neonatal mortality (Lawn et al., 2017; Prakash et al., 2016). The coverage gap of skilled birth attendance is widest in certain regions, namely sub-Saharan Africa and South Asia, where baseline coverage is lowest globally (Lawn, Lee et al., 2009However, the place of birth has changed in the last two decades. More than 80% of all births globally now occur with a skilled birth attendant. Although globally in 2016 three in every four live births were attended by a skilled health provider, increases in coverage of facility delivery and skilled attendance at birth have resulted in only limited reductions in maternal and neonatal deaths. This incongruous finding is believed to relate to the inadequate level of care that many women and children receive in health facilities (Marchant et al., 2015). There is, however, still more variability in community settings that pose challenges to effective and sustainable program implementation and scale-up as compared with facility-based settings (Wall et al., 2009).

One key impediment in low- and middle-income countries is clearly the gap in quality coverage and the lack of skilled birth attendants and newborn care providers proficient in essential newborn care and neonatal resuscitation (Ersdal & Singhal, 2013; Wall et al., 2009). Neonatal resuscitation is one of the “signal functions” of emergency obstetric and newborn care (EmONC) that must be provided by all skilled birth attendants (Lund et al., 2016). In reality, low knowledge and confidence in neonatal resuscitation among neonatal providers has been documented, and lack of initial, refresher training has been identified as a key contributor (Rule et al., 2017; Subbiah, Sarin, Jeeva, & Geetanjali, 2012). For babies born in hospitals in South Asia and sub-Saharan Africa, staff often lack training in resuscitation and equipment is not available. Assessments in six African countries showed only 2% to 12% of personnel conducting births in facilities had been trained in neonatal resuscitation, and only 8% to 22% of facilities had equipment for newborn respiratory support. A key challenge is how to seize the missed opportunity to ensure adequate provision of essential newborn care and basic resuscitation for births already occurring in facility settings, including equipment and competent personnel (Wall et al., 2009). Key constraints include cost, availability of competent trainers, and the rapid pace of updates to information, technology, and clinical recommendations (Chandrasekaran et al., 2014; Klunklin, Viseskul, Sripusanapan, & Turale, 2010).

Nurses and midwives are crucial to strengthening health systems and achieving improved maternal and child health (Bell, Rominski, Bam, Donkor, & Lori, 2013; WHO, 2011a). However, particularly in rural areas where there is the lowest coverage of skilled care at birth and neonatal mortality is highest, access to relevant up-to-date learning opportunities is difficult or impossible to obtain (Hudspeth, Curry, Sacks, & Surena, 2015; WHO, 2006). Given the challenges of organizing learning opportunities for health workers in remote settings, expanding mobile phone penetration holds promise for new technological mechanisms to reach these health workers with up-to-date information (Nilsson, Sorensen, & Sorensen, 2014); however, published robust research on newborn health learning initiatives using technology is limited (Amoakoh et al., 2017; Chandrasekaran et al., 2014).

The purpose of this scoping review was to explore what is known from the existing research literature about technology-based newborn health learning initiatives for health workers in low- and middle-income countries, with an emphasis on nurses and midwives. Kirkpatrick's model for the evaluation of training programs (Kirkpatrick, 1996) was used as the guiding framework to consider the various learning initiatives.

METHODS

Data Evaluation and Analysis

The database search identified 576 papers. Duplicates were removed and papers were screened by title and abstract, leaving 36 papers. After full-text review, 27 papers were excluded and nine studies remained. An additional three studies were identified through bibliographic ancestry review, leaving 12 studies for analysis in the final review. Search results are found on the accompanying PRISMA diagram (Figure 1). Kirkpatrick's Model was used to assess training program evaluation in the interventions reviewed (Kirkpatrick, 1996) (Table 1). Kirkpatrick considers evaluation of training programs across four levels: (a) Reaction (evaluates how participants assess the training); (b) Learning (measures if participants acquired knowledge, learned skills, or changed attitudes), (c) Behavior (considers if they are using what they learned at the training on the job), and (d) Results (evaluates if the training positively impacted organization and/or patient outcomes). Kirkpatrick noted that training evaluation becomes more meaningful as it progresses from level 1 to 4, but that ideally all four levels of any training should be evaluated.

FIGURE 1.

PRIMSA flow diagram.

TABLE 1.

Kirkpatrick Levels Measured in Included Articles

FIRST AUTHOR, YEAR	MEASURES	KIRKPATRICK LEVEL
Simulation training with HBB (n = 6) and other training methodologies (n = 3)
Subbiah et al. (2012)	Knowledge and Skills: Structured knowledge questionnaire on NR. Observational checklist on routine care and NR.	2: Learning
Goudar et al. (2012)	Results: 24-hour, 7-day, and 28-day neonatal outcomes.	4: Results
Msemo et al. (2013)	Behavior: Use of stimulation, suctioning and FMV was monitored in pre- and post-HBB intervention births.	3. Behavior
	Results: Primary outcomes were a reduction in ENM within 24 hours and rates of FSBs.	4: Results
Ersdal et al. (2013)	Skills: Two simulation scenarios, “routine care” and “neonatal resuscitation,” were performed by 39 providers, 7 months before (September 2009), and 27 providers, 7 months after (November 2010) HBB training in April of 2010. Two independent raters scored the videotaped scenarios for “pass/fail” performance and different skills.	2: Learning (Skills)
	Behavior: Observational data was collected in the delivery room during 7 months pre (n = 2,745) and 7 months post (n = 3,116) HBB training (April 2010) by observing all hospital births during the same period.	3. Behavior
Seto et al. (2015)	Reaction to training: Pre- and post confidence levels, post-HBB course satisfaction.	1: Reaction
	Knowledge and Skills test: MCQs pre- and post for NR knowledge. Bag-mask ventilation skills checklist (a seven-item checklist of skills to perform effective ventilation) and two OSCEs (A—one for routine newborn care) (B—one complex resuscitation scenario).	2: Learning
Bang et al. (2016)	Knowledge and Skills: Compared pass rates for skills and knowledge pre- and post-initial HBB training and following refresher training among birth attendants. OSCE B tested resuscitation skill retention by comparing post-initial training performance with pre-refresher training performance.	2: Learning
Bellad et al. (2016)	Results: Comparison of pre- and post-HBB training of FSR, numbers of FSBs and 7-day neonatal mortality rate giving composite PMR, among all live registry births in study clusters and registry births in HBB-trained facilities.	4: Results
Eblovi et al. (2017)	Skills: Trainee skills were evaluated by OSCE at three time points: immediately, 4 months after HBB training, and 4 months after the refresher training.	2: Learning
	Behavior: Midwives recorded in logbooks the highest level of NR performed on all newborns delivered for 1 year.	3. Behavior
	Results: The mortality rate in the first 24 hours of life reported by midwives during the 1-year study was compared with nationwide estimates of first 24-hour mortality in Ghana from 2007 to 2013.	4: Results
Vail et al. (2017)	Skills test: Skills evaluated by clinical complexity of the simulated scenario, ranging from level 1, requiring NR without a maternal complication, to level 3, requiring simultaneous management of neonatal and maternal complications. Simulations videorecorded and coded for predefined clinical skills using “Studiocode.”	2: Learning
eHealth/eLearning interventions (n = 3)
Thukral et al. (2012)	Reaction: Participant satisfaction evaluated on a 5-point Likert scale	1: Reaction
	Knowledge and Skills: Knowledge and skill changes were evaluated by administering online MCQs and on-site OSCE stations before and after completion of the course on essential newborn care.	2: Learning
Prakash et al. (2016)	Reaction: Participant satisfaction evaluated on a 5-point Likert scale	1: Reaction
	Knowledge and Skills: Knowledge and skill changes were evaluated by administering online MCQs and on-site OSCE stations before and after completion of the course. Baseline knowledge and skill scores of pediatricians working in SNCUs in the state of Tamil Nadu, India (n = 32) were assessed by 25 MCQs and by five OSCE skill stations.	2: Learning
Lund et al. (2016)	Knowledge and Skills: Knowledge and skills in NR were assessed at baseline and at 6 and 12 months after the intervention. Knowledge and skills assessed via questionnaire and objective structured assessment of technical skills using simulated scenarios with scoring on skills performance on mannequins.	2: Learning
	Results: Perinatal mortality was defined as a composite of a stillbirth or an early neonatal death. Outcome was recorded at delivery.	4: Results

Note. NR = neonatal resuscitation; FMV = face-mask ventilation; ENM = early neonatal mortality; FSB = fresh stillbirth; HBB = Helping Babies Breathe; MCQs = multiple-choice test questions; OSCE = objective structured clinical evaluation; FSR = fresh stillbirth rates; PMR = perinatal mortality rate; SNCU = special newborn care unit.

RESULTS

Twelve studies are included in this review. Studies were conducted in India (n = 5) (Goudar et al., 2012; Prakash et al., 2016; Subbiah et al., 2012; Thukral et al., 2012; Vail et al., 2017), sub-Saharan Africa (n = 4) (Eblovi et al., 2017; Ersdal et al., 2013; Lund et al., 2016; Msemo et al., 2013), both India and Africa (n = 2)(Bang et al., 2016; Bellad et al., 2016), and one study was conducted in Honduras (Seto et al., 2015). All twelve studies were published in English. Studies were all conducted using quantitative methods with the exception of one study that used mixed methods (Prakash et al., 2016). The studies collectively addressed all four of Kirkpatrick's training evaluation levels, and a number of the studies specifically noted the use of Kirkpatrick's model as part of their research approach. The majority of studies evaluated health provider knowledge and skills before and/or after training (n = 9); fewer studies evaluated the effect on change in provider practice (n = 3) and/or patient health outcomes (n = 5) (Kirkpatrick levels 3 and 4). Table 1 presents the consideration of the respective Kirkpatrick levels in the learning interventions' training evaluation methods. Only a few of the included studies utilized more robust study designs, such as randomized controlled trials (RCTs), and validated measurement instruments.

The interventions included in this review fall into two main categories (Tables 2 and 3): (a) Simulation training in routine neonatal care and neonatal resuscitation via Helping Babies Breathe (HBB) courses (American Academy of Pediatrics [AAP], n.d.-a) (n = 6) (Bang et al., 2016; Bellad et al., 2016; Eblovi et al., 2017; Ersdal et al., 2013; Msemo et al., 2013; Seto et al., 2015) and other training programs (n = 3) (Goudar et al., 2012; Subbiah et al., 2012; Vail et al., 2017) and (b) eHealth/eLearning initiatives (defined as the use of information and communication technologies for health and learning, respectively) (n = 3) (Lund et al., 2016; Prakash et al., 2016; Thukral et al., 2012; WHO, 2011b). Tables 2 and 3 present the studies per these categories, respectively.

The studies are presented by category of teaching–learning approach:

DISCUSSION

The purpose of this scoping review was to explore what is known from the existing published research literature about technology-based newborn health learning initiatives for facility-based nurses and midwives in low- and middle-income countries. In this review, we identified a number of studies focusing on simulation training for facility-based health workers using evidence-based training programs such as ENC, NRP, HBB, and PRONTO. Additionally, we identified a limited number of studies utilizing eLearning initiatives, such as Internet-based online learning, and mobile phone mLearning applications. The content of the initiatives focused primarily on essential newborn care, including basic neonatal resuscitation and one study that tested an App targeting interventions for sick newborns (Prakash et al., 2016). Collectively, these interventions center around prevention and treatment of the major causes of neonatal deaths, particularly deaths in the intrapartum period and during the first week of life when intrapartum-related “birth asphyxia” takes a heavy toll. Evaluation study designs ranged from one-group pre–post intervention studies or one-group posttest only (n = 9), to more robust cluster RCTs (n = 2) (Goudar et al., 2012; Lund et al., 2016).

Results reported from simulation training programs (HBB, ENC, NRP, and PRONTO) in this review varied widely, both in terms of skill acquisition and retention, and in terms of provider behavior and patient outcomes. Some studies reported positive and negative results within the same study. The variable results suggest that while short simulation programs improve short-term knowledge and skills and may improve longer-term practice in the simulated setting, the training may be insufficient to enable health workers to integrate and translate those skills into clinical practice. Results also suggest that performance decreases significantly over time and requires refresher training (Bang et al., 2016; Eblovi et al., 2017). These finding are consistent with other literature on the topic (Ersdal & Singhal, 2013), which highlight the importance of programs targeting ongoing refresher training and local mentoring for skill retention in order to impact clinical management and patient outcomes and to increase sustainability. However, supervised practices for skill retention can be challenging and costly in resource-limited settings (Rule et al., 2017).

This review identified a dearth of studies focused on eLearning or other technology-based interventions, which points to the relative lack of peer-reviewed studies on eLearning and mLearning interventions for newborn health workers in low- and middle-income countries, which is consistent with findings reported elsewhere (Agarwal & Labrique, 2014; Chandrasekaran et al., 2014). Thukral et al. (2012) demonstrated that online training is feasible and effective in increasing knowledge and skills and can be blended with local on-site skill learning in partnering institutions. Their findings suggest that this blended approach might be an economical and expeditious way to cover a wide geographical region with uniform standardized education. mHealth initiatives provided updated information to health workers through a training and point of care algorithm-based app for androids (Prakash et al., 2016) and through the Safe Delivery App, a learning resource for training providers on emergency obstetric and newborn care and for point of care use (Lund et al., 2016). Prakash et al. (2016) measured changes in doctors' knowledge and skill levels; Lund et al. (2016) measured changes and retention in birth attendants' knowledge and skill levels, demonstrating that the use of mobile phones is feasible for healthcare workers, irrespective of their educational level or prior training. Lund also found a nonsignificant lower perinatal mortality rate in intervention clusters compared with control clusters subsequent to Safe Delivery App use.

Simulation training is an increasingly utilized methodology for training healthcare providers to manage obstetrical and newborn emergencies such as birth asphyxia (Vail et al., 2017). Simulation-based initiatives in this review primarily used mannequin-based simulation, which has grown in healthcare education and training as a method of teaching, learning, and performance evaluation (Taekman et al., 2017). The advantages of simulation training are that it provides concrete learning opportunities within “realistic” and reproducible contexts (Fuchs, Miller, & Berkowitz, 2009). Engaging providers to synthesize and apply knowledge and tasks according to a scenario, thereby promoting theoretical, cognitive, technical, and behavioral skills is another advantage of simulation training (Ersdal et al., 2013). Simulation is suitable for both pre- and in-service settings, with the potential to provide a continuum of training throughout a professional career (Nelissen et al., 2014). Additionally, simulation allows learners to practice new skills at their own pace, since scenarios can be slowed down to adapt to the needs of the learner. Scenarios can be used to train learners to prepare for emergency situations in settings where no patient will be harmed from the initial attempts of performing a procedure unfamiliar to the learner, and/or in situations that are too rare, serious, or expensive to recreate in real life (Stitely, Cerbone, Nixon, & Bringman, 2011). The disadvantages of simulation training are that it can be costly and requires specific equipment, learners to be physically present, and capable trainers, which may not be a practical option for reaching health workers in the most remote settings.

Learning opportunities are harder to access for clinicians working in rural and remote areas, adding to challenges for remote and rural clinicians that already include professional isolation, inadequate communication with peers in the cities, and a lack of appropriate equipment and technologies that hinder timely and appropriate quality obstetric and neonatal care (Ellard et al., 2016).

mHealth and mLearning, have been linked to expanding mobile phone penetration in low- and middle-income countries. Mobile Learning (M-Learning) provides important opportunities for the delivery of health information and services to remote locations, expanding access by remote health workers to current evidence-based information in order to improve clinical decision-making and management of patients (Agarwal & Labrique, 2014; Obasola, Mabawonku, & Lagunju, 2015). Evidence from systematic reviews and other studies suggest that mHealth initiatives can contribute to improving maternal and newborn health in a variety of ways, for example, mHealth interventions can help by directing health prevention information to women, community workers, and the public, expediting referrals for management of obstetric and newborn health emergencies, monitoring patient adherence to treatment regimens, and conducting surveillance (Amoakoh-Coleman et al., 2016; Lee et al., 2016; Lund et al., 2014; Obasola et al., 2015). Advantages of mHealth initiatives are many, including reduced program cost, practicality in contexts where there are insufficient numbers of available instructors to provide quality training (Thukral et al., 2012), and ease of collaboration with individuals in remote locations at flexible times. Apps are also easy to share and can be updated easily as clinical practice guidelines evolve (Prakash et al., 2016). An additional benefit of mLearning and online learning is the opportunity for providers to participate in self-paced learning with standardized course delivery and monitoring, allowing the learner to tailor their own learning objectives and giving the learner control over content and pace of learning. The findings from Lund et al. (2016) support those of other studies that providers perceive the use of mobile tools as a promising opportunity for healthcare workers' self-improvement with effects on their motivation, self-efficacy, and enthusiasm, in contrast to passive learning models (Chandrasekaran et al., 2014; Labrique, Vasudevan, Kochi, Fabricant, & Mehl, 2013). Disadvantages of eLearning programs include lack of face-to-face interaction, poor Internet connectivity in certain regions, non-user-friendly interfaces, and digital illiteracy (Chandrasekaran et al., 2014).

The feasibility of mLearning has largely advanced beyond earlier barriers of bandwidth and the cost of Internet access (Obasola et al., 2015). However, despite the rapid growth of mHealth and eHealth programs, research and evidence on their effectiveness are scarce (Agarwal, Perry, Long, & Labrique, 2015; Batavia & Kaonga, 2014; Sondaal et al., 2016; Tomlinson, Swartz, Rotheram-Borus, & Tsai, 2013). More specifically, there is a lack of trials of mobile phone and other eHealth interventions in low- and middle-income countries that have health outcomes, or level-4 Kirkpatrick findings, as a primary outcome (Lund et al., 2016; Obasola et al., 2015). One review hypothesized that the lack of evidence related to mHealth may be because organizations piloting mHealth projects in the field are perhaps more focused on implementation of mHealth programs, and are opting to publish open-source gray literature and white papers rather than publishing in peer-reviewed papers (Batavia & Kaonga, 2014). Other challenges to overcome include a lack of standardization of eLearning evaluation methods (Chandrasekaran et al., 2014), using validated and reliable measurement instruments.

Our identification of studies in this review was limited by the selected databases and search terms that we used, which may have resulted in relevant studies being missed from other databases or from the use of associated search terms. Additionally, we were limited to analysis of results reported in the selected studies, which made comparisons between learning interventions difficult due to lack of comparable data. Future research on simulation training and eHealth/mHealth should emphasize use of robust study designs, using valid and reliable measurement instruments, and measuring higher-level outcomes (Kirkpatrick levels 3 and 4) such as behavior change and patient results (Ameh & van den Broek, 2015; Amoakoh-Coleman et al., 2016, Chandrasekaran et al., 2014). Feasibility studies in remote areas are needed to determine if certain technology-enhanced approaches are viable for the specific context or country, as well as cost analyses, to identify potentially scalable approaches leading to better coverage of skilled birth attendants to enhance proficiency in essential newborn care and neonatal resuscitation.

CONCLUSION

Given the high costs and logistical challenges of in-person training programs in low-income settings, the healthcare workers who most need support are often the least likely to receive it. Studies examined here propose innovative technology-based learning initiatives to address gaps in resource-poor locations. Learning approaches that emphasize simulation training and eHealth approaches for maternal and newborn health workers hold promise for identifying potentially scalable approaches toward establishing better coverage of skilled birth attendants with proficiency in essential newborn care and neonatal resuscitation. This is especially critical for health workers in more remote areas, where maternal and newborn mortality are the highest. However, existing simulation literature demonstrates that translation of knowledge gained during simulation into improved clinical outcomes in real births is difficult and variable. The volume of evidence demonstrating the potential benefit of mHealth strategies, especially in the neonatal period, is limited. Learning initiatives with high priority content and demonstrated effectiveness should be tested using rigorous research methods to determine long-term effectiveness and to measure cost-effectiveness in different contexts and priority areas. However, mHealth and blended learning initiatives combining distance learning and simulation training hold many advantages. Future research is suggested documenting patient outcomes due to technology-based learning interventions, in order to enhance the quality of care of nurses and midwives in maternal and newborn health.

Disclosure

The authors have no relevant financial interest or affiliations with any commercial interests related to the subjects discussed within this article.