Summary
- What is the program? Programs that train health workers on facility-based maternal and neonatal health (MNH) interventions are designed to prevent maternal and perinatal mortality, including stillbirths and neonatal mortality, by increasing coverage of evidence-based care to women and newborns through health facilities.
- What is its evidence of effectiveness? Eight randomized controlled trials (RCTs) find that on average, these types of interventions reduce maternal and perinatal mortality among women delivering at facilities. There is substantial variation across these trials in program design, setting, and effects on key outcomes. We are also highly uncertain about the quality with which the program is implemented at large scale and in real-world settings. Both of these limit our confidence in the extent to which the results of these studies apply to any specific program conducted by a charity.
- How cost-effective is it? Our best guess is that training for facility-based MNH interventions is within the range of cost-effectiveness of programs we would consider directing funding to. The high cost-effectiveness is driven by high baseline rates of neonatal mortality and stillbirths and high effectiveness of the program in reducing these outcomes. However, our cost-effectiveness estimate of this intervention relies on several assumptions about which we are highly uncertain, including the cost of implementation and the extent to which the trial findings extend to specific programs and settings we would consider.
- Does it have room for more funding? We expect that there is likely substantial room for more funding in this area, particularly in Sub-Saharan Africa and South Asia, where neonatal mortality and maternal mortality rates remain high.
- Bottom line: Training for facility-based MNH interventions appears to be promising, and we are investigating opportunities to direct funding to these programs. Because we are highly uncertain about the effectiveness of specific programs conducted by charities, we may consider including an evaluation of any program to which we recommend a grant.
Published: July 2022
What is the problem?
In 2019, around 2.4 million infants died in the first month of life,1
and neonatal mortality rates in Sub-Saharan Africa and South Asia remain particularly high.2
Many of these deaths are concentrated during the first week of life.3
In 2017, around 295,000 women died from complications arising from pregnancy and childbirth, with 86% of those deaths occurring in Sub-Saharan Africa and South Asia.4
What is the program?
The program is to train health workers to deliver a package of maternal and neonatal health (MNH) interventions in health facilities during and following birth, up to discharge. One training model is low-dose, high-frequency training (LDHF), in which short, hands-on training and ongoing practice sessions are provided at repeated intervals. This is hypothesized to result in improved skill acquisition and health provider performance.5 Packages of MNH interventions are generally aimed at preventing neonatal and maternal morbidity and mortality, and they can include:6
- Essential care for labor and birth, including: management and monitoring of labor, early identification of complications, and infection prevention during labor.7
- Essential newborn care, including: supportive breastfeeding, temperature regulation, and clean cord care and other hygienic practices.8
- Applying chlorhexidine to umbilical cords stumps.9
- Care of small and sick newborns, including: skin-to-skin care and other forms of temperature regulation, supportive breastfeeding, and referral to higher-level facilities.10
- Emergency obstetric care, including: interventions to prevent post-partum hemorrhage such as uterotonics,11 treatment for pre-eclampsia and eclampsia,12 and improving maternal and neonatal survival for pregnant women at risk of delivering a preterm baby.13
Does the program have strong evidence of effectiveness?
Overall, based on a medium-depth literature search, we believe there is moderately strong evidence that training health workers to deliver maternal and neonatal care interventions in facilities reduces neonatal mortality, stillbirths, and maternal mortality. We guess this is driven by increased uptake of interventions, such as management and monitoring of labor, essential newborn care, and emergency obstetric care, which we believe reduce neonatal mortality, stillbirths, and maternal mortality.
However, heterogeneity in program design underlying these studies as well as concerns over the quality with which a program would be conducted in real-world settings lead us to be highly uncertain about the extent to which the results of these studies apply to any specific program conducted by a charity.
Evidence of effectiveness on neonatal outcomes
Based on seven cluster-RCTs, there is moderately strong evidence that training health workers to deliver neonatal care interventions in facilities reduces perinatal mortality, including fresh stillbirths and neonatal deaths before discharge. We have not reviewed individual studies in depth.
We have done a rough subjective weighting of these cluster-RCTs,14 based on how relevant we guess those studies are to programs we may consider funding.15 Given this weighting, evidence from these seven trials suggests that training health workers in maternal and neonatal care reduces neonatal deaths before discharge by around 33%16 (with trials reporting effect sizes ranging from a non-statistically significant 3% increase to a significant 70% reduction in neonatal mortality before discharge)17 compared to no training. Evidence from 4 trials suggests that training reduces stillbirths by around 35%18 (with trials reporting effect sizes ranging from a non-significant 5% increase to a significant 52% reduction in stillbirths rate).19
The trials vary in terms of training model,20 content covered by training,21 and what additional support was given outside of training health workers.22 Given the heterogeneity in program design and effect sizes in these trials, we are highly uncertain about the appropriate weighting of different studies.
In particular, we put a relatively large weight (30%) on one study, Gomez et al. 2018, which shows a large reduction in neonatal mortality and stillbirths. We think the large effect sizes found in this study are plausible because it used a LDHF training model, in which training is more practice-based and conducted at a higher frequency than traditional training models,23 and because it used a particularly comprehensive package of maternal and neonatal interventions.24 We put the most weight on Gomez et al. 2018 because we think the intervention studied is closest to the programs we would consider funding. However, it is possible that the larger effect size found in this study is due to statistical variability or particular characteristics of the intervention in the study that won’t apply to programs that we consider funding.
Three non-experimental, pre-post studies on LDHF training for maternal and neonatal care find similar effect sizes on neonatal mortality and stillbirths.25 Because there may be biases in non-experimental evidence that are difficult to fully account for, we do not put weight on these non-experimental studies in our estimate of the pooled effect size. However, the non-experimental evidence increases our confidence in the effects found in the trials, and we factor it into our internal validity adjustments.
We plan to revisit the weighting on studies when considering any specific program conducted by a charity.
Evidence of effectiveness on maternal outcomes
Based on three cluster-RCTs, there is weak evidence that training health workers to deliver maternal care interventions in facilities reduces maternal mortality.
We have done a rough weighting of these three cluster-RCTs26 based on how relevant we guess those studies are to programs we may consider funding.27 Evidence from these three trials suggests that training health workers in maternal care reduces maternal deaths before discharge by around 32%28 (with trials reporting effect sizes ranging from a non-significant increase of 11% to a significant 89% reduction in maternal mortality)29 compared to no training. The low number of studies reporting maternal all-cause mortality and the high heterogeneity in effect sizes limits our confidence in this effect.
In addition, we found six non-experimental studies (five pre-post studies and one observational study with a non-randomized control group) that find fairly large reductions in maternal mortality and postpartum hemorrhage incidence rates.30 As with neonatal outcomes, we do not put weight on these non-experimental studies in our estimate of the pooled effect size because of concerns over potential biases in the reported effects. However, the non-experimental evidence increases our confidence in the effects found in the trials, and we factor it into our internal validity adjustments.
Overall, we believe that it’s likely that training on packages of maternal and neonatal care reduces maternal mortality, but we are highly uncertain about the magnitude of the effect size on maternal mortality. As such, we’ve applied a fairly strict internal validity adjustment to the pooled effect size (see below).
Evidence on effect of practices promoted
Training programs on maternal and neonatal care aim to promote specific practices intended to reduce maternal and perinatal mortality.
Gomez et al. 2018 found that LDHF training increased retention of knowledge of clinical content and skills that included infection prevention and control, support of labor and birth, immediate newborn care, newborn resuscitation, management of postpartum hemorrhage, and kangaroo mother care (KMC).31 We guess that the uptake of these best practices for emergency care, infection prevention, etc. also increased as a result of the training.
We have also conducted a shallow review of some trials showing the effect of the individual practices promoted on these outcomes, which give us additional confidence that a package of these interventions reduces maternal and perinatal mortality:
- Chlorhexidine: GiveWell’s meta-analysis of five cluster-RCTs finds that cord cleansing with chlorhexidine leads to a 15% reduction in neonatal mortality (see our intervention report).
- Community-based maternal and neonatal packages: A 2015 Cochrane meta-analysis of 26 randomized and quasi-randomized trials finds that community-based packages of maternal and neonatal care, which include interventions similar to the facility-based packages of care covered in this report,32 are associated with 25% reduction in neonatal mortality, 19% reduction in stillbirths, and 20% reduction in maternal mortality (see our intervention report).
- Breastfeeding: A 2013 meta-analysis of three RCTs finds that breastfeeding promotion is associated with a 31% reduction in diarrhea,33 which we expect to lead to reductions in diarrhea mortality (see our intervention report).
- KMC, which includes skin-to-skin contact and early initiation of breastfeeding for low birthweight infants: A 2016 Cochrane meta-analysis consisting of 12 studies finds that KMC reduces neonatal mortality among low birthweight infants by around 30%34 (see our intervention report). However, our impression, based on reviewing case studies of KMC, is that KMC would be difficult to implement with high quality in real-world settings, and we have those same concerns with any training package that includes components of KMC.
- Oxytocin for prevention of postpartum hemorrhage: A 2019 meta-analysis of 24 trials finds that prophylactic oxytocin used during the third stage of labor was associated with a 49% reduction in risk of postpartum hemorrhage.35
However, many of these individual maternal and neonatal interventions address the same overarching risk factors (e.g., reducing the risk of infection, protecting against intrapartum or preterm birth complications), and it’s unclear how these effects might aggregate when implemented together in a package. As such, we put the most weight on the trials measuring the effect of training for packages of interventions on these outcomes.
Internal and external validity
On the whole, we view this evidence to be moderately strong due to the size of the evidence base.
We have a high level of uncertainty about the extent to which findings from trials will generalize to future implementation, though we haven’t explored this issue in depth yet. In particular, we’re uncertain about how the training model (including frequency, duration, and intensity of training), training content, and other aspects of support in programs we may recommend funding to will compare with the programs conducted in the studies. Moreover, we think it’s likely that there are additional challenges to implementing the program with high quality in real-world settings and at a large scale that may not be reflected in the evidence base.
We’ve chosen fairly strict external validity adjustments to reflect our guess that any program conducted by a charity in a real-world setting is likely significantly less effective than the programs studied in the trials.
See here for our internal and external validity adjustments and rationale.
How cost-effective is the program?
We conducted a preliminary cost-effectiveness analysis. We found that training of health workers in facility-based maternal and neonatal care has a cost-effectiveness within the range of programs we would recommend funding.
Note that our cost-effectiveness analyses are simplified models that do not take into account a number of factors. There are limitations to this kind of cost-effectiveness analysis, and we believe that cost-effectiveness estimates such as these should not be taken literally due to the significant uncertainty around them. We provide these estimates (a) for comparative purposes and (b) because working on them helps us ensure that we are thinking through as many of the relevant issues as possible.
This cost-effectiveness analysis is at an early stage, and we think it’s likely that our bottom line cost-effectiveness estimate will change with further review.
Neonatal mortality and stillbirth rates are quite high in low-resource settings, and conducting training of facility-based MNH care leads to substantial declines in neonatal mortality, stillbirths, and maternal mortality. This leads to high cost-effectiveness.
A sketch of the cost-effectiveness model is below:
- Prevalence of the problem: Across ten high-burden countries, we estimate that there are an average of approximately 32 neonatal deaths (of which 25 occur in the first week of life), approximately 12 fresh stillbirths, and approximately 6.3 maternal deaths per 1,000 facility births.36
- Effect of the intervention on reducing the problem: Based on our evidence review, we estimate that training health workers in the package of MNH facility-based care leads to:
- IV/EV: We apply downward adjustments to the effect sizes found in the literature to account for potential biases in the evidence base and our guess that the program was conducted with higher quality in RCTs than it would be in real-world settings. The size of these adjustments is 62% for early neonatal mortality, 58% for fresh stillbirths and 36% for maternal mortality.40
- Additional benefits: We include a rough overall adjustment of 125% to account for child disability and developmental effects, infant and maternal morbidity, cost savings from averting treatment, and increased clinic utilization.41
- Cost of the program: We estimate that it costs roughly $15 to an NGO implementer and $21 to all contributors (NGO and government) per facility birth covered.42 NGO costs may include costs of designing and conducting training sessions for health workers, supportive supervision, and improving infrastructure or providing equipment necessary for the program. Government costs may include things like health worker time to attend training sessions and deliver maternal and neonatal care, and space and equipment provided by health facilities for this program.43
- Cost-effectiveness: Using an estimate of the overall cost to all contributors of $21 per delivery, our best guess is that training of health workers in the package of MNH facility-based care averts a death (including neonatal deaths, stillbirths, and maternal deaths) for an average of $3,024.44
However, we have high uncertainty about the costs (both to an NGO and to governments) of a program that trains health workers in facility-based maternal and neonatal care, as well as the effect size of a program in practice and baseline mortality rates in different contexts.
- Effect size.
- We expect the treatment effect size to differ significantly between implementers of different programs. As noted above, the literature on facility-based MNH worker training involves very heterogeneous training content, frequency and methods, and the programs in the trials may also be implemented with higher quality than would be feasible at larger scale and in real-world settings. If we investigate a specific opportunity, we will refine our effect-size estimates according to how similar the opportunity is to interventions in the literature.
- Moreover, we are uncertain about how the reductions in pre-discharge neonatal mortality reported in the studies translate to reductions in overall neonatal deaths. We currently assume that the estimated reduction in mortality applies to counterfactual deaths that would occur during the first week of life. Although most newborns are discharged from hospital sooner than this, we think it is likely that improved newborn healthcare likely averts some deaths that would have occurred soon after discharge. It is also possible that averting intrapartum or early neonatal complications provides a protective effect that averts late neonatal deaths. This possibility is not currently represented in our model and we are unsure how large this effect may be.
- Program costs. We have high uncertainty about the cost estimates that we use. We used estimates from one academic cost-effectiveness analysis, but we are very unsure about how generalizable these estimates are across different non-trial settings. In particular, the costs per birth found in the analysis varied considerably depending on the number of births per facility.45 While we have limited confidence in these estimates, we believe they are the best ones at our disposal at this stage of our investigation. We expect to be able to refine our estimates if we investigate a specific giving opportunity.
- Costs to the government and health facilities. We are similarly unsure about the share of costs that would be incurred by governments and health facilities, which may include things like government staff capacity for conducting trainings of health workers, health worker time to attend trainings and deliver maternal and neonatal care, and space provided by facilities for this program. As such, we do not have a strong sense of the extent to which potential leveraging and funging effects will impact our cost-effectiveness analysis. We are also uncertain about the extent to which the program might increase the workload of healthcare providers, which may be harmful in settings where there is already a shortage of staff.
- Baseline mortality. We are moderately uncertain about the baseline facility mortality rates across different contexts. We have adjusted national stillbirth, neonatal, and maternal mortality rates by some constant factors to estimate facility-based fresh stillbirth, early neonatal, and maternal mortality rates.46 It is likely that these adjustment factors are different across countries, but we think they are a reasonable best-guess at this stage. If baseline mortality rates are higher than expected in our model, the program would be more cost-effective than currently estimated (and vice versa).
Is there room for more funding?
We expect that there is likely substantial room for more funding in this area, particularly in Sub-Saharan Africa and South Asia, where neonatal mortality and maternal mortality rates remain high. Among the 10 countries with highest total facility-based neonatal mortality, we estimate total annual room for more funding to be around $470 million, assuming a cost of $15 per childbirth targeted.47
How feasible is implementation of the program?
This seems to be a fairly complex program, and we are uncertain about how challenging it is to implement the program with high quality on a large scale. We plan to investigate this further as we talk with implementers about their specific programs.
Key questions for further investigation
Our biggest uncertainties for this opportunity are around the effectiveness of any particular program in reducing maternal and perinatal mortality in the particular settings we would consider supporting and the costs of the program, both to an NGO implementer and other partners.
We plan to identify and talk to charities currently implementing facility-based MNH programs to understand:
- What is the nature of their support? Including:
- What are the barriers to high-quality maternal and neonatal care (e.g. Are there gaps in healthcare providers' skills and knowledge or other factors related to the ability of health facilities to implement these practices)? How would their support be addressing these barriers?
- What does their support look like at the national, facility, and health-worker level?
- Training model: Who would conduct the training for health workers? What’s the frequency, duration, and intensity of training? What are their plans for ongoing supportive supervision or coaching?
- Are they providing support to ensure that the necessary commodities, equipment, and infrastructure are in place for the program to be conducted successfully in health facilities?
- Which interventions are included in their package of care? How much weight should we give to different studies in our estimates of mortality effects given variation in program implementation and setting?
- How does the reduction in pre-discharge neonatal deaths relate to the reduction in overall neonatal mortality due to the intervention?
- Where are there opportunities to expand the program?
- What is the quality of care when this program is implemented in real-world settings?
- What are barriers and risks to scaled implementation?
- What counterfactual care would infants receive in the absence of this training in settings where we would consider funding?
- What are the costs of a facility-based MNH training program?
- What are the costs saved or borne by a hospital that participates in MNH training? What are other costs borne by the government or other partners in implementing this kind of program?
Sources
- 1
“Globally 2.4 million children died in the first month of life in 2019.” World Health Organization, "Newborns: improving survival and well-being," 2020
- 2
“Sub-Saharan Africa had the highest neonatal mortality rate in 2019 at 27 deaths per 1,000 live births, followed by Central and Southern Asia with 24 deaths per 1,000 live births. A child born in sub-Saharan Africa or in Southern Asia is 10 times more likely to die in the first month than a child born in a high-income country.” World Health Organization, "Newborns: improving survival and well-being," 2020
- 3
“The majority of all neonatal deaths (75%) occurs during the first week of life, and about 1 million newborns die within the first 24 hours.” World Health Organization, "Newborns: improving survival and well-being," 2020
- 4
“Maternal mortality is unacceptably high. About 295 000 women died during and following pregnancy and childbirth in 2017…
Sub-Saharan Africa and Southern Asia accounted for approximately 86% (254 000) of the estimated global maternal deaths in 2017.” World Health Organization, "Maternal mortality," 2019
- 5
"LDHF is a capacity-building approach that promotes maximal retention of clinical knowledge, skills, and attitudes through short, targeted in-service simulation-based learning activities, which are spaced over time and reinforced with structured, ongoing practice sessions on the jobsite," @Jhpiego, Low Dose, High Frequency: A Learning Approach to Improve Health Workforce Competence, Confidence, and Performance, 2016@, pg. 2.
- 6
See GiveWell, Facility-based MNH Evidence Review, 2022, "Content Covered" columns for components of care included in the studies we reviewed.
- 7
For example, see Jhpiego, Modules at a Glance: Essential Care for Labor & Birth, 2020: “Helping Mothers Survive Essential Care for Labor & Birth (HMS ECL&B) is a learning module designed to improve the quality and experience of care for women and newborns during labor, birth and the immediate postpartum period.”
“Key competencies covered:
Respectful care
Infection prevention
Classification of labor: latent, active or false labor
Management and monitoring of labor
Early identification of complications
Identification of poor progress of labor
Supportive care during all stages of labor and birth" - 8
"Studies have shown that many newborn lives can be saved by the use of simple low technological interventions. Interventions such as:
supporting breastfeeding
providing adequate warmth
ensuring good hygiene and cord care,
recognizing early signs of danger and providing prompt treatment and
referral, giving extra care to small babies, and
having skilled health workers attend mothers and babies at delivery
and in the immediate postpartum period can all increase a newborn baby’s chances of survival.” @World Health Organization, "Newborn Health"@. - 9
“Umbilical cord hygiene prevents sepsis, a leading cause of neonatal mortality. The World Health Organization recommends 7.1% chlorhexidine digluconate (CHX) application to the umbilicus after home birth in high mortality contexts. In Bangladesh and Nepal, national policies recommend CHX use for all facility births,” Zaman et al. 2021.
- 10
“Some newborns require additional attention and care during hospitalization and at home to minimize their health risks.
Low-birth-weight and preterm babies:
If a low-birth weight newborn is identified at home, the family should be helped in locating a hospital or facility to care for the baby.
increased attention to keeping the newborn warm, including skin-to-skin care, unless there are medically justifiable reasons for delayed contact with the mother;
assistance with initiation of breastfeeding, such as helping the mother express breast milk for feeding the baby from a cup or other means if necessary;
extra attention to hygiene, especially hand washing;
extra attention to danger signs and the need for care; and
additional support for breastfeeding and monitoring growth.Sick newborns
Danger signs should be identified as soon as possible in health facilities or at home and the baby referred to the appropriate service for further diagnosis and care;
If a sick newborn is identified at home, the family should be helped in locating a hospital or facility to care for the baby.”
World Health Organization, "Newborns: improving survival and well-being," 2020 - 11
“Bleeding after Birth Complete 2.0 is a module designed to reduce maternal deaths caused by post-partum hemorrhage and incorporates the latest WHO recommendations for use of tranexamic acid (TXA), uterotonics for prevention such as carbetocin and heat stable carbetocin, (HSC). The module builds capacity of the entire team of health workers who care for women at birth. With HMS BABC 2.0, skills are built around communication and teamwork, active management of the third stage of labor, early detection and management of PPH including management of shock, uterine balloon tamponade, manual removal of the placenta, repair of cervical lacerations, and use of the non-pneumatic anti-shock garment.” Jhpiego, Bleeding after Birth Complete 2.0, 2022
- 12
“Pre-eclampsia & Eclampsia is a module designed to address the second leading cause of maternal mortality globally, pre-eclampsia & Eclampsia. Key content in this module includes: correct assessment and classification of hypertensive disorders of pregnancy, administration of loading and maintenance doses of magnesium sulfate and antihypertensive medications, and management of convulsions.” Jhpiego, Pre-eclampsia & Eclampsia: Managing hypertensive disorders of pregnancy, 2022
- 13
“Threatened Preterm Birth Care is part of the Helping Mothers Survive and Helping Babies Survive suite of training modules. It is a two-day course designed for teams of health workers who provide care for women and newborns. The content is based on the 2015 WHO recommendations on interventions to improve preterm birth outcomes and focuses on identifying women likely to deliver a preterm baby within 7 days and the actions that can be taken prior to birth to improve survival. It is designed to be used at higher level facilities where trained staff and supplies are available to:
Accurately estimate gestational age (GA)
Accurately diagnose conditions leading to preterm birth
Provide advanced care for preterm newborns, including resuscitation, thermal care, feeding support, infection treatment, and safe oxygen use
Reliably identify and treat maternal infections” Jhpiego, Threatened Preterm Birth Care: Improving Survival for Preterm Infants, 2022 - 14
See GiveWell, Facility-based MNH Evidence Review, 2022 for details on study design, effect sizes and their confidence intervals, and the weights we use the pool effect sizes.
- 15
Our guess is that we’d be most likely to fund programs that provide training on a more comprehensive package of interventions, use a training model with more practice-oriented and frequent training sessions on clinical skills (such as a LDHF training model), and include key practices recommended by WHO such as essential newborn care. We put more weight on studies with programs that were closer to these criteria.
- 16
See GiveWell, Facility-based MNH Evidence Review, 2022, "Weighting of Evidence" sheet, "Justification for weight on study" column. Most of the studies we put the most weight on report neonatal deaths before discharge or within 24 hours. See “Timing of reported mortality outcomes” column in GiveWell, Facility-based MNH Evidence Review, 2022, for more information.
- 17
See "Effect sizes (% reduction)" column, GiveWell, Facility-based MNH Evidence Review, 2022.
- 18
See GiveWell, Facility-based MNH Evidence Review, 2022, "Weighting of Evidence" sheet, "Justification for weight on study" column.
- 19
See GiveWell, Facility-based MNH Evidence Review, 2022, "Weighting of Evidence" sheet, "Effect sizes (% reduction)" column.
- 20
Training models differed in terms of how practice-oriented the training was, as well as the frequency and duration of training and practice sessions (ranging from daily or weekly practice sessions to no follow-up sessions after initial training). See “Description of Intervention,” “Initial training,” and “Frequency of follow up practice sessions” columns in GiveWell, Facility-based MNH Evidence Review, 2022.
- 21
- 22
For instance, some studies provided equipment to facilities such as simulators, resuscitation equipment, and delivery kits. See “Equipment supplied.”
- 23
“Traditional in-service training typically takes place in a workshop setting, where providers leave their workplaces for variable lengths of time to receive didactic classroom training in a large group. (For example, basic emergency obstetric care or essential obstetric care is often taught over two or three weeks, and HIV/AIDS/infectious disease is taught over one or two weeks.)...
…This review identified that didactic instruction, such as relying on reading and lecture, often results in no to low learning outcome. In contrast, the use of interactive, practice-heavy techniques, such as clinical simulation, case-based learning, hands-on practice with anatomic models, and immediate feedback on performance, results in better learning outcomes. In addition, repeated frequency is preferable to one-time training interventions; workplace learning may be superior for skill acquisition, and multiple media modes of delivery can be used to deliver training more efficiently.
LDHF is a capacity-building approach that promotes maximal retention of clinical knowledge, skills, and attitudes through short, targeted in-service simulation-based learning activities, which are spaced over time and reinforced with structured, ongoing practice sessions on the jobsite,” @Jhpiego, Low Dose, High Frequency: A Learning Approach to Improve Health Workforce Competence, Confidence, and Performance, 2016@, pgs. 1-2.
- 24
See columns X-AK.
- 25
See “LDHF Training” section of the “Non-experimental studies” tab. Average risk ratio effect size among these pre-post studies is 0.54 and 0.59 for neonatal mortality and stillbirths, respectively.
- 26
See "Randomized studies" sheet in the GiveWell, Facility-based MNH Evidence Review, 2022 for details on study design, effect sizes and their confidence intervals, and the weights we use the pool effect sizes.
- 27
Our guess is that we’d be most likely to fund programs that provide training on a more comprehensive package of interventions and use a training model with more practice-oriented and frequent training sessions on clinical skills (such as a LDHF training model). We put more weight on studies with programs that were closer to these criteria.
- 28
See GiveWell, Facility-based MNH Evidence Review, 2022, "Weighting of evidence" sheet, "Justification for weight on study" column.
- 29
See GiveWell, Facility-based MNH Evidence Review, 2022, "Weighting of Evidence" sheet, "Effect sizes (% reduction)" column.
- 30
See GiveWell, Facility-based MNH Evidence Review, 2022 for maternal outcomes reported in these studies.
- 31
- See Table 2, “Low-Dose Session Content. Lists the clinical content presented in each low-dose session," on page 3 of Gomez et al. 2018 for the content of the low-dose training sessions.
- “Objective structured clinical examination (OSCE) tools and knowledge tests, based on global and national guidelines, were developed to assess SBAs in management of normal birth, immediate newborn care, and maternal and newborn complications. These assessments were used before and immediately after low-dose sessions, and again 1 year later, to evaluate knowledge and skill acquisition and retention,” Gomez et al. 2018, pg. 3.
- “Pre-test scores on the low-dose session 1 (LD1) knowledge test and OSCE were 76% and 44%, respectively (Table 6). Low-dose session 2 (LD2) pre-test scores were slightly higher at 91% and 52%, respectively. Post-test scores increased by 11% (95% confidence interval [CI] 9–12; p < 0·001) and 4% (95% CI 2–6; p = 0·001) on LD1 and LD2 knowledge tests, respectively, and 44% (95% CI 40–48; p < 0·001) and 40% (95% CI 37–43; p < 0·001) on the LD1 and LD2 OSCEs, respectively. After 1 year, participants retained most knowledge and skills gains. Participants scored 31% and 28% higher on OSCE content from LD1 and LD2 at 1 year, respectively, compared to pre-test scores (LD1 95% CI 27–36; p < 0·001; LD2 95% CI 25–32; p < 0·001), and 8% and 2% higher, respectively, on knowledge content from LD1 and LD2 at 1 year, compared to pre-test scores (LD1 95% CI 6–10; p < 0·001; LD2 95% CI 0–4; p = 0·025),” Gomez et al. 2018, pgs. 6-7.
- 32
Intervention packages included in this evidence assessment may include some combination of:
- Basic antenatal, natal, and postnatal care
- Preventive essential newborn care including: ensuring warmth, immediate skin-to-skin care, early breastfeeding, umbilical cord care, eye care, Vitamin K administration, and immunization
- Breastfeeding counseling
- Management and referral of sick newborns
- Skills development in behavior change communication
- Community mobilization strategies to promote birth and newborn care preparedness.
See GiveWell's intervention report on community-based packages of maternal and neonatal care for more information.
- 33
"[W]e identified three randomized trials of breastfeeding promotion; diarrhea morbidity was lower in the group receiving the intervention [pooled relative risk: 0.69 (95% confidence interval: 0.49; 0.96)]." @Horta and Victora 2013@, p. 2.
- 34
- ”Kangaroo mother care was associated with a statistically significant reduction in risk of mortality at discharge or at 40 to 41 weeks’ postmenstrual age (3.2% vs 5.3%; RR 0.60, 95% CI 0.39 to 0.92; I2 = 0%; NNTB = 47, 95% CI 31 to 236; eight trials, 1736 infants) (Analysis 1.1), and at latest follow-up (4.0% vs 6.0%; RR 0.67, 95% CI 0.48 to 0.95; I2 = 0%; NNTB = 50, 95% CI 32 to 331; 12 trials, 2293 infants; moderate-quality evidence) (Analysis 1.4) (Figure 3).” Conde-Agudelo and Díaz-Rossello 2016, Pg. 16.
- ”Between enrolment and 28 days, 73 infants died in 4423 periods of 28 days in the intervention group and 90 deaths in 3859 periods of 28 days in the control group (hazard ratio [HR] 0·70, 95% CI 0·51–0·96; p=0·027). Between enrolment and 180 days, 158 infants died in 3965 periods of 180 days in the intervention group and 184 infants died in 3514 periods of 180 days in the control group (HR 0·75, 0·60–0·93; p=0·010). The risk ratios for death were almost the same as the HRs (28-day mortality 0·71, 95% CI 0·52– 0·97; p=0·032; 180-day mortality 0·76, 0·60–0·95; p=0·017).” Mazumder et al. 2019, Pg. 1.
- For standard KMC at latest follow-up, 1 minus RR 0.67 equals 0.33 (i.e., a reduction of 33%). For community-based KMC at 180-day follow-up, 1 minus RR 0.76 equals 0.24 (i.e., a reduction of 24%).
- 35
“This review includes 24 trials, with 23 trials involving 10,018 women contributing data…
Prophylactic oxytocin compared with no uterotonics or placebo may reduce the risk of blood loss of 500 mL after delivery (average risk ratio (RR) 0.51, 95% confidence interval (C) 0.37 to 0.72; 4162 women; 6 studies; Tau² = 0.10, I² = 75%; low-quality evidence), and blood loss 1000 mL after delivery (RR 0.59, 95% CI 0.42 to 0.83; 4123 women; 5 studies; low-quality evidence).” Salati et al. 2019.
- 36
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet. For neonatal deaths, a simple average was taken of the values in the "Baseline NMR per 1000 live facility births” row. For fresh stillbirths, a simple average was taken of the product of the values in the "Baseline stillbirth rate (facility births)" row, the values in the “Fresh stillbirth as percent of all stillbirths” row, and 1,000. For maternal deaths, a simple average was taken of the values in the "Baseline MMR per 100,000 facility births" and divided by 100.
- 37
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet, “% pre-discharge neonatal mortality reduction, adjusted” row.
- 38
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet, “% fresh stillbirth mortality reduction, adjusted” row.
- 39
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet, “% maternal mortality reduction, adjusted” row.
- 40
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet, rows 33-34, 47-48 and 63-64 respectively. For early neonatal mortality, .8 (IV) x .77 (EV) = ~.62. For fresh stillbirths averted, .75 (IV) x .77 (EV) = ~.58. For maternal mortality, .6 (IV) x .6 (EV) = .36.
- 41
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet, “Total adjustment factor for excluded effects” row.
- 42
See GiveWell, Facility-based MNH BOTEC, 2022, “Supplementary” sheet, “Cost per delivery” row. We estimate $15 per delivery based on the number of births covered by the Gomez et al. 2018 trial and related costs as detailed in Willcox et al. 2017. We guess that NGO costs account for 70% of total program costs.
- 43
See Willcox et al. 2017 for a discussion of program costs.
- 44
See GiveWell, Facility-based MNH BOTEC, 2022, “CEA” sheet. We used a simple average of the values in the “Cost per death averted, after downside adjustments” row.
- 45
Across four different study waves in Gomez et al. 2018, we estimate (based on costs as detailed in Willcox et al. 2017) that costs per birth ranged from $9.75 to $26.07 (in inflation-adjusted terms). See GiveWell, Facility-based MNH BOTEC, 2022, “Supplementary” sheet.
- 46
According to a meta-analysis by Tura, Fantahun, and Worku 2013, facility neonatal mortality rates are around 29% lower than home birth neonatal mortality rates: "Based on the random effects model, the final pooled effect size in the form of relative risk was 0.71 (95% CI: 0.54, 0.87) for health facility delivery as compared to home delivery." We apply this adjustment factor to estimate facility-based neonatal mortality rates. We also apply this adjustment factor to mortality rates for stillbirths and maternal mortality.
We use the share of neonatal deaths that are considered early (based on 2019 mortality estimates from the Institute for Health Metrics and Evaluation, GBD Results Tool) as a proxy for pre-discharge neonatal mortality.
Finally, we apply a 49% adjustment to represent the share of stillbirths that are considered fresh (i.e. occurring during labor). See sub-Saharan Africa estimate in Figure 5, UNICEF, A Neglected Tragedy: The global burden of stillbirths, 2020 pg. 18. - 47
See GiveWell, Facility-based MNH BOTEC, 2022, “RFMF” sheet.