Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Results from a Dose-Response Analysis Authors: Shivam Gupta Abigail Kazembe Tatenda Mupfudze Angella Mtimuni Reena Sethi Lolade Oseni Tambudzai Rashidi Evelyn Zimba Stella Abwao Barbara Rawlins Fannie Kachale www.mcsprogram.org September 2014 This report was made possible by the generous support of the American people through the United States Agency for International Development (USAID), under the terms of the Leader with Associates Cooperative Agreement GHS-A-00-08-00002-00 and Cooperative Agreement AID-OAA-A-14-00028. The contents are the responsibility of The Maternal and Child Health Integrated Program (MCHIP) and The Maternal and Child Survival Program (MCSP), and do not necessarily reflect the views of USAID or the United States Government. Table of Contents Abbreviations .................................................................................................................... iv Acknowledgments ............................................................................................................. v Executive Summary ......................................................................................................... vi Background......................................................................................................................... 1 Methodology....................................................................................................................... 6 Results...............................................................................................................................13 Routine Newborn and Maternal Care...........................................................................26 Results Summary.............................................................................................................47 Discussion .........................................................................................................................50 Recommendations and Next Steps ...............................................................................52 References ........................................................................................................................54 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi iii Abbreviations HBB Helping Babies Breathe MCHIP Maternal and Child Health Integrated Program MDG Millennium Development Goals MNCH maternal, newborn, and child health MOH ministry of health RHD Reproductive Health Division USAID United States Agency for International Development iv Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Acknowledgments The primary objective of this study is to evaluate the quality and coverage of the Helping Babies Breathe (HBB) newborn resuscitation intervention scale-up at the facility level in Malawi over time. An evaluation like this one requires contributions and the dedication of many people, and the authors would like to recognize those who made this study possible. First and foremost, we thank the Malawi Ministry of Health. We would also like to thank the maternity clients who agreed to allow our study teams to observe their care and the care of their newborns, even under emergency circumstances. Without their participation, this study would not have been possible. We also extend our thanks to all the health facility directors and medical personnel in the health facilities assessed for their participation. We understand that they are extremely busy, and we very much appreciate the time they devoted to this evaluation. In addition, we are grateful to the HBB implementation team for their continued support to this evaluation. The study modified MCHIP quality of care survey tools and the HBB simulation tools developed by the HBB Global Development Alliance (GDA). Service Provision Assessment tools informed the inventory tool and provider interviews. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi v Executive Summary Global evidence suggests that improving the quality of obstetric and newborn care can directly reduce maternal and neonatal deaths. Helping Babies Breathe (HBB) is an educational program that teaches essential neonatal resuscitation techniques to health workers in resource-limited areas. The Malawi Ministry of Health (MOH), Reproductive Health Division (RHD), has been working with partners since March 2011 to scale up the HBB approach nationally. Accurate and reliable data are needed on the quality of the Helping Babies Breathe program and the resulting services offered in health facilities in Malawi. With support from the United States Agency for International Development (USAID), and in collaboration with the Malawi MOH, the Maternal and Child Health Integrated Program (MCHIP) and the Support for Service Delivery Integration Program (SSDI) conducted a performance evaluation of the Helping Babies Breathe program in Malawi. The primary objective of the evaluation was to assess the quality of care and coverage of the HBB newborn resuscitation intervention at the facility level in Malawi. The results of this evaluation will help to inform and guide the ongoing scale-up of this program in Malawi and in other countries. A similar evaluation of the HBB initiative in Bangladesh is currently under way and will complement impact evaluations being conducted by the United States’ National Institute of Child Health and Human Development in India and Kenya. The evaluation consisted of data collection at two points in time: the first round was in September 2012, after the intervention had begun in 13 districts; and the second round was in September 2013, when the intervention had been expanded to all districts in Malawi. The methodological strength of this evaluation is that it employs direct observation of labor and delivery and management of newborns who are not breathing at birth, complemented by data collected through interviews, facility audits, and qualitative key informant interviews. Seven data collection tools were used to gather data during health facility visits: (1) health worker interviews and knowledge assessments on essential newborn care; (2 and 3) skills assessments of health providers using clinical simulations of newborn resuscitation; (4) a structured clinical observation checklist for labor and delivery (L&D); (5) a structured clinical observation checklist for the management of asphyxiated newborns; (6) in-depth interviews with key stakeholders who either made or influenced policy decisions related to newborn health in Malawi; and (7) a health facility supply and equipment inventory checklist. The evaluation used a quasi-experimental design for the initial analysis of round 1, whereby 13 districts were classified as “intervention” districts if they were part of the first phase of HBB scale-up in Malawi. The remaining districts were classified as “comparison” districts. However, by the time of the round 2 data collection, all districts in Malawi had been exposed to the HBB intervention. Thus, the final comparative analysis of the round 1 and round 2 data used a dose-response analysis. For this analysis, districts were classified into three dose-response groups based on their level of exposure to the different elements of the HBB intervention—high, medium, and low dose—based on a score created by combining responses for 16 items related to health provider training, practice with the NeoNatalie anatomic model, supervision, and availability of equipment and supplies for newborn care at facility level. vi Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Results This report presents results from the first and second rounds of data collection.1 The first round of the study included 81 health facilities and the second round included 90. A total of 190 and 202 facility-based health workers who attend labor and deliveries were interviewed in round 1 and 2 respectively. In both rounds, about 60% of all health providers interviewed were nurse/midwife technicians. The mean number of years that respondents had provided delivery services was 5.9 in round 1 and 6.0 in round 2, with nurse/midwife technicians having an average of 4.3 years and 4.5 years of delivery services in rounds 1 and 2 respectively. There were no significant differences between the high-, medium-, and low-dose districts in the mean number of years of service. More health workers interviewed had received training in subjects related to newborn care in the past two years in round 2 (68.4%) than in round 1 (59.7%). The trained individuals largely came from the nurse/midwife technician group (34.8% and 41.8% in rounds 1 and 2 respectively). There were no differences in health provider training by dose group in either round of the study. The majority of supervisors checked the records of their supervisees (81.7% and 88.8% in rounds 1 and 2, respectively) and observed their work (78.6% and 87.3% in rounds 1 and 2, respectively). However, only slightly more than one-third (39.7% and 41.5% in rounds 1 and 2 respectively) of supervisees had received any written comments from their supervisors. Most supervisees were able to discuss any problems encountered with their supervisor (86% and 90% overall in rounds 1 and 2, respectively) and were given verbal feedback (82.2% and 88.6% overall in rounds 1 and 2, respectively). More supervisors observed work in the medium-dose group in round 1 compared to the other groups; however, there were no obvious differences in supervision by dose in round 2. Quality of resuscitation care using clinical simulations with the NeoNatalie model Among the health workers who participated in a clinical simulation of bag and mask ventilation using the NeoNatalie anatomic model, the mean number of steps that were correctly completed (out of a total possible score of 10) was higher in round 2 (mean of 7.1) compared to round 1 (mean of 6.2). In both rounds, the steps completed by the fewest respondents were squeezing the bag harder if the newborn’s chest did not move and testing the function of the bag and mask. The step completed by the highest proportion of respondents in both rounds was extending the newborn’s head. The overall mean score was higher in the high-dose arm than in the medium-dose arm (p=0.041) in round 1. The overall mean score was not significantly different by dose in round 2. However, health workers in the high-dose arm scored higher at testing the function of the bag and mask and squeezing the bag harder. Respondents were asked to complete two role-play case scenarios with the newborn simulator (NeoNatalie). In the first scenario, the mean number of steps performed correctly was lower in round 1 (mean 10.1; SD 3.3) than round 2 (mean 11.0; SD 2.9) out of 16 possible total correct steps, with a range of two to 16 steps performed correctly in both rounds. In both rounds, most respondents dried the baby thoroughly while few made an emergency plan. In the second scenario the mean number of steps performed correctly was higher in round 2 (mean 19.5; SD 5.8; range 4–29) than in round 1 (mean 17.4; SD 6.6; range 2–29). Most providers remembered to thoroughly dry the baby, and as with the first scenario, the fewest respondents called for help in both round 1 and 2. 1 Results from the first round alone were presented earlier in a separate report. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi vii Overall, the proportion of facilities that had the equipment required for newborn care was not different by dose group. One exception was suction bulbs for mucus extraction, which were more common in the high-dose group in round 1 and more common in the medium-dose group in round 2. In both rounds, the proportion of facilities that had the equipment required for delivery services was not different by dose group except for single-use hand-drying towels, which were more common in the medium-dose group. Observed quality of routine labor and delivery care All stages of labor were observed in 175 cases in round 1 and 193 cases in 2. The second, third, and immediate postpartum stages of labor were observed in 1,417 cases in round 1 and 1,842 cases in round 2 to assess whether women and newborns were receiving evidence-based interventions such as screening and prevention of pre-eclampsia/eclampsia, postpartum hemorrhage, and newborn asphyxia and hypothermia. Although most mothers were checked for their client card, fetal presentation, and fetal heart rate and most received a vaginal exam, only about half in each round had their temperature or pulse taken, and only 5.7% in round 1 and 3.6% in round 2 were tested for urine protein. Only 32.6% and 24.4% of mothers were checked for fundal height in rounds 1 and 2, respectively. A higher percentage of providers in the low-dose group asked about at least one danger sign, performed general examination (anemia, edema, etc.), and checked fundal height in round 1. In round 2, a higher percentage of providers in the medium-dose group than in the other dose groups took the woman’s temperature and blood pressure and performed a general examination. More than 85% of mothers in both rounds were greeted respectfully, but less than 20% were encouraged to have someone in attendance at delivery. Less than 30% were asked whether they had any questions for the health provider or provided with drapes. Less than 60% were encouraged to hydrate and eat during labor. The only differences in performance by dose group were that providers in the high-dose arm more often informed the pregnant woman of findings in round 1 and more often explained procedures to the woman in round 2. As part of active management of the third stage of labor, 91% and 98% of women in round 1 and 2, respectively, were given a uterotonic immediately after birth. More nonbeneficial behaviors were reported in round 1 (5.6%) than in round 2 (2.7%). Very few nonindicated practices were reported in either round. The most frequently reported nonbeneficial and nonindicated behaviors were holding the newborn upside-down, applying fundal pressure, and stretching the perineum. Data were complete for a total of 1,303 and 1,800 observations of immediate newborn care in rounds 1 and 2, respectively. Close to 70% of newborns had delayed cord clamping after birth as recommended, over 95% were immediately dried, and almost all newborns were either placed skin-to-skin or wrapped in a dry towel in both rounds. In round 1, 42% of women started breastfeeding their newborns within one hour after birth compared to 78% in round 2. Newborns were more likely to be placed skin-to-skin or wrapped in a dry towel and breastfed within the first hour in the medium-dose group in round 1. Delayed cord clamping and cutting the cord with a clean blade were more prevalent in the high-dose group in round 1. In round 2, immediately drying the baby, delayed cord clamping, and initiation of breastfeeding in the first hour were more common in the low-dose group than in the other groups. viii Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Observed quality of management of newborns not breathing at birth A key component of this evaluation was observing the management of newborns who were not breathing at birth. Out of the 1,747 valid observations in round 1, 88 newborns in the high-dose group, 46 in the medium-dose group, and 59 in the low-dose group were found not to be breathing at birth (a total of 193 newborns). Among the newborns who were found not to be breathing at birth, 84 survived in the high￾dose arm (95.5%), 42 survived in the medium-dose arm (91.3%), and 54 survived in the low dose arm (91.5%). Stimulation was given to 66 newborns in the high-dose arm (75%), 32 newborns in the medium-dose arm (69.6%), and 47 newborns in the low-dose arm (79.7%). After stimulation, 63.6%, 37.5% and 52.2% of the newborns in the high-, medium-, and low-dose groups, respectively, started breathing. Among the babies who were not revived by initial stimulation, 23 of 24 in the high dose group (95.8%), 19 of 20 in the medium￾dose group (95%), and 20 of 22 in the low-dose group (90.9%) received a bag-and-mask intervention. After the bag-and-mask intervention, 21of 23 babies survived in the high-dose group (91.3%), 18 of 19 (94.7%) survived in the medium-dose group, and 18 of 20 (90%) survived in the low-dose group. The differences in proportion between the dose groups were not statistically significant for any outcome. Out of the 2,093 valid observations in round 2, 91 newborns in the high-dose group, 90 in the medium-dose group, and 99 in the low-dose group were found not to be breathing at birth (a total of 291 newborns). Among the babies found not to be breathing at birth, 88 survived in the high-dose group (96.7%), 87 survived in the medium-dose group (95.6%), and 92 survived in the low-dose group (92.9%). Stimulation was given to 68 newborns in the high-dose arm (74.7%), 74 in the medium-dose arm (81.1%), and 82 in the low-dose (82.9%). After stimulation, 48.5%, 63.5%, and 52.4% of the newborns in the high-, medium-, and low-dose groups, respectively, started breathing. Among the babies who were not revived after stimulation, 29 of 35 in the high-dose group (82.9%), 25 of 26 in the medium-dose group (96.2%), and 32 of 39 in the low-dose group (82.1%) received a bag-and-mask intervention. Among babies who received a bag-and-mask intervention, 28 of 29 survived in the high-dose group (91.1%), 22 of 25 survived in the medium-dose group (88%), and 27 of 32 survived in the low-dose group (84.4%). The differences in proportion between the arms were not statistically significant for any outcome. In summary, there was an overall improvement in health worker training, knowledge, equipment availability, and management of labor and delivery, including newborn care, over the two rounds of data collection in Malawi. However, there were no significant differences between the high-, medium-, and low-dose groups in any of the two rounds. Although health worker performance in round 2 was better than in round 1, a majority of the findings in this report indicate the absence of significant differences by dose group in the two rounds. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi ix x Evaluation of the Helping Babies Breathe (HBB) Initiative Scale -Up in Malawi Background A review of progress toward Millennium Development Goal Four (MDG 4) indicates that while under-five mortality has declined worldwide, neonatal mortality has not experienced a similar decline. An estimated 3.6 million neonates die each year globally, 99% of them in developing countries (Lawn et al. 2010; Black et al. 2010). Neonatal deaths represent an increasing proportion of under-five deaths—an estimated 41% globally in 2008 compared to 38% in 2000—due to the stagnation in neonatal mortality rates. Neonatal mortality is largely attributable to three preventable conditions: birth asphyxia, prematurity, and infections. Globally, about one-tenth of the under-five deaths, an estimated 814,000 deaths per year, are caused by birth asphyxia (intrapartum-related deaths) (Bryce et al. 2005). Even in low-resource settings, many of these deaths can be prevented by improving early recognition of newborn asphyxia and access to appropriate and timely resuscitation. A Delphi estimation reported that immediate newborn assessment and stimulation could lead to a 10% drop in intrapartum-related and preterm deaths, and to an additional 30% drop with facility-based resuscitation (Lee et al. 2011). Birth asphyxia occurs as a consequence of interrupted placental blood flow In low- and middle-income countries, birth asphyxia is defined as failure to initiate respiration at birth. Despite efforts to improve outcomes, morbidity and mortality rates associated with birth asphyxia have remained unchanged. This reflects several factors, including a lack of essential basic resuscitation equipment and a failure to initiate resuscitation in a timely manner. Although maternal, newborn, and child health (MNCH) programs are being scaled up in many low-resource settings, limited guidance has been provided to health workers regarding identification and management of newborn asphyxia. Furthermore, health workers may find it difficult to sustain and improve their resuscitation skills because the rarity of asphyxia provides too few cases for adequate training and practice. Helping Babies Breathe (HBB) is an educational program developed by a group of stakeholders that include: • American Academy of Pediatrics • U.S. Agency for International Development • Saving Newborn Lives/Save the Children • Eunice Kennedy Shriver National Institute of Child Health and Human Development • Laerdal Global Health • Johnson & Johnson • Latter Day Saint Charities HBB is used to train health workers in essential neonatal resuscitation knowledge, skills, and techniques to manage asphyxiated newborns in resource-limited areas. An important goal of this initiative is to have at least one person skilled in neonatal resuscitation present at the birth of every baby. The HBB curriculum was designed to be used as part of a coordinated educational approach to early neonatal care and can be effectively combined with other curricula. It can be used locally for training birth attendants in diverse venues and locations. HBB focuses on practices that everyone who cares for newborns can learn and use to assist babies who do not breathe on their own at birth. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 1 To accomplish this goal, the HBB approach to newborn resuscitation was developed as a practical and easy-to-use training solution, which includes: • An evidence-based educational program, based on the International Liaison Committee on Resuscitation (ILCOR) consensus on science conclusions that have undergone a World Health Organization scientific technical review; • Culturally sensitive, pictorial-based learning materials, including a learner workbook, an action plan wall poster, and a facilitator flip chart; • Realistic newborn NeoNatalie anatomic simulator (developed by Laerdal Medical, a Norwegian medical device company), which simulates breathing during resuscitation with bag and mask ventilation, imitates an umbilical pulse, and comes with a newborn size bag-mask ventilator and penguin suction bulb that can be cleaned and disinfected by boiling (all equipment has been tested for durability in a variety of climates and teaching conditions and has been made available at cost to MDG countries); and • An ongoing mentorship program to provide expert assistance, implementation guidance, knowledge exchange, integration and evaluation support, and continuous quality improvement for sustained practice outcomes and decreased infant mortality. Elements of the HBB tool kit Action plan 2 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Facilitator flip chart Exercises with neonatal simulator (low-cost mannequin) Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 3 Learner workbook The HBB toolkit also includes the following: • Performance evaluation (OSCE) • Written/verbal evaluation • Video Situated in the southeastern part of Africa, Malawi has a population of 14 million people and ranks as one of the poorest countries in the world. The country is administratively divided into three regions (Northern, Central, and Southern), which are further divided into 28 districts. Each district is divided into 250 traditional authorities and 110 administrative wards. Malawi’s health profile is characterized by a high prevalence of communicable diseases. The major burden of disease is due to HIV/AIDS, malaria, tuberculosis, sexually transmitted infections, diarrhea, and acute respiratory infections. The neonatal mortality rate in Malawi is 31 per 1,000 live births Macro (NSOMaO 2011). Major causes of neonatal mortality include neonatal sepsis (29%), prematurity (29%), and asphyxia (23%) (Lawn et al. 2010). Ensuring health workers who attend births have resuscitation skills is a crucial aspect of efforts to prevent asphyxia-related deaths in Malawi. Study objectives The primary objective of this study was to evaluate the quality and coverage of HBB interventions at the facility level in Malawi. This evaluation included a facility survey and direct observation of labor and delivery. Complementary qualitative research was conducted to elicit additional information and context for the evaluation. The report provides guidance and recommendations for further strengthening HBB implementation and scale-up and improving newborn care in Malawi. It includes key information for policymakers regarding the value of integrating HBB into the strategy for basic emergency obstetric and 4 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi neonatal care. Furthermore, it sheds light on the availability and quality of neonatal health services in Malawi, highlighting strengths and areas for improvement. The evaluation addressed the following questions: • Are newborn resuscitation service delivery guidelines and supplies available and compliant with the national rollout plan? • Are the HBB-trained health workers able to apply the newborn resuscitation and essential newborn care skills after 18 and 30 months of training? • What recommendations can be made for improved availability and quality of neonatal health services in Malawi? Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 5 Methodology Study design and analysis plan HBB was rolled out in Malawi starting in early 2011. According to the original scale-up plan developed by stakeholders, the rollout was to be financed and led by the Ministry of Health, with support from USAID/MCHIP, Johnson & Johnson/Save the Children International, and other development partners. According to the initial plan, implementation of HBB was to begin in February 2011 in a phased manner in 13 districts (MCHIP), with implementation in the remaining districts during 2012 and 2013. The first round of data collection and analysis was based on the original scale-up plan, whereby the first 13 districts were classified as intervention districts and the remaining 14 as comparison districts.2 The first round of data was collected in August and September 2012 and results were presented in July 2013. Health provider knowledge improved more in intervention districts than in comparison districts; however, this improvement was not associated with a corresponding improvement in provider performance. Need for dose-response analyses HBB had been scaled up all over Malawi by the time of the second round of data collection in August and September 2013, and there were no comparison districts available to continue with the original analysis plan. Moreover, due to several financial and logistical difficulties (documented in the recent process documentation conducted by Saving Newborn Lives), the Ministry of Health could not lead the scale-up as proposed in the original rollout plan. Therefore, the districts were grouped according to the “dose” of HBB for analysis after the second round of data collection. The purpose of the dose-response analysis was to measure the extent to which districts were exposed to different domains of HBB intervention and assess whether the difference in dose was associated with difference in availability of equipment, supplies, health worker training, knowledge, and performance. Use of dose-response analyses allowed classification of districts according to the actual strength of the HBB intervention, rather than as intervention and comparison groups, which had limited utility as a measure of exposure to HBB intervention. The four domains that constituted the HBB dose in each district were: • Training and capacity-building • Practice with newborn simulator, “NeoNatalie” • Supervision • Availability of equipment, supplies, and guidelines Sixteen items (variables) were used to constitute these four domains. Each of the 16 items was scored as a percentage, calculated for each district, for a total possible score of 1600. The scores were ranked from highest to lowest, and the 27 districts were divided into three groups—high, medium, and low dose—based on the score created from the 16 items (Table 1). Overall scores and dose-response districts are presented in Tables 2 and 3 for round 1 and 2, respectively. 2 Likoma district was excluded from this evaluation. 6 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 1: Items in Dose Calculation for Each District Training and capacity-building 1 Received any training on how to resuscitate a newborn with bag and mask in the past 2 years 2 Received any training on resuscitation of newborns not breathing at birth in the past 2 years 3 Received this newborn resuscitation training as part of the HBB initiative Practice with model 4 Facility has an anatomic model to use to practice resuscitation of newborns that do not breathe (for example, a "NeoNatalie") 5 Opportunity to practice resuscitating a newborn using a newborn anatomic model/doll after you were trained 6 Opportunity to practice resuscitating a newborn using a newborn anatomic model/doll in the past 3 months Supervision 7 Supervised in the past 3 months 8 Supervisor observed performing newborn resuscitation with a newborn anatomic model/doll (NeoNatalie) Availability of equipment, supplies, and guidelines 9 Facility performed newborn resuscitation 10 Facility performed newborn resuscitation in the past 3 months with bag and mask 11 Bag and mask (infant size) for resuscitation—Size 0 12 Bag and mask (infant size) for resuscitation—Size 1 13 Suction bulb for mucus extraction 14 Towel or blanket to wrap baby 15 Helping Babies Breathe guidelines for newborns not breathing at birth 16 Facility has the Helping Babies Breathe action plan posted on the wall in L&D 7 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 2: Aggregated Score and Classification of Districts into Dose Groups in Round 1 8 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 3: Aggregated Score and Classification of Districts into Dose Groups in Round 2 Data collection tools The following tools were developed for this evaluation: • Health worker interview and knowledge and skill assessment. Three tools, including both quantitative and qualitative components, were used for health care worker interviews. A short structured interview assessed each provider’s knowledge and practice of HBB. The goal was to collect information on the constraints in and facilitators of delivering quality care, and recommendations for ways to improve quality of care. The quantitative instruments included two observational clinical case studies/ simulations, which were the providers completed and which assessed the providers’ clinical decision-making pertaining to screening, management, and treatment of newborn birth asphyxia. • Direct observation of deliveries. Two quantitative observation checklists were used for observation of deliveries—specifically, the provider’s performance of labor, delivery, newborn, and immediate postpartum care and newborn resuscitation. Provider practice during labor, delivery, and the immediate postpartum period were observed and documented by study staff in the selected facilities. The labor and delivery and newborn resuscitation observation checklists were adapted from the MCHIP maternal and newborn health quality of care facility survey, which has been conducted in multiple countries in sub-Saharan Africa. The newborn resuscitation checklist documented adherence to American Academy of Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 9 Pediatrics (AAP)-developed HBB protocols for screening, management, and treatment of birth asphyxia in newborns; age, gravidity, and parity of the mother; qualification of the provider; and level of health facility. • Key informant interview. A key informant interview guide was used with stakeholders at the national, district, and facility levels to determine the status of HBB rollout and the quality of maternal and newborn care. The respondents were asked questions related to strengths, weaknesses, and recommendation for improvement. Respondents were selected so that there would be representation from the following organizations: the national Ministry of Health, the national professional society of obstetrics/gynecology, the national association of midwives, and hospital maternity ward in-charges and district health officers/district nursing officers from zonal MOHs. The study co-investigators selected the key informants based on their influence, knowledge, and work related to the HBB initiative in Malawi. • Record review. This tool captured the number of antenatal care consults, deliveries, births, deaths, and obstetric complications at each facility for the last year from facility records, including maternity registers. • Facility inventory. The facility inventory tool assessed infrastructure conditions and verified the availability of and storage conditions for medications, supplies, and equipment. Through an interview with the head of each health facility or their designee, a listing was generated of all health workers who attend deliveries and/or provide antenatal care. Sample A total of 90 (data available for 81) facilities were randomly selected to be representative of the 27 high-, medium-, and low-dose districts in round 1. In round 2, only two out of 90 facilities were replaced due to low patient volume or non-availability of health workers providing labor and delivery services. Two or three health workers working in the maternity and labor ward in each sampled facility were included in the study. A total of 190 and 202 health workers were sampled from the selected facilities in round 1 and 2, respectively. The sample sizes of the high-, medium-, and low-dose arms in round 1 were 55, 70, and 65 health workers, respectively (Table 1). The sample sizes for the high-, medium-, and low-dose arms in round 2 were 66, 68, and 68 health workers, respectively. Direct observation of patient care was conducted only in facilities in which at least five deliveries were conducted per day on average and that were capable of managing obstetric emergencies. Therefore, data on labor and delivery observations were collected in only 44 facilities in each round. Table 4: Health Facility Sample, Health Workers Interviewed, and Labor and Delivery Observations Round 1 Round 2 High Medium Low Total High Medium Low Total Health facility 23 28 30 81 30 30 30 90 Health providers 55 70 65 190 66 68 68 202 Women in labor 527 567 476 1,570 635 742 732 2,109 Newborns not breathing at birth 98 70 57 225 96 93 102 291 10 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 5: Components of Labor and Delivery (L&D) Observed Round 1 Round 2 High Medium Low Total High Medium Low Total Components of labor and delivery observed Number of cases Number of cases Number of cases Number of cases Number of cases Number of cases Number of cases Number of cases Initial client assessment 66 62 65 193 59 67 67 193 First stage of labor 66 62 65 193 62 75 71 208 Second and third stage of labor 527 567 476 1,570 635 742 732 2,109 Immediate newborn care and postpartum care 527 567 476 1,570 635 742 732 2,109 Total number of L&D observations 527 567 476 1,570 635 742 732 2,109 Data collection procedure A week of training for 34 data collectors was conducted in August 2012 and August 2013 in Lilongwe, Malawi. The data collectors were previously trained health care workers who had experience in basic emergency obstetric and neonatal care. Most of these data collectors had also been trained previously in the Helping Babies Breathe curriculum. The data collector training included briefings on the background and rationale of the study and a description of and technical instructions for completion of data collection tools. The trainees also had an opportunity to go to the field to practice using actual tools for data collection at health facilities and to observe actual maternity clients. Data collection was monitored on an ongoing basis by a trained team of supervisors and was overseen by a study coordinator. The supervisors checked the data collection plans, observed data collection, and reviewed completed forms during their monitoring visit. The fieldwork occurred between August 20 and September 22, 2012, for round 1 and between September 1 and September 30, 2013, for round 2. Direct observation of patient care was conducted at 44 facilities that had an average of five deliveries per day and were capable of managing obstetric emergencies. At each of these facilities, one data collector spent 10–12 days. In addition to conducting the facility inventory, record review, health worker interview, and knowledge assessment, the data collector observed the care provided during all types of deliveries: normal vaginal delivery, assisted (vacuum or forceps) or cesarean section. At each of the remaining facilities (46 in round 1 and round 2), a team of two data collectors spent one day conducting only the facility inventory, the record review, and the health worker interview/knowledge assessment. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 11 Data analysis Data were collected on paper-based forms and then entered by trained data entry clerks into a Microsoft Access database that was designed for this evaluation. Data were analyzed using Stata analytical software (Stata SE), version 12. Descriptive statistics, including means and percentage distributions, were calculated. All three components of the study’s primary objectives were assessed: quality, coverage, and outcomes of the Helping Babies Breathe newborn resuscitation intervention at the facility level. Tabulations were run on the collected data. For each topic or subtopic area for which data were collected (i.e., initial client assessment during delivery and labor, and partograph use), tabulations were only conducted for those individuals who had answers for all observational items in that particular section/sub-section of the clinical observation checklist. This allowed us to have a consistent denominator for each section/subsection. The tabulations are reported stratified by group (intervention/non-intervention) or dose (high/medium/low) and also combined (total). Ethical considerations The study protocol was approved by the College of Medicine Research Ethics Committee in Malawi and the Institutional Review Board of the Johns Hopkins Bloomberg School of Public Health (JHSPH). Informed consent was obtained from all participating health providers and patients, and permission to visit the health facilities was obtained from facility directors. If a woman was incapacitated, consent was to be obtained from next of kin or guardian. However, this situation did not occur in the course of the study. Limitations There are several important limitations that should be noted. Simulation of newborn resuscitation on models to assess health worker skills has limitations; for example, extension of the model’s neck may need to be exaggerated to achieve simulated breathing, and a standardized “mother” or “guardian” should be used to test the provider’s interpersonal skills during resuscitation. Some providers stopped after the stimulation and suction portion and did not perform bag and mask resuscitation. In addition, in interpreting the results from the simulation, the fact that the model was new for some providers, especially those who had not yet received training in HBB, should be considered. However, providers were given an opportunity to examine the model and familiarize themselves with it before the simulation began. 12 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Results Health personnel and years of service As part of the evaluation, health providers were interviewed about their educational background, years providing services, training received, working conditions, and knowledge of maternal and newborn care. A total of 188 and 202 health workers were interviewed in rounds 1 and 2, respectively. More than 59% and 62% of all health providers interviewed classified themselves as nurse/midwife technicians in rounds 1 and 2, respectively (Table 6). Table 6: Health Providers Interviewed, by Type Round 1 Round 2 High (n=61) Medium (n=67) Low (n=60) Total (n=188) High (n=66) Medium (n=68) Low (n=68) Total (n=202) Medical assistant 8 (13.1%) 8 (11.9%) 9 (15.0%) 25 (13.3%) 13 (19.7%) 10 (14.7%) 13 (19.1%) 36 (17.85) Clinical officer 1 (1.6%) 1 (1.5%) - 2 (1.1%) - - 3 (4.4%) 3 (1.5%) Registered midwife 7 (11.5%) 11 (16.4%) 7 (11.7%) 25 (13.3%) 5 (7.6%) 4 (5.9%) 5 (7.4%) 14 (6.9%) Enrolled nurse/midwife 4 (6.6%) 8 (11.9%) 9 (15.0%) 21 (11.2%) 3 (4.6%) 9 (13.2%) 6 (8.9%) 18 (8.9%) Nurse/midwife technician 40 (65.6%) 38 (56.7%) 33 (55.0%) 111 (59.0%) 43 (65.2%) 43 (63.2%) 40 (58.8%) 126 (62.45) Other 1 (1.6%) 1 (1.5%) 2 (3.3%) 4 (2.1%) 2 (3.0%) 2 (2.9%) 1 (1.5%) 5 (2.5%) The profile of maternity providers included in this study was comparable between dose groups and reflects the anticipated distribution of provider cadres (Tables 7a and 7b). Overall, the mean number of years that the health worker had been providing any type of health service was 5.9 years in round 1 and 6.6 years in round 2 (range 0 to 35 years). Providers in the largest group (nurse/midwife technician) had on average been providing services for 4.3 years in round 1 and 4.5 years in round 2. There were no significant differences in the mean years of service by dose group in either round 1 or round 2. Health worker training and knowledge More health workers interviewed received training in subjects related to newborn care in the past two years in round 2 (68.4%) compared to round 1 (59.7%) (Table 8). The trained individuals largely came from the nurse/midwife technician group (34.8 and 41.8% in round 1 and 2 respectively). There were no differences in health provider training by dose group in both round 1 and 2. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 13 Table 7a: Years of Experience in Service Delivery, by Type of Health Worker, Round 1 Cadre Total years of providing delivery services High Medium Low Total p-value n Mean Range n Mean Range n Mean Range N Mean Range Medical assistant 8 7 0–21 8 5.1 2–19 8 2.4 1–5 24 4.8 0–21 0.459 Clinical officer 1 31 31–31 1 9 9–9 - - - 2 20 9–31 0.157 Registered midwife 7 0.9 0–2 11 2.4 0–17 7 1.6 0–5 25 1.7 0–17 0.258 Enrolled nurse/midwife 4 18 0–30 8 20.8 5–42 9 25.4 1–40 21 22.2 0–42 0.237 Nurse/midwife technician 40 3.2 0–24 37 4.3 0–21 33 5.4 0–21 110 4.2 0–24 0.067 Other 1 53 0–53 1 12 12–12 2 16 12–20 4 24.3 12–53 0.287 Total 61 5.6 0–53 66 6.3 0–42 59 7.9 0–40 186 6.6 0–53 0.091 Table 7b: Years of Experience in Service Delivery, by Type of Health Worker, Round 2 Cadre Total years of providing delivery services High Medium Low Total p-value n Mean Range n Mean Range n Mean Range N Mean Range Medical assistant 12 4.3 0–20 9 7.3 2–20 13 4 1–7 34 5 0–20 0.430 Clinical officer - - - - - - 3 7.3 0–20 3 7.3 0–20 - Registered midwife 5 1.6 0–4 4 0.5 0–2 5 1.2 0–4 14 1.1 0–4 0.643 Enrolled nurse/midwife 3 24.3 17–31 8 18.8 8–26 6 18.8 0–28 17 19.8 0–31 0.383 Nurse/midwife technician 43 3.7 0–25 43 3.8 0–18 40 6.1 0–30 126 4.5 0–30 0.731 Other 2 14.5 0–29 2 2.5 0–5 1 35 35–35 5 13.8 0–35 0.277 Total 65 5.0 0–31 66 5.9 0–26 68 6.9 0–35 199 5.9 0–35 0.538 14 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 8: Proportion of Health Providers Who Received Training in Newborn Care in Past Two Years Cadre Round 1 Round 2 High (n=60) Medium (n=63) Low (n=58) Total (N=118) p-value High (n=63) Medium (n=66) Low (n=67) Total (n=196) p-value n % n % n % n % n % n % n % N % Medical assistant 2 3.3 4 6.4 4 6.9 10 5.5 0.561 9 14.3 5 7.6 6 9.0 20 10.2 0.333 Clinical officer 0 0.0 - - - - 0 0.0 - - - - - 2 3.0 2 1.0 - Registered midwife 4 6.7 7 11.1 6 10.3 17 9.4 0.499 5 7.9 4 6.1 5 7.5 14 7.1 - Enrolled nurse/midwife 4 6.7 4 6.4 6 10.3 14 7.7 0.189 2 3.2 7 10.6 3 4.5 12 6.1 0.535 Nurse/midwife technician 26 43.3 17 27.0 20 34.5 63 34.8 0.205 30 47.6 24 36.4 28 41.8 82 41.8 0.227 Other 1 1.7 1 1.6 2 3.5 4 2.2 - 1 1.6 2 3.0 1 1.5 4 2.0 0.392 Total 37 61.7 33 52.4 38 65.5 108 59.7 0.314 47 74.6 42 63.6 45 67.2 134 68.4 0.394 15 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Knowledge and skills during simulation of bag-and-mask ventilation Table 9 presents assessment data for health workers who participated in the bag and mask ventilation assessment using the NeoNatalie anatomic model. The mean number of steps that were completed correctly was higher in round 2 (mean 7.1; SD 2.0) than round 1 (mean 6.2; SD 2.4), out of a total possible score of 10. In both rounds, the steps completed by the lowest proportion of respondents were squeezing the bag harder if the newborn’s chest did not move and testing the function of the bag and mask. The step completed by the highest proportion of respondents in both rounds was extending the newborn’s head. The overall mean score was higher in the high-dose arm than the medium dose arm (p=0.041) in round 1. The overall mean score was not different by dose in round 2, although health workers in the high-dose arm scored higher on testing the function of the bag and mask and squeezing the bag harder. The respondents were asked to complete two case scenarios using the NeoNatalie anatomic model. In the first case scenario, the respondent was asked to perform the steps necessary to deliver a term baby without complications in pregnancy (Table 10). The mean number of steps performed correctly was lower in round 1 (mean 10.1; SD 3.3) than round 2 (mean 11.0; SD 2.9), out of 16 steps, with a range of 2 to 16 steps performed correctly in both rounds. In both rounds most respondents dried the baby thoroughly, and the fewest respondents made an emergency plan. Providers were asked to complete a second case scenario with the NeoNatalie anatomic model in which they were to pretend that they were assisting at the birth of a baby at 34 weeks’ gestation. The mean number of steps performed correctly out of 29 steps was higher in round 2 (mean 19.5; SD 5.8; range 4–29) than round 1 (mean 17.4; SD 6.6; range 2–29). In both round 1 and round 2, most providers remembered to thoroughly dry the baby and, as with the first scenario, the fewest respondents called for help. The results of the second case scenario are presented in Table 11. In round 1, health workers in the high-dose arm had higher overall scores than those in the medium- and low-dose arms. In round 2, health workers in the medium-dose group scored higher on recognizing that baby was not crying, calling for help, continuing ventilation, but the overall mean score was not statistically significantly different from the high- or low-dose groups. 16 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 9: Clinical Simulation of Bag-and-Mask Ventilation Using NeoNatalie Model Round 1 Round 2 High (n=53) % Medium (n=57) % Low (n=61) % Total (N=171) % p-value High (n=64) % Medium (n=63) % Low (n=66) % Total (N=194) % p-value Checks equipment and selects the correct mask Tests function of bag and mask 55.6 34.5 44.3 44.5 0.081 59.4 43.8 37.9 46.9 0.041 Makes sure mask fits the baby’s face 68.5 55.2 70.5 64.7 0.17 85.9 79.7 87.9 84.5 0.404 Applies the mask to make a firm seal Extends the head 88.9 84.5 93.4 89 0.295 93.8 93.8 92.4 93.3 0.941 Places mask on the chin, then mouth and nose 83.3 82.8 86.9 84.4 0.798 85.9 85.9 86.4 86.1 0.997 Ensures a firm seal to permit chest movement when the bag is squeezed 66.7 63.8 68.9 66.5 0.842 78.1 85.9 77.3 80.4 0.393 Ventilates at 40 breaths per minute 61.1 39.7 55.7 52 0.058 62.5 54.7 65.2 60.8 0.448 Looks for chest movement 75.9 67.2 80.3 74.6 0.251 84.4 87.5 81.8 84.5 0.669 Improves ventilation if the chest does not move Reapplies mask and repositions head 63 63.8 52.5 59.5 0.374 81.2 71.9 74.2 75.8 0.436 Clears secretions and opens the mouth 53.7 53.4 45.9 50.9 0.628 59.4 56.2 62.1 59.3 0.793 Squeezes the bag harder 40.7 20.7 26.2 28.9 0.055 42.2 39.1 22.7 34.5 0.043 Mean number steps performed correctly 6.6 5.7 6.2 6.2 0.041a & 0.404b 7.3 7.0 6.8 7.1 0.379a & 0.235b SD 2.5 5.7 2.4 2.4 2.0 1.9 2.1 2.0 Range 1–10 1–10 1–10 1–10 0–10 1–10 1–10 0–10 a High dose compared to medium dose b High dose compared to low dose 17 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 10: Results of Clinical Simulation to Assess Health Worker Knowledge: Initial Stimulation Round 1 Round 2 High (n=56) % Medium (n=60) % Low (n=60) % Total (n=176) % p-value High (n=63) % Medium (n=66) % Low (n=64) % Total (n=193) % p-value Prepares for birth Identify a helper 64.3 16.7 36.7 38.6 0 36.5 43.9 29.7 36.8 0.242 Makes an emergency plan 53.6 21.7 28.3 34.1 0.001 30.2 45.5 26.6 34.2 0.054 Prepares area for delivery 83.9 71.7 73.3 76.1 0.248 76.2 87.9 79.7 81.4 0.215 Cleans hands and maintains clean technique throughout 75 31.7 28.3 44.3 0 44.4 51.5 43.8 46.6 0.617 Prepares an area for ventilation 71.4 60 68.3 66.5 0.399 68.3 71.2 64.1 67.9 0.681 Checks equipment 53.6 36.7 50 46.6 0.153 61.9 47.0 45.3 51.3 0.119 Keeps baby warm Dries thoroughly 96.4 90 96.7 94.3 0.205 95.2 97.0 95.3 95.9 0.855 Removes wet cloth 83.9 70 80 77.8 0.173 76.2 90.9 81.3 82.9 0.078 Covers baby with dry cloth 78.6 71.7 83.3 77.8 0.302 77.8 84.9 78.1 80.3 0.520 Evaluates crying Recognizes baby is not crying 83.9 70 63.3 72.2 0.042 90.5 92.4 90.6 91.2 0.909 Clears airway and stimulates breathing Positions head and clears airway 75 68.3 61.7 68.2 0.305 76.2 83.3 76.6 78.8 0.533 Simulates breathing by rubbing the back 58.9 53.3 50 54 0.624 55.6 68.2 57.8 60.6 0.291 Evaluates breathing Recognizes baby is breathing well 85.7 78.3 68.3 77.3 0.08 90.5 89.4 95.3 91.7 0.431 Clamps or ties and cuts cord 78.6 73.3 73.3 75 0.756 79.4 69.7 78.1 75.7 0.376 Positions skin-to-skin on mother's chest 80.4 65 70 71.6 0.176 68.2 72.7 75.0 72.0 0.690 Communicates with mother 57.1 33.3 28.3 39.2 0.003 50.8 51.5 51.6 51.3 0.995 Mean number done correctly 11.8 9.1 9.6 10.1 <0.001a & <0.001b 10.8 11.5 10.7 11.0 0.181a & 0.862b 18 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Round 1 Round 2 High (n=56) % Medium (n=60) % Low (n=60) % Total (n=176) % p-value High (n=63) % Medium (n=66) % Low (n=64) % Total (n=193) % p-value SD 2.9 3.2 3.3 3.3 2.8 3.1 2.9 2.9 Range of correct responses 6–16 2–16 3–16 2–16 2–16 3–16 2–16 2–16 a High dose compared to medium dose b High dose compared to low dose 19 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 11: Results from Clinical Simulation to Assess Health Worker Knowledge: Case Scenario 2 Round 1 Round 2 High (n=49) % Medium (n=53) % Low (n=55) % Total (N=157) % p-value High (n=60) % Medium (n=64) % Low (n=63) % Total (N=187) % p-value Prepares for birth Identifies a helper 46.9 11.3 25.5 27.4 <0.001 45 37.5 23.8 35.3 0.044 Prepares area for delivery 71.4 73.6 72.7 72.6 0.97 73.3 79.7 73 75.4 0.617 Cleans hands 63.3 fxz 27.3 37.6 <0.001 43.3 43.8 36.5 41.2 0.651 Prepares an area for ventilation 71.4 64.2 60 65 0.47 73.3 68.8 61.9 67.9 0.392 Checks equipment 59.2 30.2 36.4 41.4 0.008 55 50 46 50.3 0.609 Keeps baby warm Dries thoroughly 93.9 92.5 92.7 93 0.957 93.3 95.3 95.2 94.7 0.859 Removes wet cloth 77.6 71.7 69.1 72.6 0.617 75 81.2 74.6 77 0.609 Covers baby with dry cloth 79.6 67.9 80 75.8 0.259 73.3 85.9 76.2 78.6 0.196 Evaluates crying Recognizes baby is not crying 83.7 77.4 61.8 73.9 0.032 86.7 93.8 98.4 93 0.036 Clears airway and stimulates breathing Positions head and clears airway 49 50.9 47.3 49 0.93 61.7 71.9 69.8 67.9 0.44 Clears airway 69.4 66 67.3 67.5 0.936 78.3 90.6 85.7 85 0.156 Simulates breathing by rubbing the back 67.3 54.7 49.1 56.7 0.162 48.3 60.9 54 54.5 0.368 Evaluates breathing Recognizes baby is not breathing 85.7 73.6 58.2 72 0.007 76.7 85.9 84.1 82.4 0.361 Ventilates with bag and mask Cuts cord 71.4 67.9 69.1 69.4 0.927 66.7 73.4 71.4 70.6 0.699 Moves to area for ventilation 75.5 64.2 74.5 71.3 0.362 75 75 79.4 76.5 0.802 Starts ventilation within the Golden Minute 57.1 47.2 49.1 51 0.568 63.3 67.2 66.7 65.8 0.888 Ventilates at 40 breaths per min 61.2 49.1 49.1 52.9 0.368 68.3 57.8 68.3 64.7 0.363 Looks for chest movement 75.5 64.2 69.1 69.4 0.46 63.3 79.7 73 72.2 0.125 20 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Round 1 Round 2 High (n=49) % Medium (n=53) % Low (n=55) % Total (N=157) % p-value High (n=60) % Medium (n=64) % Low (n=63) % Total (N=187) % p-value Evaluate breathing Recognizes baby is not breathing 77.6 73.6 61.8 70.7 0.181 75 85.9 69.8 77 0.089 Calls for help 34.7 17 25.5 25.5 0.122 25 46.9 27 33.2 0.016 Continues ventilation 87.8 75.5 67.3 76.4 0.048 71.7 87.5 77.8 79.1 0.09 Improves ventilation Head-reposition neck 61.2 50.9 50.9 54.1 0.487 61.7 70.3 71.4 67.9 0.449 Reapplies mask 67.3 56.6 65.5 63.1 0.479 65 79.7 73 72.7 0.185 Mouth-clears secretion, opens mouth slightly 42.9 39.6 40 40.8 0.937 51.7 56.2 57.1 55.1 0.808 Bag-squeezes bag harder 38.8 22.6 27.3 29.3 0.186 45 37.5 28.6 36.9 0.167 Evaluates breathing and heart rate Recognizes baby is breathing 81.6 69.8 67.3 72.6 0.223 78.3 89.1 71.4 79.7 0.045 Stops ventilation 79.6 84.9 74.5 79.6 0.41 81.7 93.8 79.4 85 0.051 Monitors baby 59.2 56.6 60 58.6 0.933 45 64.1 44.4 51.3 0.043 Communicates with mother 57.1 45.3 54.5 52.2 0.445 65 51.6 55.6 57.2 0.302 Mean number done correctly 19.5 16.4 16.5 17.4 0.020a & 0.023b 18.9 20.6 18.9 19.5 0.097a & 0.935b SD 6.3 5.4 7.6 6.6 6.4 5.4 5.9 5.9 Range of correct responses 4–29 5–27 2–28 2–29 4–28 4–29 4–28 4–29 a High dose compared to medium dose b High dose compared to low dose 21 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Availability of Supplies, Equipment, and Guidelines An inventory was taken of available essential equipment, mediations, and supplies for delivery and newborn care, and results are shown for 76 study facilities that had complete data in round 1 and 87 facilities in round 2 (Tables 12 and 13). In both rounds, the proportion of facilities that had the equipment required for delivery services was not different by dose group, except for single-use hand-drying towels, which were more commonly available in the medium-dose groups. The proportion of facilities that had the equipment required for newborn care was not different by dose group, except for suction bulbs for mucus extraction, which were more commonly found in the high-dose group in round 1 and in the medium-dose group in round 2. Overall, guidelines were more commonly available in round 2 compared to round 1. There were no significant differences in the availability of guidelines by dose in round 1. In round 2, 84% of the high-dose facilities had a copy of the HBB guidelines compared to 57% of the medium-dose facilities and 37% of the low-dose facilities (p=0.001). Health providers were asked what actions their supervisors had taken the last time they were supervised. Table 15 shows that the majority of supervisors checked the records of their supervisees (81.7% and 88.8%, respectively, in rounds 1 and 2and observed their work (78.6% and 87.3%, respectively, in rounds 1 and 2). However, only slightly more than one-third of supervisees (39.7% and 41.5%, respectively, in rounds 1 and 2) had received any written comments from their supervisors. Most supervisees were able to discuss any problems encountered with their supervisor (86% and 90%, respectively, in rounds 1 and 2) and were given verbal feedback (82.2% and 88.6%, respectively, in rounds 1 and 2). More supervisors observed work in the medium-dose group in round 1 compared to the other groups; there were no obvious differences in supervision by dose in round 2. 22 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 12: Availability of Equipment for Delivery Services Equipment Round 1 Round 2 High (n=23) Medium (n=26) Low (n=27) Total (N=76) p-value High (n=29) Medium (n=29) Low (n=29) Total (N=87) p￾Observed value % Observed % Observed % Observed % Observed % Observed % Observed % Observed % Spotlight for pelvic exam 34.8 42.3 44.4 41.8 0.772 69.0 41.4 44.8 51.7 0.073 Table or bed for delivery 100 100 100 100 - 100 100 100 100 - Clean or sterile gloves 95.7 100 96.3 97.5 0.580 100 97.0 97.0 97.7 0.599 Sharps container 100 96.2 100 98.7 0.377 96.6 93.1 86.2 92.0 0.337 At least 5+ 2ml or 3ml syringes 91.3 96.2 92.6 92.4 0.773 96.6 93.1 86.2 92 0.337 Already mixed decontaminating solution 82.6 80.8 81.5 79.8 0.986 83.0 90.0 82.8 85.1 0.696 Hand disinfectant 21.7 19.2 14.8 20.2 0.813 20.7 27.6 7.0 18.4 0.117 Waste receptacle with lid & plastic liner 73.9 65.4 51.9 63.3 0.262 51.7 58.6 65.5 58.6 0.566 Soap for hand-washing 73.9 69.2 81.5 73.4 0.582 83.0 90.0 90.0 87.4 0.660 Single-use hand-drying towel 13.0 15.4 3.7 11.4 0.343 <0.001 3.5 20.7 8.1 0.008 Water for hand-washing 100 92.3 100 97.5 0.139 100 97.0 100 98.9 0.364 23 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 13: Availability of Equipment for Newborn Care Equipment Round 1 Round 2 High (n=23) Medium (n=28) Low (n=30) Total (N=81) p-value High (n=29) Medium (n=28) Low (n=28) Total (N=85) p￾Observed value % Observed % Observed % Observed % Observed % Observed % Observed % Observed % Bag and mask (infant size) for resuscitation 78.3 67.9 58.6 67.5 0.323 89.7 89.3 71.4 83.5 0.108 Tube and mask 52.4 71.4 66.7 64.6 0.368 41.4 53.6 53.6 49.4 0.567 Incubator 26.1 21.4 6.7 17.3 0.139 24.1 14.3 10.7 16.5 0.366 Other source of heat for premature infant 38.1 25 30 30.4 0.614 20.7 32.1 21.4 24.7 0.536 Infant scale 91.3 92.9 100 95.1 0.281 100 100 100 100 - Suction bulb for mucus extraction 81.8 67.9 40 61.3 0.006 89.7 96.4 71.4 85.9 0.021 Suction apparatus for use with catheter 81 82.1 83.3 82.3 0.976 86.2 78.6 85.7 83.5 0.688 Resuscitation table for baby 87 67.9 69 73.8 0.233 69 75 60.7 68.2 0.515 Disposable cord ties or clamps 78.3 85.7 93.3 86.4 0.281 100 100 100 100 - Towel or blanket to wrap baby 26.1 37 10 23.8 0.054 3.4 28.6 21.4 17.6 0.037 24 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 14: Availability of Guidelines Round 1 Round 2 High (n=21) Medium (n=25) Low (n=28) Total (N=74) p-value High (n=31) Medium (n=30) Low (n=28) Total (N=89) p￾Observed value % Observed % Observed % Observed % Observed % Observed % Observed % Observed % Guidelines for normal delivery 47.6 28.0 28.6 33.8 0.285 38.7 60.7 40.0 46.1 0.171 Guidelines for emergency obstetric care 81.0 84.0 96.4 82.9 0.201 93.6 92.9 96.7 96.4 0.795 Blank partographs 95.2 88.0 89.3 90.5 0.677 93.6 96.4 90.0 93.3 0.619 Helping Babies Breathe guidelines 52.4 40.0 17.9 35.1 0.036 83.9 57.1 36.7 59.6 0.001 Table 15: Supervisor's Action during Last Supervision Visit, as Reported by Supervisee Round 1 Round 2 High (n=43) Medium (n=48) Low (n=40) Total (n=133) p￾value High (n=48) Medium (n=45) Low (n=47) Total (n=140) p-value n (%) n (%) n (%) N (%) n (%) n (%) n (%) N (%) Checked records 39 (90.7) 36 (75) 32 (80) 107 (81.7) 0.146 43 (90.0) 39 (86.7) 42 (89.3) 124 (88.6) 0.887 Observed work 37 (86) 40 (83.3) 26 (65) 103 (78.6) 0.04 40 (83.3) 42 (93.3) 40 (85.1) 122 (87.1) 0.311 Gave verbal feedback 41 (95.3) 33 (85.4) 41 (82.5) 115 (87.8) 0.166 44 (91.7) 42 (93.3) 38 (80.9) 124 (88.6) 0.121 Provided written comments 18 (41.9) 16 (33.3) 18 (45) 52 (39.7) 0.505 16 (33.3) 24 (53.3) 18 (38.3) 58 (41.4) 0.128 Provided updates on technical and administrative issues 30 (69.8) 27 (56.2) 23 (57.5) 80 (61.1) 0.359 33 (68.8) 35 (77.8) 30 (63.8) 98 (70.0) 0.336 Discussed problems you encountered 37 (86) 40 (83.3) 36 (90) 113 (86.3) 0.663 42 (87.5) 42 (93.3) 42 (89.4) 126 (90.0) 0.634 Participated in quality of care improvement activities 43 (74.4) 48 (62.5) 40 (70) 90 (68.7) 0.462 33 (68.8) 37 (82.2) 30 (63.8) 100 (71.4) 0.131 25 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Routine Newborn and Maternal Care Performance of initial client assessment Complete answers for all questions regarding performance of initial client assessment tasks for mothers in labor were provided for 175 observations of deliveries in round 1 and 193 observations of deliveries in round 2 (Tables 16a and 16b). Although most mothers were checked for their client card, fetal presentation, fetal heart rate, and were given a vaginal exam, only about half of the mothers in both round 1 and round 2 had their temperature or pulse taken, and as few as 5.7 % in round 1 and 3.6% in round 2 were tested for urine protein. Only 32.6% and 24.4% of mothers were checked for fundal height in rounds 1 and 2, respectively. In round 1 a higher percentage of providers in the low-dose group asked about at least one danger sign, performed a general examination (anemia, edema, etc.), and checked fundal height. In round 2, a higher percentage of providers in the medium-dose group took the woman’s temperature and blood pressure and performed a general examination. Table 16a: Performance of Initial Client Assessment Tasks Observed in Round 1 Tasks High (n=64) Medium (n=51) Low (n=60) Total (n=175) p￾value n % n % n % N % Checks client card; asks age, length of pregnancy, and parity 63 98.4 50 98.0 58 96.7 171 97.7 0.791 Asks about at least one danger sign 52 81.2 37 72.6 57 95.0 146 83.4 0.006 Takes temperature 28 43.8 27 52.9 26 43.3 81 46.3 0.526 Takes pulse 30 46.9 27 52.9 30 50.0 87 49.7 0.810 Takes blood pressure 41 64.1 33 64.7 37 61.7 111 63.4 0.938 Asks about/notes urine output 15 23.4 4 7.8 5 8.3 24 13.7 0.018 Tests urine for protein 6 9.4 1 2.0 3 5.0 10 5.7 0.225 Performs general examination (anemia, edema, etc.) 47 73.4 25 49.0 46 76.7 118 67.4 0.004 Checks fundal height 17 26.6 12 23.5 28 46.7 57 32.6 0.015 Checks fetal presentation 62 96.9 49 96.1 58 96.7 169 96.6 0.972 Checks fetal heart rate with fetoscope/ ultrasound 64 100 48 94.1 58 96.7 170 97.1 0.164 Performs vaginal exam 62 96.9 50 98.0 59 98.3 171 97.7 0.848 Asks about at least one complication in previous pregnancies* 35 81.4 29 78.4 36 90.0 100 83.3 0.359 *The denominators were different for this question, as this was the first pregnancy for some women. 26 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 16b: Performance of Initial Client Assessment Tasks Observed in Round 2 Tasks High (n=59) Medium (n=67) Low (n=67) Total (n=193) p￾value n % n % n % N % Checks client card; asks age, length of pregnancy, and parity 58 98.3 64 95.5 63 94.0 185 95.9 0.479 Asks about at least one danger sign 46 78.0 49 73.1 46 68.7 141 73.1 0.501 Takes temperature 27 45.8 37 55.2 22 32.8 86 44.6 0.033 Takes pulse 37 62.7 41 61.2 24 35.8 102 52.9 0.003 Takes blood pressure 40 67.8 50 74.6 32 47.8 122 63.2 0.004 Asks about/notes urine output 7 11.9 16 23.9 11 16.4 34 17.6 0.200 Test urine for protein 1 1.7 4 6.0 2 3.0 7 3.63 0.414 Performs general examination (anemia, edema, etc.) 21 35.6 49 73.1 42 62.7 112 58.0 <0.001 Checks fundal height 15 25.4 13 19.4 19 28.4 47 24.4 0.470 Checks fetal presentation 53 89.8 62 92.5 65 97.0 180 93.3 0.264 Checks fetal heart rate with fetoscope/ ultrasound 59 100 65 97.0 66 98.5 190 98.5 0.401 Performs vaginal exam 57 96.6 67 100 67 100 191 99.0 0.101 Asks about at least one complication in previous pregnancies* 30 83.3 35 74.5 32 80.0 97 out of 123 78.9 0.604 *The denominators were different for this question, as this was the first pregnancy for some women. Provider communication and support Complete answers for communication and respectful, humanistic care before and during labor were provided for 188 and 199 observations in rounds 1 and 2, respectively (Tables 17a and 17b). In both round 1 and round 2, more than 85% of mothers were greeted respectfully, but less than 20% received encouragement to have someone in attendance at delivery. Less than 30% were provided with drapes or asked whether they had any questions for the health provider. Less than 60% were encouraged to hydrate and eat during labor. The only differences in performance by dose group were that providers in the high-dose arm in round 2 more often informed the pregnant woman of findings and in round 2 more often explained procedures to the woman. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 27 Table 17a: Communication and Support Tasks Before and During Labor, Round 1, by Dose Group High (n=64) Medium (n=59) Low (n=65) Total (N=188) p-value n % n % n % N % Initial assessment Respectfully greets pregnant woman 61 92.4 46 78.0 54 85.7 161 85.6 0.071 Encourages the woman to have a support person present throughout labor and birth 7 10.6 2 3.4 11 17.5 20 10.6 0.042 The woman has support person at some point during labor 19 28.8 13 22.4 9 14.8 41 22.2 0.163 Asks woman (and support person, if present) if she has any questions 14 21.2 19 32.2 18 28.6 51 27.1 0.367 Explains procedures to woman (and support person) before proceeding 55 83.3 45 76.3 53 84.1 153 81.4 0.473 Informs the pregnant woman of findings 62 93.9 46 78.0 55 87.3 163 86.7 0.031 Communication and support tasks for first stage of labor At least once, explains what will happen in labor to pregnant woman and her support person 40 62.5 34 57.6 49 75.4 123 65.4 0.096 At least once, encourages woman to consume fluids/food throughout labor 34 53.1 30 50.9 37 56.9 101 53.7 0.789 At least once, encourages/assists woman to ambulate and assume different positions during labor 44 67.7 44 74.6 44 74.6 132 70.2 0.670 Drapes woman 12 18.8 15 25.4 19 29.2 46 24.5 0.375 Table 17b: Communication and Support Tasks Before and During Labor, Round 2, by Dose Group High (n=61) Medium (n=69) Low (n=69) Total (N=199) n % N % n % N % p-value Initial assessment Respectfully greets pregnant woman 52 85.2 61 88.4 60 87.0 173 86.9 0.867 Encourages the woman to have a support person present throughout labor and birth 14 23.0 10 14.5 2 11.6 32 16.1 0.193 The woman has a support person at some point during labor 20 32.8 26 37.7 25 36.2 71 35.7 0.839 Asks woman (and support person, if present) if she has any questions 23 37.7 18 26.1 14 20.3 55 27.6 0.081 Explains procedures to woman (support person) before proceeding 58 95.1 54 78.3 53 76.8 165 82.9 0.010 Informs the pregnant woman of findings 55 90.2 64 92.8 60 87.0 179 90.0 0.525 28 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi High (n=61) Medium (n=69) Low (n=69) Total (N=199) n % N % n % N % p-value Communication and support tasks for first stage of labor At least once, explains what will happen in labor to pregnant woman and her support person 54 87.1 53 74.7 58 77.3 165 79.3 0.182 At least once, encourages woman to consume fluids/food throughout labor 37 59.7 47 62.7 54 76.1 138 66.4 0.096 At least once, encourages/assists woman to ambulate and assume different positions during labor 40 64.5 53 70.7 57 80.3 150 72.1 0.122 Drapes woman 19 30.7 22 29.3 15 21.1 56 26.9 0.392 29 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Prevention of Postpartum Hemorrhage Complete answers for all items pertaining to active management of the third stage of labor and other activities associated with prevention of postpartum hemorrhage were obtained for 1,417 observations in round 1 and 1,842 observations in round 2 (Tables 18a and 18b). Compliance was high, with approximately 91% and 98% of women receiving an uterotonic in rounds 1 and 2, respectively, and 94% and 97% of women receiving traction to the cord in rounds 1 and 2, respectively. In round 1 a higher percentage of women in the medium-dose group had a health worker apply traction to the cord while applying suprapubic counter traction and had a uterine massage immediately following the delivery of the placenta (p<0.001). In round 2, 90.8% of women in the high-dose arm had the placenta and membranes assessed, compared to 75% and 81% in the medium- and low-dose arms, respectively (p<0.001). Table 18a: Observation of Tasks Performed to Prevent Postpartum Hemorrhage, Round 1 Task High (n=469) Medium (n=445) Low (n=503) Total (n=1,417) p-value n % n % n % N % Administers uterotonic 444 96.7 413 92.8 427 84.9 1,284 90.6 <0.001 Applies traction to the cord while applying suprapubic counter traction 430 91.7 422 94.8 490 97.4 1,342 94.7 <0.001 Performs uterine massage immediately following the delivery of the placenta 406 86.6 398 89.4 329 65.4 1,133 80.0 <0.001 Assesses completeness of placenta and membranes 353 75.3 345 77.5 403 80.1 1,101 77.7 0.191 Assesses for perineal and vaginal lacerations 453 96.6 438 98.4 495 98.4 1,386 97.8 0.086 Table 18b: Observation of Tasks Performed to Prevent Postpartum Hemorrhage, Round 2 Task High (n=535) Medium (n=690) Low (n=617) Total (n=1, 842) p-value n % n % n % N % Administers uterotonic 527 98.5 680 98.6 609 98.7 1,816 98.6 0.955 Applies traction to the cord while applying suprapubic counter traction 514 96.1 676 98.0 603 97.7 1,793 97.3 0.094 Performs uterine massage immediately following the delivery of the placenta 489 91.4 619 89.7 558 90.4 1,666 90.5 0.607 Assesses completeness of placenta and membranes 486 90.8 517 74.9 502 81.4 1,505 81.7 <0.001 Assesses for perineal and vaginal lacerations 517 96.6 674 97.7 608 98.5 1,799 97.7 0.102 30 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Partograph use Complete answers for all items regarding partograph use were obtained for 986 observations in round 1 and 1,180 observations in round 2 (Tables 19a and 19b). More than 90% of the partographs had birth information recorded (time, delivery method, weight) in both rounds, but less than 1% of the partographs were properly/fully completed in round 2, compared to 9.4% in round 1. Only 1.8% had fetal heart rate, frequency/duration of contractions, and maternal pulse plotted every half-hour in round 2, compared to 12.2% in round 1. The high-dose group performed better on most indicators in round 1. In round 2, the high-dose group performed better than the other dose groups on plotting data at least every half-hour during labor and at recording birth information after delivery. Table 19a: Observed Use of Partographs, Round 1 Task High (n=359) Medium (n=275) Low (n=352) Total (n=986) p-value n % n % n % N % Partograph initiated at appropriate time 288 80.2 236 85.8 275 78.1 799 81.0 0.045 Data plotted at least every half￾hour during labor 19 5.3 1 0.4 100 28.4 120 12.2 <0.001 Birth information recorded after delivery 350 97.5 253 92.0 348 98.9 951 96.5 <0.001 Blood pressure recorded at least every four hours 86 24.0 20 7.3 108 30.7 214 21.7 <0.001 Filled in completely (1) 13 3.6 0 0 80 23.7 93 9.4 <0.001 Table 19b: Observed Use of Partographs, Round 2 Task High (n=320) Medium (n=414) Low (n=446) Total (n=1,180) p-value n % n % n % N % Partograph initiated at appropriate time 303 94.7 382 92.3 417 93.5 1,102 93.4 0.423 Data plotted at least every half￾hour during labor 12 3.8 7 1.7 2 0.5 21 1.8 0.003 Birth information recorded after delivery 306 95.6 399 96.4 392 87.9 1,097 93.0 <0.001 Blood pressure recorded at least every four hours 22 6.9 45 10.9 47 10.5 114 9.7 0.140 Filled in completely (1) 4 1.3 4 1.0 2 0.5 10 0.85 0.465 Provider performance of infection prevention practices before and after delivery Complete responses for all items regarding infection prevention behavior before delivery were obtained for 183 observations in round 1 and 202 observations in round 2 (Tables 20a and 20b). In both rounds more than 30% of providers washed their hands before conducting an examination during the first stage of labor, and more than 80% wore sterile gloves for vaginal examination. In round 1 53% wore protective clothing for delivery, compared to 30% in round 2. Round 1 providers in the low-dose group were more likely than providers in the other groups to wash their hands before performing an examination during the first stage of Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 31 labor and to wear protective clothing for delivery. In round 2, providers in the low-dose group were more likely to wear sterile gloves for vaginal examinations, and providers in the high dose group were more likely to wear protective clothing for delivery. Table 20a: Performance of Infection Prevention Practices before Delivery, Round 1 Provider Practice High (n=64) Medium (n=59) Low (n=60) Total (n=183) p￾value n % n % n % N % Washes hands before any examination in first stage of labor 15 23.4 23 39.0 31 51.7 69 37.7 0.005 Wears high-level disinfected or sterile gloves for vaginal exam 58 90.6 46 78.0 45 75.0 14 9 81.4 0.058 Puts on clean protective clothing in preparation for birth 25 39.1 31 52.5 41 68.3 97 53.0 0.005 Table 20b: Performance of Infection Prevention Practices before Delivery, Round 2 Provider Practice High (n=61) Medium (n=72) Low (n=69) Total (n=202) p￾value n % n % n % N % Washes hands before any examination in first stage of labor 16 26.2 29 40.3 18 26.1 63 31.2 0.116 Wears high-level disinfected or sterile gloves for vaginal exam 50 82.0 56 77.8 66 95.7 17 2 85.2 0.008 Puts on clean protective clothing in preparation for birth 26 42.6 11 15.3 24 34.8 61 30.2 0.002 Complete data regarding provider performance of infection prevention behavior after delivery were obtained for 1,411 observations in round 1 and 1,967 observations in round 2 (Tables 21a and 21b). Notably, a much larger percentage of providers washed their hands after delivery than before the delivery. While almost all providers in both rounds disposed of sharps and waste appropriately and decontaminated equipment, only about 28% in round 1 and 60% in round 2 wiped their aprons with chlorhexidine wipes. Nonbeneficial and nonindicated practices Among the 1,747 observations in rounds 1 and 2,109 observations in round 2, more nonbeneficial behaviors were reported in round 1 (5.6%) compared to round 2 (2.7%) (Tables 22a and 22b). The most frequently observed nonbeneficial practices were holding the newborn upside down, applying fundal pressure, and stretching the perineum. There were very few reports of health workers conducting nonindicated practices in either round (0.7% in round 1 and 0.3% in round 2. More nonbeneficial practices were reported in the high￾dose group in round 1 compared to the other two groups. There were no significant differences in reported nonbeneficial or nonindicated behaviors by dose in round 2. 32 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 21a: Performance of Infection Prevention Practices after Delivery, Round 1, by Dose Group Provider Practice High (n=471) Medium (n=431) Low (n=509) Total (n=1,411) p-value n % n % n % N % Disposes of all sharps in puncture-proof container immediately after use 470 99.8 428 99.3 479 94.1 1377 97.6 <0.001 Disposes of all contaminated waste in leakproof containers 427 90.7 376 86.2 488 95.9 1291 91.5 <0.001 Decontaminates all reusable instruments in 0.5% chlorine solution 411 87.3 396 91.9 472 92.7 1279 90.6 <0.008 Removes apron and wipes with 0.5% chlorine solution 106 22.5 62 14.4 237 46.6 405 28.7 <0.001 Washes hands thoroughly with soap and water 360 76.4 320 74.3 423 83.1 1103 78.2 <0.002 Table 21b: Performance of Infection Prevention Practices after Delivery, Round 2, by Dose Group Practice High (n=604) Medium (n=695) Low (n=668) Total (n=1967) p-value N % N % N % N % Disposes of all sharps in puncture-proof container immediately after use 599 99.2 689 99.1 656 98.2 1944 98.8 0.179 Disposes of all contaminated waste in leak proof containers 590 97.7 598 86.0 574 85.9 1762 89.6 <0.001 Decontaminates all reusable instruments in 0.5% chlorine solution 559 92.6 483 69.5 626 93.7 1668 84.8 <0.001 Removes apron and wipes with 0.5% chlorine solution 398 65.9 413 59.4 369 55.2 1180 60.0 0.001 Washes hands thoroughly with soap and water 565 93.5 611 87.9 599 89.7 1775 90.2 0.002 33 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Table 22a: Performance of Nonbeneficial and Nonindicated Practices, Round 1, by Dose Group Negative practices High Medium Low Total n % n % n % N % p-value n=577 n= 551 n=619 N=1747 Nonbeneficial practices Use of enema 0 0 1 0.2 0 0 1 0.1 0.338 Pubic shaving 0 0 0 0 0 0 0 0 - Applying fundal pressure 5 0.9 26 4.7 2 0.3 33 1.9 <0.001 Lavage of the uterus after delivery 2 0.4 0 0 1 0.2 3 0.2 0.371 Slapping newborn 1 0.2 1 0.2 2 0.3 4 0.2 0.830 Holding newborn upside down 25 4.3 5 0.9 5 0.8 35 2.0 <0.001 Milking the newborn’s chest 8 1.4 2 0.4 4 0.7 14 0.8 0.135 Stretching of the perineum 7 1.2 3 0.5 9 1.5 19 1.1 0.306 Had at least one non-beneficial practice 46 8.0 37 6.7 21 3.4 104 5.6 0.002 Nonindicated practices Manual exploration of the uterus after delivery 2 0.4 0 0 1 0.2 3 0.2 0.371 Use of episiotomy 1 0.2 0 0 4 0.7 5 0.3 0.098 Aspiration of newborn mouth and nose at birth 0 0 1 0.2 0 0 1 0.1 0.338 Restricting food and fluids in labor 1 0.2 1 0.2 2 0.3 4 0.2 0.830 At least one non-indicated practice 4 0.7 2 0.4 7 1.1 13 0.7 0.308 Table 22b: Performance of Nonbeneficial and Nonindicated Practices, Round 2, by Dose Group Negative practices High Medium Low Total n % n % n % N % p-value n=635 n= 742 n=732 N=2,109 Nonbeneficial Use of enema - - - - - - - - - Pubic shaving - - - - - - - - - Applying fundal pressure 1 0.2 11 1.5 7 1.0 19 0.9 0.034 34 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Negative practices High Medium Low Total n % n % n % N % p-value n=635 n= 742 n=732 N=2,109 Lavage of the uterus after delivery - - - - 1 0.1 1 0.1 0.390 Slapping newborn - - 5 0.5 4 0.6 8 0.4 0.177 Holding newborn upside down 4 0.6 9 1.2 8 1.1 21 1.0 0.525 Milking the newborn’s chest 3 0.5 2 0.3 6 0.8 11 0.5 0.334 Stretching of the perineum 3 0.5 5 0.7 5 0.7 13 0.6 0.857 Had at least one nonbeneficial practice 11 1.7 26 3.5 20 2.7 57 2.7 0.129 Nonindicated practices Manual exploration of the uterus after delivery - - 1 0.1 2 0.3 3 0.1 0.408 Use of episiotomy - - - - 2 0.3 2 0.1 0.152 Aspiration of newborn mouth and nose at birth - - - - 1 0.1 1 0.1 0.390 Restricting food and fluids in labor - - 1 0.1 - - 1 0.1 0.398 At least one nonindicated practice - - 2 0.3 5 0.7 7 0.3 0.085 35 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Immediate newborn care tasks Complete data were obtained for 1,303 observations of immediate newborn care in round 1 and 1,800 observations in round 2 (Tables 23a and 23b). Close to 70% of newborns had delayed cord clamping after birth, as recommended, more than 95% were immediately dried, and almost all newborns were either placed skin-to-skin or wrapped in a dry towel in both round 1 and round 2. In round 1, 42% of women started breastfeeding their newborns within one hour after birth, compared to 78% in round 2. Newborns in the medium-dose group in round 1 were more likely to be placed skin-to-skin or wrapped in a dry towel and breastfed within the first hour. Delayed cord clamping and cutting the cord with a clean blade were more prevalent in the high-dose group in round 1. In round 2, immediately drying the baby, delayed cord clamping, and initiation of breastfeeding in the first hour were more common in the low-dose group than the other groups. Table 23a: Performance of Newborn Care Tasks for Babies Breathing at Birth, Round 1, by Dose Group Provider Task High (n=339) Medium (n=409) Low (n=495) Total (n=1303) p￾value n % n % n % N % Immediately dries baby with towel 383 96.0 393 96.1 473 95.6 1,249 95.9 0.911 Discards wet towel 316 79.2 321 78.5 408 82.4 1,045 80.2 0.279 Places newborn skin-to-skin with mother 234 58.7 301 73.6 339 68.5 874 67.1 <0.001 Places skin-to-skin or wrapped with towel 393 98.5 406 99.3 493 99.6 1,292 99.2 0.194 Ties/clamps cord when pulsations stop, or within 2–3 minutes after birth (but not immediately after birth) 292 73.2 260 63.6 343 69.3 895 68.7 0.012 Cuts cord with clean blade 388 97.2 358 87.5 474 95.8 1,220 93.6 <0.001 Initiate breast feeding within the first hour 141 35.3 279 56.4 132 32.3 552 42.4 <0.001 Table 23b: Performance of Newborn Care Tasks for Babies Breathing at Birth, Round 2, by Dose Group Provider Task High (n=550) Medium (n=637) Low (n=613) Total (n=1,800) p￾n % n % n % N % value Immediately dries baby with towel 541 98.4 598 93.9 610 99.5 1,749 97.2 <0.001 Discards wet towel 521 94.7 548 86.0 546 89.1 1,615 89.7 <0.001 Places newborn skin-to￾skin with mother 382 69.5 457 71.7 447 72.9 1,286 71.4 0.417 Placed skin-to-skin or wrapped with towel 543 100. 0 633 99.4 611 99.7 1,775 99.0 0.003 36 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Provider Task High (n=550) Medium (n=637) Low (n=613) Total (n=1,800) p￾n % n % n % N % value Ties/clamps cord when pulsations stop, or within 2–3 minutes after birth (but not immediately after birth) 371 67.5 412 64.7 488 79.6 1,271 70.6 <0.001 Cuts cord with clean blade 437 79.5 532 83.5 436 71.1 1,405 78.1 <0.001 Initiate breast feeding within the first hour 388 70.6 479 75.2 543 88.6 1,410 78.3 <0.001 Newborn asphyxia management Observation of neonatal asphyxia management in round 1 Among the 1,747 babies that had valid answers pertaining to whether they were breathing at birth and the management of birth asphyxia, 88, 46, and 59 newborns in the high-, medium-, and low-dose intervention groups were found not to be breathing at birth (a total of 193 newborns). Among the babies that were not breathing at birth, 84 (95.5%) newborns survived in the high-dose arm, 42 (91.3%) survived in the medium-dose arm, and 54 (91.5%) survived in the low-dose arm (Figures 1, 2, and 3). Sixty-six newborns in the high-dose group (75%), 32 in the medium-dose group (69.6%), and 47 in the low-dose group (79.7%) received stimulation. Among the newborns receiving stimulation, 63.6%, 37.5%, and 52.2% in the high-, medium-, and low-dose groups, respectively, started breathing after stimulation. Bag and mask were administered to 23 of 24 non-breathing babies who were not revived after initial stimulation (95.8%) in the high-dose group, 19 out of 20 non-breathing babies (95%) in the medium-dose group, and 20 out of 22 non-breathing babies (90.9%) in the low-dose group. After the bag-and-mask intervention, 21 of 23 babies (91.3%) in the high-dose group, 18 of 19 (94.7%) in the medium-dose group, and 18 of 20 (90%) in the low-dose group survived. The differences in proportion between the groups were not statistically significant for any outcome. Observation of newborn asphyxia management in round 2 Out of the 2,093 babies that had valid answers pertaining to whether they were breathing at birth and the management of birth asphyxia, 91, 90, and 99 newborns in the high-, medium-, and low-dose intervention groups, respectively, were found not to be breathing at birth (a total of 280 newborns). Among the newborns that were not breathing at birth, 88 (96.7%) in the high-dose arm survived, 87 (95.6%) in the medium-dose arm survived, and 92 (92.9%) in the low-dose arm survived (Figures 5, 6, and 7). Sixty-eight newborns in the high-dose group (74.7%), 74 in the medium-dose group (81.1%), and 82 in the low-dose group (82.9%) received stimulation. Among the newborns receiving stimulation, 48.5%, 63.5%, and 52.4% of the newborns in the high-, medium-, and low-dose group, respectively, started breathing after stimulation. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 37 Bag and mask were administered to 29 out of 35 non-breathing babies who were not revived after initial stimulation (82.9%) in the high-dose group, 25 out of 26 non-breathing babies (96.2%) in the medium-dose group. and 32 out of 39 non-breathing babies (82.1%) in the low-dose group. After the bag-and-mask intervention, 28 of 29 babies (91.1%) in the high-dose group, 22 of 25 babies (88%) in the medium-dose group, and 27 of 32 babies (84.4%) survived. The differences in proportion between the dose groups were not statistically significant. 38 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 1: Management of Asphyxiated Newborns in High-Dose Intervention Facilities in Round 1 88 not breathing at birth (95.5%) 66 received stimulation 42 breathing 24 not breathing 22 (25%) received no stimulation 23 with bag and mask 1 without bag and mask 21 alive 2 neonatal deaths 1 alive 20 alive 2 neonatal deaths 39 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 2: Management of Asphyxiated Newborns in Medium-Dose Intervention Facilities in Round 1 46 not breathing at birth (91.3%) 32 received stimulation 12 breathing 20 not breathing 14 (30.4%) received no stimulation 19 with bag and mask 1 without bag and mask 18 alive 1 neonatal death 1 alive 11 alive 3 neonatal deaths 40 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 3: Management of Asphyxiated Newborns in Low-Dose Intervention Facilities in Round 1 59 not breathing at birth (91.5%) 46 received stimulation 24 breathing 22 not breathing 13 (22%) received no stimulation 20 with bag and mask 2 without bag and mask 18 alive 2 neonatal deaths 2 neonatal deaths 12 alive 1 neonatal death 41 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 4: Overall Management of Asphyxiated Newborns in Round 1 193 not breathing at birth 144 received stimulation 78 breathing 66 not breathing 49 (26.5%) received no stimulation 62 with bag and mask 4 without bag and mask 57 alive 5 neonatal deaths 2 alive 2 neonatal deaths 43 alive 6 neonatal deaths 42 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 5: Management of Asphyxiated Newborns in High-Dose Intervention Facilities in Round 2 91 not breathing at birth (96.7%) 68 received stimulation 33 breathing 35 not breathing 23 (25.3%) received no stimulation 29 with bag and mask 6 without bag and mask 28 alive 1 neonatal death 6 alive 21 alive 2 neonatal deaths 43 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 6: Management of Asphyxiated Newborns in Medium-Dose Intervention Facilities in Round 2 90 not breathing at birth (95.6%) 73 received stimulation 47 breathing 26 not breathing 17 (18.9%) received no stimulation 25 with bag and mask 1 without bag and mask 22 alive 3 neonatal deaths 1 alive 16 alive 1 neonatal death 44 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 7: Management of Asphyxiated Newborns in Low-Dose Intervention Facilities in Round 2 99 not breathing at birth (92.9%) 82 received stimulation 43 breathing 39 not breathing 17 (17.2%) received no stimulation 32 with bag and mask 7 without bag and mask 27 alive 5 neonatal deaths 7 alive 15 alive 2 neonatal deaths 45 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Figure 8: Overall Management of Asphyxiated Newborns in Round 2 280 not breathing at birth 223 received stimulation 123 breathing 100 not breathing 57 (20.4%) received no stimulation 86 with bag and mask 14 without bag and mask 77 alive 9 neonatal deaths 14 alive 52 alive 5 neonatal deaths 46 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Results Summary Consistent with the national distribution of health workers providing maternal and newborn health services in Malawi, about 60% of all health providers interviewed in both round 1 and round 2 were nurse/midwife technicians. More round 2 health workers (68.4%) than round 1 health workers (59.7%) received training in subjects related to newborn care in the past two years. Efforts by the MOH and the Nurses and Midwives Council of Malawi resulted in the inclusion of HBB training in nurse/midwife colleges between 2012 and 2013. Many of the trained health workers are nurse/midwife technicians (34.8% and 41.8% in rounds 1 and 2, respectively). There were no differences in health provider training by dose group in either round 1 or round 2. The majority of supervisors checked the records of their supervisees (81.7% and 88.8% in rounds 1 and 2, respectively) and observed their supervisees’ work (78.6% and 87.3% in rounds 1 and 2, respectively). However, only slightly more than one- third of supervisees (39.7% and 41.5% in rounds 1 and 2, respectively) had received any written comments from their supervisors. Most supervisees were able to discuss problems with their supervisor (86% and 90% in rounds 1 and 2, respectively) and were given verbal feedback (82.2% and 88.6% in rounds 1 and 2, respectively). In round 1 more supervisors in the medium-dose group than the other groups observed their supervisees’ work. There were no obvious differences in supervision by dose group in round 2. In clinical simulations, the mean number of steps (out of 10 possible) that were correctly completed was higher in round 2 (mean 7.1; SD 2.0) compared to round 1 (mean 6.2; SD 2.4). The overall mean score was higher in the high-dose arm than the medium-dose arm (p=0.041) in round 1. The overall mean score was not different by dose group in round 2; however, health workers in the high-dose arm scored higher on testing the function of the bag and mask and squeezing the bag harder. The providers were asked to complete two role-play case scenarios using the NeoNatalie model (newborn simulator). The mean number of steps performed correctly was lower in round 1 than in round 2 in both scenarios. Health workers in the high-dose arm had higher overall scores compared to the medium- and low-dose arms in round 1. In round 2, while the health workers in the medium-dose group scored higher on recognizing when a baby was not crying, calling for help, continuing ventilation, the overall mean score was not statistically significantly different from the high- or low-dose groups. The proportion of facilities that had HBB guidelines and the equipment required for newborn care increased over time. This is likely related to provision of newborn supplies and equipment as part of the national rollout of HBB. The distribution was not different by dose group, except for suction bulbs for mucus extraction, which were more common in the high-dose group in round 1 and more common in the medium-dose group in round 2. Health workers in the high-dose group performed better on some aspects of infection prevention (disposing of all contaminated waste, wiping their aprons, and washing their hands), partograph use (plotting data at least every half-hour during labor and recording birth information after delivery), prevention of postpartum hemorrhage (assessing placenta and membranes), and communication and support (explaining procedures to the woman). Observations of communication and respectful, humanistic care before and during labor indicated low quality in this aspect of care in general. In both round 1 and round 2 more than 85% of mothers were greeted Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 47 respectfully, but less than 20% received encouragement to have a support person in attendance at delivery. In both rounds, less than 30% of women were asked whether they had any questions for the health provider and less than 30% were provided with drapes. Less than 60% were encouraged to hydrate and eat during labor. The only differences in performance by dose group were that providers in the high-dose arm more often informed the pregnant woman of findings in round 1 and more often explained procedures to the woman in round 2. Very few nonbeneficial and nonindicated practices were reported, with the most frequent being holding the newborn upside down, applying fundal pressure, and stretching the perineum. Again, there were no significant differences in reported nonbeneficial or nonindicated behaviors by dose group in either round, although more nonbeneficial practices were reported in round 1 than round 2. Contrary to expectations, the key elements of immediate newborn care—immediately drying the baby, delayed cord clamping, and initiation of breastfeeding in the first hour—were more observed more frequently in the low-dose group than in the other groups in round 2. Similarly, in round 1, newborns were more likely to be placed skin-to-skin or wrapped in a dry towel and breastfed within the first hour in the medium-dose group compared to the high-dose group. Out of the 1,747 babies in round 1 for whom valid answers were obtained pertaining to whether they were breathing at birth and the management of birth asphyxia, 88, 46, and 59 the high-, medium-, and low-dose intervention groups, respectively, were found not to be breathing at birth (a total of 193 newborns). Among the newborns who were not breathing at birth in round 1, 84 (95.5%) in the high-dose arm, 42 (91.3%) in the medium-dose arm, and 54 (91.5%) in the low-dose arm survived (Figures 1, 2, and 3). Sixty-six (75%) in the high-dose arm, 32 (69.6%) in the medium-dose in the medium-dose arm, and 47 (79.7%) in the low-dose received stimulation. After stimulation, 63.6%, 37.5%, and 52.2% of the newborns in the high-, medium-, and low-dose group, respectively, started breathing after stimulation. Among those who were not breathing after initial stimulation, 23 of 24 babies (95.8%) in the high-dose group, 19 of 20 (95%) in the medium-dose group, and 20 of 22 (90.9%) in the low-dose group received bag-and-mask. After the bag-and-mask intervention, 21 of 23 (91.3%) babies in the high-dose group, 18 of 19 (94.7%) babies in the medium-dose group, and 18 of 20 (90%) babies in the low-dose group survived. The differences in proportion between the groups were not statistically significant for any outcome. Out of the 2,093 babies in round 2 for whom valid answers were obtained pertaining to whether they were breathing at birth and the management of birth asphyxia, 91, 90, and 99 newborns in the high-, medium-, and low-dose intervention groups were found not to be breathing at birth (a total of 280 newborns). Among the newborns who were not breathing at birth in round 2, 88 (96.7%) in the high-dose arm, 87 (95.6%) in the medium-dose arm, and 92 (92.9%) in the low-dose arm survived (Figures 5, 6, and 7). Sixty-eight in the high-dose group (74.7%), 74 in the medium-dose group (81.1%), and 82 in the low-dose group (82.9%) received stimulation. After stimulation, 48.5%, 63.5%, and 52.4% of the newborns in the high,- medium-, and low-dose groups, respectively, started breathing. 48 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Among those who were not breathing after stimulation, 29 of 35 (82.9%) babies in the high-dose group, 25 of 26 (96.2%) in the medium-dose group, and 32 of 39 (82.1%) in the low-dose group received a bag-and-mask intervention. After the bag-and-mask intervention, 28 of the 29 babies in the high-dose arm (91.1%), 22 of the 25 babies in the medium-dose arm (88%), and 27 of the 32 babies in the low-dose arm (84.4%) survived. The differences in proportion between the groups were not statistically significant for any outcome. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 49 Discussion Overall improvement in health worker training and knowledge, equipment availability, and management of labor and delivery, including newborn care, occurred over the two rounds of data collection in Malawi. However, there were no significant differences between the high-, medium-, and low-dose groups in the two rounds. Although health worker performance in round 2 was better than round 1, a majority of findings in this report indicate the absence of significant differences by dose group between the two rounds. These results were consistent irrespective of how districts were grouped. We conducted sensitivity analyses by regrouping the districts and analyzing the data in the following two ways, in addition to comparing the dose-response groups: 1. “Intention to treat” analyses were conducted in which the 13 districts identified as intervention districts during round 1 were grouped as intervention districts for round 2 as well, with the remaining 14 districts grouped comparison districts. 2. “Process documentation” analyses were conducted in which the districts were classified into two groups based on the timing of their HBB training and availability of training and implementation equipment. One group included districts that received training and implementation equipment immediately after training (IAT) or received training equipment IAT but implementation equipment one to 24 months after training. The other group included districts that received training equipment IAT but never received implementation equipment or received both training and implementation equipment one to 12 months after initial HBB training. The data presented in this report were collected for the first and second round of evaluation of the nationwide scale-up of the HBB initiative in Malawi. The overall findings of second round are similar to the first round, with significant differences found between comparison groups in some aspects of health worker knowledge and skills, but no difference in actual performance while managing newborns who were not breathing at birth. There are a few possible explanations for these findings: • The scale-up was implemented in a phased manner, but as identified in the HBB process documentation report, a majority of the facilities had a relative short duration of exposure to the HBB initiative due to delays in training and equipment availability. The nationwide scope and short duration of exposure presented logistical and managerial challenges to ensuring an adequate supply of the required equipment, guidelines, and appropriately trained staff in the facilities, which in turn might be reflected in the lower scores even in the high-dose group. • More specifically, the short duration of implementation and limited supervision are likely to be associated with limited opportunities for health providers to practice the newly acquired skills and assimilate the new knowledge and skills into actual performance. • The lack of significant differences between different dose groups might also be attributable to the transfer of HBB-trained health workers between high-, medium-, and low-dose facilities. This “contamination” might be associated with the similar distribution of HBB skills among the three dose groups. Several studies have reported on the effectiveness of training health workers in newborn resuscitation in reducing neonatal mortality attributable to birth asphyxia (Deorari et al. 2001; Carlo 2010; Vakrilova, Elleau, and Sluncheva 2005; Msemo et al. 2013). However, these studies also highlighted the need for implementation research as they were not nested within national health systems in developing countries (Lawn et al. 2011). While other studies use facility records instead of direct observations to evaluate HBB, we used direct observation to measure health workers performance in managing newborns who were not 50 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi breathing at birth. Findings from this study provide evidence on the performance of national-level delivery systems in achieving adequate intervention coverage and quality of a potentially efficacious intervention to reduce the number of newborn deaths due to birth asphyxia. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 51 Recommendations and Next Steps The findings from this evaluation suggest some recommendations for the Ministry of Health and donors to improve newborn care in Malawi. Due to the lack of major differences by dose group, the following recommendations and proposed steps apply to all districts and have been organized according to different domains of HBB intervention. Training • Partners should continue incorporating HBB into pre-service training at the remaining health worker training institutions in Malawi and expand the training to all skilled birth assistants (instead of focusing only on nurses and midwives) to ensure complete coverage. All cadres providing services to women in labor (medical assistants, clinical officers, doctors, nurse-midwives) should be included in the HBB training and mentoring programs to ensure maximum coverage and prevent missed opportunities in HBB. • During training, facilitators should emphasize to the trainees and District Health Management Team members the importance of transferring their skills to other providers at their health facilities who might not receive formal HBB training. This will help to ensure that all providers in the facilities implement HBB. Follow-up of providers after training • Following the HBB trainings, district-level plans should be developed to assess provider skills and provide onsite coaching and mentoring at six weeks, three months, and six months to promote retention of skills. The plans should be structured to include coaching and mentoring sessions for providers who did not attend training. • More emphasis needs to be placed on routine care of the newborn. During onsite coaching and mentoring, emphasis should be placed on providers internalizing the importance of discarding the wet towel, placing the baby in skin-to-skin contact with the mother, covering the baby with a dry towel, and initiating breastfeeding within one hour of birth. These are low-cost, high-impact interventions that contribute to reducing neonatal mortality. During resuscitation, emphasis should be placed on stimulation of the newborn, which seemed to be a major challenge in all facilities across the dose groups. • District nursing officers and district health officers should be involved in introducing the training program to the districts, and agreements should be made regarding the rotation of the trained providers and handling of new health care workers. The districts should be helped to develop and implement schedules for weekly practice sessions for new providers coming to the maternity department. • The district health officer, district nursing officer, and Safe Motherhood coordinator should include HBB in their routine supervision and coaching visits. • The Reproductive Health Division (RHD) of the MOH and partners should work toward ensuring that each facility has a NeoNatalie for practice sessions to promote the acquisition and maintenance of skills. • Mentors from the RHD and partners should structure integrated quarterly follow-up and supportive visits to the districts to support the district-based mentors and service providers. The follow-up should include assessing provider skills using the NeoNatalie or actual observation of and coaching on deliveries and assessing documentation, reporting, and utilization of data for decision making at district level. 52 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi Availability of guidelines and equipment • The MOH through RHD and partners should advocate for procurement and distribution of HBB equipment and supplies through Central Medical Stores Trust to ensure that HBB equipment is always available. • The RHD and partners should work together to make available HBB guidelines and guidelines for normal delivery in all facilities with maternity services. This will give providers reference materials to use when they are providing care. Supervision • It is important to have close supervision and mentorship at the district and facility levels, not only from the national headquarters. • Offer mentoring and capacity strengthening of the facility in-charges and supervisors on how to conduct supervision, as opposed to routine ward count and rounds. This training could be reinforced through a job aid/tool for supervisors. Supervisors’ HBB knowledge and skills also need to be improved to ensure that the supervisors are adequately knowledgeable about HBB and know the areas to look out for during supervision. Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi 53 References 1. Lawn JE, Kerber K, Enweronu-Laryea C, Cousens S. 2010. 3.6 million neonatal deaths—What is progressing and what is not? Seminars in Perinatology 34(6): 371–386. 2. Black RE, Cousens S, Johnson HL, et al. 2010. Global, regional, and national causes of child mortality in 2008: A systematic analysis. Lancet 375(9730): 1969–1987. 3. Bryce J, Boschi-Pinto C, Shibuya K, Black RE. 2005. WHO estimates of the causes of death in children. Lancet 365(9465): 1147–1152. 4. Lee AC, Cousens S, Wall SN, et al. 2011Neonatal resuscitation and immediate newborn assessment and stimulation for the prevention of neonatal deaths: A systematic review, meta-analysis and Delphi estimation of mortality effect. BMC Public Health 11 Suppl 3: S12. 5. Macro NSOMaO. 2010. Malawi Demographic and Health Survey 2010. Zomba, Malawi and Calverton, Maryland, USA: National Statistical Office and ORC Macro. 6. Deorari AK, Paul VK, Singh M, Vidyasagar D. 2001. Impact of education and training on neonatal resuscitation practices in 14 teaching hospitals in India. Ann Trop Paediatr 21(1): 29–33. 7. Carlo WA, Chomba E, Chakraborty H, Hartwell T, Harris H, Lincetto O, Wright LL. 2010. Newborn care training of midwives and neonatal and perinatal mortality rates in a developing country. Pediatrics 126(5): e1064–1074. 8. Vakrilova L, Elleau C, Sluncheva B. 2005. French-Bulgarian program “Resuscitation of the Newborn in a Delivery Room”: Results and perspectives. Akush Ginekol (Sofiia) 44(3): 35–40. 9. Msemo G, Massawe A, Mmbando D, Rusibamayila N, Manji K, et al. 2013. Newborn mortality and fresh stillbirth rates in Tanzania after Helping Babies Breathe training. Pediatrics 131: e353. 10. Lawn JE, Bahl R, Bergstrom S, Bhutta ZA, Darmstadt GL, et al. 2011. Setting research priorities to reduce almost one million deaths from birth asphyxia by 2015. PLoS Med 8(1): e1000389. 54 Evaluation of the Helping Babies Breathe (HBB) Initiative Scale-Up in Malawi